combining ict with eco-driving concepts to …...combining ict with eco-driving concepts to improve...

Combining ICT with Eco-driving Concepts to Improve Corporate Car Drivers’ Fuel Efficiency

D I S S E R T A T I O N of the University of St. Gallen,

School of Management, Economics, Law, Social Sciences

and International Affairs to obtain the title of

Doctor of Philosophy in Management

submitted by

Johannes Tulusan

from

Germany

Approved on the application of

Prof. Dr. Elgar Fleisch

and

Prof. Dr. Rolf Wüstenhagen

Dissertation no. 4246

Difo-Druck GmbH, Bamberg, 2013

The University of St. Gallen, School of Management, Economics, Law, Social Sci-ences and International Affairs hereby consents to the printing of the present disser-tation, without hereby expressing any opinion on the views herein expressed.

St. Gallen, October 21, 2013

The President:

Prof. Dr. Thomas Bieger

Acknowledgements I

Acknowledgements The research for this dissertation was jointly conducted with the Bits to Energy Lab (University of St. Gallen and ETH Zurich) and SAP Research from January 2010 to January 2013. The cross-institutional approach helped to enhance my research from both a theoretical and practical perspective through the application of theoretical concepts to a true global problem. Colleagues from both institutions provided valu-able feedback and guidance, which enabled me to complete my work.

I would like to thank my supervisor, Prof. Dr. Elgar Fleisch, for his ongoing sup-port. He helped create a rewarding work environment that nurtured this form of re-search. My sincere gratitude to Prof. Dr. Thorsten Staake, the Bits to Energy lab director, who enabled me to narrow down my research area, thoroughly discussed the experiments, and who provided regular feedback in a resourceful manner. And to Prof. Dr. Rolf Wüstenhagen, who already offered guidance in the first year of my doctoral studies.

From the industry I would like to thank my manager, Dr. Uli Eisert, the CFO of SAP Switzerland AG, Thomas Scheer, and the sustainability manager, Jonas Dennler; without their support and facilitation, research with the case study compa-ny, including access to the 340 corporate car drivers’ data sets would not have been possible.

I had fruitful discussions with, and received immense support from my colleagues at the Bits to Energy Lab and SAP Research. My sincere thanks to: Dr. Claire Michelle-Look, Dr. Tobias Graml, Michael Baeriswyl (from the Bits to Energy Lab), and Dr. Ali Dada, Dr. Marc Brogle, and Dr. Felix von Reischach (SAP Re-search). I hope these friendships will remain beyond my days as a doctoral student, and that we continue to inspire one another.

Finally, to my parents and two siblings, which have been invaluable during this challenging and timely journey. And to my fiancée, Barbara Kuzman, thank you for your unconditional encouragement, understanding, and love.

Munich, May 2013 Johannes Tulusan

II Abstract

Abstract The steady increase in energy consumption, largely due to the road transport sector, necessitates immediate action to curtail CO2 emissions. Regulations implemented by the European Commission dictate that the automotive industry should manufacture vehicles with reduced emissions output. In addition to these governmental regula-tions, individuals can address this problem by adopting a more eco-friendly driving style, as up to 30% of the fuel consumption rate is determined by their driving prac-tices.

This thesis aims to outline a way to influence corporate car drivers to improve their fuel efficiency and reduce CO2 emissions by using eco-feedback technologies. This sample is unique because unlike private car drivers who respond well to financial incentives for adopting sustainable driving behavior, corporate drivers are not fi-nancially motivated, as their fuel costs are reimbursed by their employer. Two feed-back technologies that provided details about the fuel consumption of each driver, an eco-driving smartphone application and a mobility information system, were ap-plied in an experimental setting. These technologies were tested in two field tests over a period of three months, involving a total of 340 corporate car drivers.

Analysis of the datasets, which included more than 20,000 fuel efficiency data points, showed an average improvement in fuel efficiency of 3.23% in the experi-ment with the eco-driving smartphone application and 1.24% respectively 1.69% in the second experiment with a mobility information system. Since no further mana-gerial guidelines or financial incentives were given, these findings indicate that eco-feedback technologies can play an important role in improving the overall fuel effi-ciency of a large corporate fleet. In total, the case study company could save up to 1.4 million Euros on fuel costs and reduce emissions by up to four tons of CO2 per year.

In 2013, the initiative will be rolled out to 11,000 corporate car drivers based in Germany. Furthermore, the technology will be prototyped, implemented as a service offered in fleet management software, and commercialized. This solution will be attractive to other corporations with large or expanding car fleets that are eager to reduce energy consumption and fuel costs by modifying the driving behavior of their employees without making large financial investments.

Zusammenfassung III

Zusammenfassung Der anhaltende Anstieg des Energieverbrauchs durch den Verkehrssektor verlangt nach sofortigen Maßnahmen, um den steigenden Ausstoß von CO2-Emissionen zu senken. Regulierungen seitens der Europäischen Kommission verlangen von der Automobilindustrie eine Produktion von Kraftfahrzeugen mit einem geringeren CO2-Ausstoß. Neben diesen Regulierungen können Autofahrer dem Problem entge-genwirken, indem sie einen nachhaltigeren Fahrstil erlernen. 30% des Benzin-verbrauchs kann vom Fahrverhalten jedes Einzelnen abhängen.

Das Forschungsvorhaben dieser Doktorarbeit ist es zu evaluieren, ob es möglich ist, Firmenwagenfahrer durch Feedback Technologien zu beeinflussen, den Benzin- und CO2-Verbrauch zu senken. Firmenwagenfahrer stellen eine besondere Gruppe dar, da die Benzinkosten von der Firma bezahlt werden und daher kein tatsächlicher An-reiz für ein nachhaltigeres Fahren besteht. Bei Privatwagenfahrern ist die finanzielle Einsparung der wichtigste Grund für eine nachhaltigere Fahrweise. In den Experi-menten wurden zwei Feedback Technologien, die individuelle Angaben zum Ben-zinverbrauch für jeden einzelnen Fahrer anzeigten, verwendet. Diese Technologien wurden in zwei Feldtests über eine Zeitspanne von drei Monaten und insgesamt 340 Firmenwagenfahrern getestet.

Die Analyse der Datensätze beinhaltet mehr als 20.000 Tankvorgänge. Eine Ver-besserung des durchschnittlichen Benzinverbrauchs um 3,23% wurde in dem ersten Experiment mit einer eco-driving Smartphone Applikation und 1,24% bzw. 1,69% in dem zweiten Experiment durch ein Mobilitätsinformationssystem erzielt. Da es keine Richtlinien seitens des Managements und finanzielle Anreize für die Teil-nehmer gab, zeigen diese Ergebnisse, dass eco-feedback Technologien eine wichti-ge Rolle bei der Verbesserung des durchschnittlichen Benzinverbrauchs großer Firmenwagenflotten spielen können. Insgesamt könnten bei der Fallstudie bis zu 1,4 Millionen Euro an Benzinkosten und bis zu vier Tonnen an CO2-Emissionen pro Jahr eingespart werden.

In diesem Jahr (2013) wird die Initiative bei 11.000 Firmenwagenfahrern in Deutschland erweitert. Zusätzlich wird ein Prototyp dieser Technologie entwickelt und als ein weiteres Serviceangebot in der bestehenden Flottenmanagement Soft-warelösung angeboten. Dieses Angebot wird vor allem für Firmen mit einem immer größer werdenden Fuhrpark relevant sein. Insbesondere kann - ohne eine hohe fi-nanzielle Investition zu tätigen - die Senkung des Energieverbrauchs und der Ben-zinkosten durch die Veränderung des Fahrverhaltens der Mitarbeiter erzielt werden.

IV Table of Contents

Table of Contents I Introduction ...................................................................................................... 1

I.1 Motivation ................................................................................................. 1 I.2 Objectives and Research Questions ............................................................ 7 I.3 Research Methodology in relation to Research Questions .......................... 8 I.4 Scope of the Thesis ...................................................................................16 I.5 Structure of the Thesis...............................................................................22

II Eco-driving and Eco-Feedback Technologies Related Work ............................25

II.1 Overview of Eco-driving ...........................................................................25 II.2 Differences between Private and Corporate Car Drivers ............................28 II.3 Eco-Feedback Technologies ......................................................................30 II.4 Smartphones Market Share........................................................................36 II.5 Impact of Eco-Feedback Technologies on the Driving Behavior ...............37 II.6 Feedback Intervention Theory ...................................................................39 II.7 Research Gap, ...........................................................................................41

III Private and Corporate Car Drivers’ Preferences for Eco-feedback Types .....43

III.1 Overview and Research Question ..............................................................43 III.2 Research Design and Research Methodology ............................................43 III.3 Data Evaluation and Findings ....................................................................44 III.4 Discussion .................................................................................................56 III.5 Conclusion ................................................................................................58 III.6 Limitations and Future Research ...............................................................58

IV Direct Feedback by an Eco-driving Smartphone Application, .......................61

IV.1 Overview ..................................................................................................61 IV.2 State of the Art and Related Work .............................................................62 IV.3 Research Questions and Hypothesis ..........................................................68 IV.4 Research Design and Research Methodology ............................................70 IV.5 Data Evaluation and Findings ....................................................................73 IV.6 Discussion .................................................................................................84 IV.7 Conclusion ................................................................................................86 IV.8 Limitations and Future Research ...............................................................86

V Indirect Feedback by a Mobility Information System ....................................89

V.1 Overview ..................................................................................................89 V.2 State of the Art and Related Work .............................................................90 V.3 Research Questions and Hypothesis ..........................................................98 V.4 Research Design and Research Methodology .......................................... 101 V.5 Data Evaluation and Findings .................................................................. 104

Table of Contents V

V.6 Discussion ............................................................................................... 106 V.7 Conclusion .............................................................................................. 107 V.8 Limitations and Future Research ............................................................. 108

VI Recommendations to Increase Corp. Car Drivers’ Intrinsic Motivations .... 109

VI.1 Overview and Research Questions .......................................................... 109 VI.2 Extrinsic and Intrinsic Motivation ........................................................... 109 VI.3 Research Design and Research Methodology .......................................... 110 VI.4 Data Evaluation and Findings .................................................................. 112 VI.5 Conclusion .............................................................................................. 119 VI.6 Limitations and Future Research ............................................................. 119

VII Development of an Eco-driving Dashboard System .................................... 121

VII.1 System Architecture, Data Structure, and Data Processing ...................... 121 VII.2 Mock-ups ................................................................................................ 123 VII.3 Mock-up incorporated in SAP´s Portal .................................................... 125 VII.4 Drawbacks of the current eco-driving System ......................................... 126

VIII Conclusion, Recommendations, Outlook and Future Research ................ 129

VIII.1 Key Findings ........................................................................................... 129 VIII.2 Theoretical Contributions ........................................................................ 133 VIII.3 Practical Implications .............................................................................. 135 VIII.4 Recommendations and High Level Concept ............................................ 137 VIII.5 Limitations .............................................................................................. 140 VIII.6 Future Research ...................................................................................... 141

Appendix ............................................................................................................. 143

References ........................................................................................................... 171

Curriculum Vitae ................................................................................................. 181

VI List of Figures

List of Figures Figure 1: EU CO2 Emissions per sector (left) and transport mode (right) ................. 1

Figure 2: Research Process and Relation of Research Questions to Methods ......... 13

Figure 3: 2011 SAP AG’s global distribution of CO2 Emissions ............................ 17

Figure 4: Fuel Efficiency Calculation..................................................................... 19

Figure 5: Research Perspectives ............................................................................. 20

Figure 6: Structure of the Thesis ............................................................................ 22

Figure 7: Feedback Technologies and Response of Driver ..................................... 31

Figure 8: Ford’s SmartGauge in-car Interface14 ...................................................... 31

Figure 9: Eco-driving Applications Market Share .................................................. 32

Figure 10: Bliss Trek Application Interface ........................................................... 33

Figure 11: greenMeter Application Interface .......................................................... 34

Figure 12: Color changing Feedback ...................................................................... 34

Figure 13: Green Driving App ............................................................................... 35

Figure 14: Smartphone Market Share Outlook ....................................................... 37

Figure 15: Research Design ................................................................................... 44

Figure 16: Categorization Process of Open-ended Questions ................................. 45

Figure 17: Awareness of Feedback Devices ........................................................... 46

Figure 18: Changes while using a Feedback Device ............................................... 47

Figure 19: Preferences for Feedback Types ............................................................ 48

Figure 20: Preferences for Feedback Technologies ................................................ 50

Figure 21: Reasons for Favoring the Color-changing Feedback System ................. 51

Figure 22: Reasons for Not Favoring the PC Application Feedback System .......... 52

Figure 23: Suggestions for Additional Eco-efficient Feedback ............................... 53

Figure 24: Average Kilometers driven per Month .................................................. 54

Figure 25: Preference of Feedback Types to improve Fuel Efficiency .................... 55

Figure 26: Understanding of the Meaning of 100 grams of CO2 per km ................. 56

Figure 27: Preferences for Feedback Types, Private vs. Corporate Car Drivers ...... 57

Figure 28: Overview of DriveGain Application Interface ....................................... 65

Figure 29: Settings Screen ..................................................................................... 66

Figure 30: Select Vehicle Screen ........................................................................... 66

Figure 31: Advanced Savings Feedback ................................................................. 66

Figure 32: Fuel Savings Feedback ......................................................................... 66

Figure 33: Detailed Journey View shown in Online Driving Portal ........................ 68

Figure 34: Between-subjects Experimental Design ................................................ 71

List of Figures VII

Figure 35: Influence on Driving Behavior .............................................................. 76

Figure 36: Direct – vs. Indirect Feedback ............................................................... 80

Figure 37: Frequency of Journey Score reviewed in the ODP ................................ 81

Figure 38: Preference of when Feedback should be received ................................. 82

Figure 39: Preference of how Feedback should be provided................................... 83

Figure 40: Preference of Social Comparisons ......................................................... 84

Figure 41: Verbal Descriptor Scale (VDS) ............................................................. 94

Figure 42: Visual Analogue Scale (VAS) ............................................................... 95

Figure 43: Verbal Graphic Rating Scale (GRS) ...................................................... 95

Figure 44: Verbal and Numerical Graphic Rating Scale (GRS) .............................. 95

Figure 45: Normative Feedback to improve Energy Consumption ......................... 96

Figure 46: Energy Efficiency Rating Scale (EU directive 2002/91/EC) .................. 96

Figure 47: Categorical and Continual Feedback Scale Formats .............................. 97

Figure 48: Experimental Design ........................................................................... 101

Figure 49: Monthly Feedback Email with Discrete or Continual Feedback .......... 102

Figure 50: Interview Data Analysis ...................................................................... 111

Figure 51: Eco-driving Dashboard Architecture ................................................... 122

Figure 52: Driver´s Master Data Types ................................................................ 122

Figure 53: Eco-driving Dashboard Architecture ................................................... 123

Figure 54: Corporate Car Driver´s Eco-driving Dashboard .................................. 124

Figure 55: Management´s Corporate Fleet Dashboard.......................................... 125

Figure 56: Management´s Corporate Fleet Dashboard.......................................... 126

Figure 57: Audi navigation system (under development) ..................................... 144

Figure 58: Ford Gauge ......................................................................................... 145

Figure 59: greenMeter iPhone application ............................................................ 145

Figure 60: Gauge in cars dashboard ..................................................................... 147

Figure 61: Screen on top of the car dashboard ...................................................... 148

Figure 62: iPhone Application ............................................................................. 148

Figure 63: PC Application ................................................................................... 149

Figure 64: iPhone Application ............................................................................. 149

Figure 65: Audi navigation system (under development) ..................................... 150

Figure 66: Description in iTunes Store ................................................................. 154

Figure 67: Main Screen ........................................................................................ 154

Figure 68: Setting Screen ..................................................................................... 154

Figure 69: Add Vehicle ........................................................................................ 154

Figure 70: Enable Upload of Data ........................................................................ 155

VIII List of Figures

Figure 71: Shift to N Gear .................................................................................... 155

Figure 72: Shift to 1st Gear ................................................................................... 155

Figure 73: Shift to 2nd Gear .................................................................................. 155

Figure 74: Shift to 3rd Gear .................................................................................. 156

Figure 75: Fuel Savings Meter ............................................................................. 162

Figure 76: Advanced Savings Meter .................................................................... 162

List of Tables IX

List of Tables Table 1: Mixed-method Research Methodologies .................................................... 9

Table 2: SAP Corporate Car Drivers’ Data Set Example ........................................ 19

Table 3: Advantages and Disadvantages of Eco-Driving ........................................ 26

Table 4: Differences between Private and Corporate Car Drivers ........................... 29

Table 5: Overview of Eco-driving Applications ..................................................... 35

Table 6: Five Arguments Feedback Intervention Theory ........................................ 39

Table 7: Private Car Driver’s Awareness of CO2 Emissions ................................... 49

Table 8: Sample Demographics ............................................................................. 74

Table 9: Group Statistics ........................................................................................ 78

Table 10: Independent Samples Test (t-test for Equality of Means)........................ 78

Table 11: Independent Samples Test (Levene’s Test for Equality of Variances) .... 78

Table 12: Correlation of Direct or Indirect Feedback vs. Duration ......................... 79

Table 13: Illustration of Hypothesis 1 .................................................................... 99

Table 14: Illustration of Hypothesis 2 .................................................................... 99

Table 15: Illustration of Hypothesis 3 .................................................................. 100

Table 16: Fuel Efficiency Rating and Intervals .................................................... 103

Table 17: Paired Samples T-Test TG1 ................................................................. 104



Table 20: Categories of Structured Analysis ........................................................ 112

Table 21: Summary of Feedback Intervention Theory Arguments ....................... 115

Table 22: Recommendations and High-Level Concept ......................................... 138

X Abbreviations

Abbreviations

Cronbach's Alpha ANOVA Analysis of Variance AT Attitude Avg_fe_y Average fuel efficiency total year Avg_fe_trtmt_t Average fuel efficiency treatment time Baseline Baseline BI Behavioral Intention BIC Behavioral Intention to Change B2E Bits to Energy Lab CAN-bus Controller Area Network Bus CO2 Carbon Dioxide CG Control Group CPM Computer Performance Monitoring date_fil Date of the filling Df Degree of freedoms in a t-test diff_fe_percentage Percentile difference between fe_filling and fe_oem DS Driving Style EMEA Europe, the Middle East and Africa ETH Swiss Federal Institute of Technology EU European Union EUR Euro FE Fuel Efficiency fem1 First Feedback EMail fem2 Second Feedback EMail fem3 Third Feedback EMail fe_filling Fuel Efficiency in Liter per 100km (I_fil / km_drivn) x

100=fe

fe_filling_oem_difference Difference between fe_filling and fe_oem fe_oem Average Fuel Efficiency defined by car manufacturer fe_type Fuel Type of the Filling FI Feedback Intervention FIT Feedback Intervention Theory FT Feedback Technologies F&S Frost and Sullivan

Abbreviations XI

Gart Gartner gCO2/km Gram CO2 per kilometer GPS Global Positioning System GWh Gigawatt hours HCI Human Computer Interaction HSG University of St. Gallen ICT Information Communication Technology Id ID number of a driver IS Information System IT Information Technology km_fil Total amount of km at the time of the filling kTons Kilo Tons l/100km Liters of fuel needed for 100km l_fil Amount of liters filled up (in liters) Mil Million Mtoe Million Tons of Oil Equivalent N Total amount n (%) Total amount in percent n_fillings Number of fillings per driver pers_max Maximizer Personality pers_satis Satisficer Personality P Significance level PBC Perceived Behavioral Control PRQ Principal Research Question Q Quarter RPM Revolution per Minute RQ Research Question Sig. Significance SQ Sub Question Std. Dev. Standard Deviation survey_answd Survey answered T Time TG1 Treatment Group 1 TG2 Treatment Group 2 TPB Theory of Planned Behavior TRA Theory of Reasoned Action

XII Abbreviations

Trtmt Treatment USD US-Dollar

Introduction 1

I Introduction

I.1 Motivation

I.1.1 Practical Relevance

In 2010, energy consumption increased by 5.5% to 12.9 Mtoe (Million Tons of Oil Equivalent) and reached the highest consumption rate to date (Enerdata, 2011). One approach to address this alarming trend would be to reduce the consumption of fos-sil fuels in the transport sector (European Commission, 2010). In Europe, for exam-ple, 23% of carbon dioxide (CO2) emissions were attributable to the transport sector (see Figure 1, left pie chart), and 71% of these resulted from road transport (see Figure 1, right pie chart). In comparison, marine and aviation transport only con-tribute 15% and 12%, respectively.

Figure 1: EU CO2 Emissions per sector (left) and transport mode (right)1

Within road transport, much of the contribution comes from drivers who use their car for business-related travel; this group is known as corporate car drivers. As highlighted in the latest report by Dataforce (2011) (“Fuhrparkmanagement in deutschen Unternehmen 2011”), one quarter of the three million newly registered

1European Commission. (2010), “Road transport: Reducing CO2 emissions from light-duty vehicles.” Re-

trieved from http://ec.europa.eu/clima/policies/transport/vehicles/index_en.htm

2 Introduction

cars in Germany are corporate cars, representing a rise from 700,000 in 2011 to 750,000 in 2012. In Switzerland the ratio is even higher: in 2011 one third of the 319,000 newly registered cars were corporate cars, a rise of 8.4% compared to 2010 (Bahnmüller, 2012; Ullmenstein, 2012). The number of corporate cars is predicted to rise further in Europe in the coming years (Dataforce, 2011). Corporate car driv-ers in Europe drive 35,000 kilometers per year on average, which is 21,500 kilome-ters per year more than private car drivers. Their fuel costs are reimbursed by their company (DTLR, 2001), so they lack the financial motivation that influences pri-vate car users to improve their fuel efficiency.

Government legislation is one efficient means of reducing CO2 emissions in the personal transportation sector. In Europe, Environmental Zones help reduce traffic congestion in city centers. Germany has forty-two Environmental Zones spread across city centers; an obligatory environmental badge specific to the vehicle type, known as a green badge, is required to enter these designated ‘green’ zones (Cli-mate Company, 2011). In 2007, a new European regulation set limits on the average CO2 emissions permitted from the yearly total sum of new cars produced by each car manufacturer. By 2012, each car manufacturer should not exceed more than 120 grams per km (gCO2/ km) for their yearly total sum of produced cars. This is a re-duction of 25% when compared to emission rates in 2006 (European Commission, 2010). A further reduction target for manufacturer’s average fleet emissions has been set at 70 gCO2/ km by 2025.

A report from the National Research Council (2010) appraised technologies that could improve vehicles’ fuel economy in order to meet the European Commission’s CO2 fleet regulations. Factors affecting fuel consumption include vehicle weight, engine efficiency, transmission type, automatic start / stop function, brake regenera-tion, tire type, air conditioning, and cruise control (National Research Council, 2010). Accommodating all of these features would require purchasing a new vehi-cle, a step many private car drivers may be unwilling or unable to undertake solely for the benefit of reducing overall CO2 emissions (Gardner and Stern, 2008). In cor-porations such as SAP, eligible corporate-car-driving employees are allowed to pur-chase a new car every four years and drivers must purchase a car falling under a maximum CO2 emissions threshold of 190 gCO2/ km. Drivers willing to buy a car

Introduction 3

with emissions below 150 gCO2/ km receive a financial incentive in the form of reduced price for the car (e.g., ‘EcoBonus’ offers a 5% reduction in list price).

In addition to government and corporate legislations and improvements in vehicle technology, drivers can influence their CO2 emissions rate without purchasing a new car by adopting a more eco-friendly driving style. Driving style can influence fuel consumption either positively or negatively; aggressive driving habits, for example, can lead to a 30% increase in fuel consumption (Romm and Frank, 2006). One ex-ample of an intervention aimed at improving individuals’ driving styles is an eco-driving training program in which drivers learn techniques to minimize fuel con-sumption (Boriboonsomsin et al., 2010). Eco-driving is an economical approach to reducing fuel consumption that can be utilized regardless of the vehicle type and is typically taught through classroom instruction and a practical training session over the course of a single day. The underlying ethos of eco-driving is to introduce posi-tive attributes into an individual’s daily driving habits, thus encouraging an average reduction in fuel consumption of up to 15% (Onoda, 2009). Additional benefits could include reductions in greenhouse gas emissions, petrol expenses, wear and tear of the vehicle, and improved road safety resulting in fewer vehicle collisions (GreenRoad, 2008; Onoda, 2009). However, some studies have uncovered a 5 - 10% deterioration in fuel efficiency one month after participating in eco-driving training, as drivers resorted back to their old driving habits (Wahlberg, 2007). Con-sequently, considerable post-training measures are needed after eco-driving courses to ensure that over the long term drivers can change actions that have become au-tomatic and are performed without conscious thought.

In addition to classroom and onsite training, eco-driving advice can be provided through brochures, driving simulators, websites, and in-vehicle eco-driving systems. The results of a field test conducted with twenty drivers who received instantaneous feedback from an in-vehicle eco-driving system indicated an improvement in fuel efficiency of 1% on highways and 6% on city streets (Boriboonsomsin et al., 2010). With the increasing popularity of smartphone technologies (as demonstrated, for example, by an estimated increase in mobile-phone market share from 11% in 2008 to 40% in 2014 (Gartner, 2010)), one cost-effective intervention may be eco-driving smartphone applications. These applications inform drivers about their driving style either following eco-driving training or in the absence of such training, and those

4 Introduction

currently on the market cost less than five Euros per installation. The latest smartphone processors and sensing technologies, such as GPS and accelerometer sensors, are able to provide feedback according to the driving context.

Another way to provide feedback is by increasing drivers’ awareness of fuel con-sumption via email, to which every employee has access. This could be beneficial in a corporate environment in which drivers are normally not informed about their av-erage fuel consumption, fuel costs, or CO2 emissions unless they independently note such details. Some corporations also have fleet management systems in place to track their fleet’s contractual agreements (e.g., fuel costs, car maintenance cycles, fuel tank data, and accident reports) but do not share this information with their em-ployees.

Evidence suggests that a reduction in CO2 emissions brings both ecologic and eco-nomic advantages. The German Federal Environmental Agency (2009) estimated that an improvement in fuel efficiency of one liter per 100 km leads to total finan-cial savings of 2,500 to 4,200 Euros (for yearly travel of 15,000 km or 25,000 km, respectively) over the twelve-year lifespan of a car. This calculation assumed an estimated petrol price of 1.40 Euro per liter; if fuel prices per liter increase in the upcoming years, the savings would also increase. These savings are particularly relevant for companies with large car fleets.

In summary, to reduce the CO2 emissions by lowering fossil fuel consumption in the road transport sector, corporate car drivers should be targeted because their mileage exceeds that of other groups. Even if vehicles with the latest fuel efficiency tech-nologies are purchased, a major fuel consumption factor that remains unaddressed is drivers’ driving style (Romm and Frank, 2006). To attain a better understanding of how to motivate corporate car drivers to drive more sustainably with the support of an eco-driving feedback information system (IS), it is important to consider the in-creasing size of this driver group in Europe.

Introduction 5

I.1.2 Theoretical Relevance2

From a theoretical perspective, studies exist that describe social, psychological, and structural factors influencing changes in transportation modes (see Bamberg et al., 2003; Forward, 2004; Graham et al., 2011), but difficulties persist in making a clear link between a given feedback and a behavioral modification. Kluger and DeNisi (1996) used the Feedback Intervention Theory (FIT) to evaluate how feedback in-terventions (FI) can be applied to influence employees’ task performance and initi-ate changes in their behavior within an organizational setting. They defined the term ‘feedback-standard gap,’ which denotes the comparison between the given feedback and the existing standard, such as a common practice in an organization. If the giv-en feedback does not align with the given standard, employees tend to reject the FI (Kluger and DeNisi, 1996). Kluger and DeNisi (1996, p.278) highlighted that fur-ther research needs to be conducted “to establish the circumstances under which

positive FI effects on performance are also lasting and efficient.”

Alder (2007) extended the FIT by using computer-mediated feedback through a computer performance monitoring information system (CPM IS) to evaluate chang-es in employee performance in the workplace. He found that changes were only ef-fective if the CPM FI was not perceived as a control mechanism but rather as a sup-port tool (Alder, 2007). A CPM IS has key benefits for employers and employees: it is objective and improves performance evaluation (Angel, 1998; Worsnop, 1993). On the downside, a CPM IS may be seen as an invasion of employees’ privacy, con-tribute to decreased job satisfaction, and increase stress (Greengard, 1996). In a study conducted by Van Mierlo et al. (2004), subjects demonstrated the ability to reduce their fuel consumption by up to 6% with the support of a prototype feedback device that provided clear and concise advice on a screen without creating more work for the driver. Although this example is limited to a simulated test environ-ment, it offers important evidence of the potential of feedback technologies for promoting fuel efficiency. These technologies could enable drivers to gain a better understanding of eco-driving practices without attending a daylong training session.

2 Published paper: Tulusan, J., Steggers, H., Staake, T., Fleisch, E., Supporting eco-driving with eco-feedback technologies: Recommendations targeted at improving corporate car drivers’ intrinsic motivation to drive more sustainable, Energy Informatics 2012 (EI 2012), Atlanta, Georgia, United States, October 2012.

6 Introduction

In contrast, Lee et al. (2010) found a deterioration in fuel efficiency as drivers’ task loads were increased as a consequence of using an eco-driving support system.

In addition to feedback technologies, the appropriate time at which feedback should be given to corporate car drivers needs to be further explored. The literature on en-ergy savings in households shows that direct or real-time feedback has a positive impact on residents, as they get an immediate response to their energy consumption and are able to act accordingly (Graml et al., 2010; Winkler and Winett, 1982). In relation to driving, direct feedback can improve fuel efficiency but also distract the driver, increase the risk level, and even lead to increased fuel consumption (Donmez, 2007; Meschtscherjakov et al., 2009). Indirect feedback, on the other hand, refers to decontextualized types of feedback such as a detailed breakdown of a driver’s fuel consumption, CO2 emissions, or acceleration/braking patterns. This type of feedback is offered after driving behavior has been monitored and analyzed. A study of college students by Graham et al. (2011) targeted a reduction in their daily car usage, and the results revealed that providing them with an online IS showing their financial and CO2 emissions savings facilitated an overall reduction in their car usage.

Finally, a driver’s motivation has a large impact on reducing their overall fuel con-sumption. Among the studies researching the effectiveness of eco-driving interven-tions, Johansson et al. (2003) found that the pre-existing motivation of individuals to reduce their fuel efficiency was essential. Since improvements in fuel efficiency do not directly result in financial benefits for corporate car drivers, different motiva-tions need to be evaluated in order to understand which factors encourage these drivers to adapt their driving style.

To summarize, this research builds on the extant literature by evaluating the feed-back provided by feedback technologies. Feedback through information communi-cation technology (ICT) is recognized as part of the concept of a CPM IS, as evalu-ated by Nussbaum and duRivage (1986) and Alder (2007). A CPM IS enables or-ganizations to provide their employees with information about their personal per-formance by collecting relevant data when they are working. This improves their awareness of areas for development and enables them to modify their behavior when necessary (Grant & Higgins, 1989). The FIT from Kluger and DeNisi (1996)

Introduction 7

can be used to evaluate changes in task performance (i.e. changes in driving behav-ior) as they relate to feedback technologies. Yet there remains a gap in this litera-ture, as it does not yet consider how feedback technologies such as smartphone ap-plications or a CPM IS can promote eco-driving amongst corporate car drivers by informing them about their individual fuel consumption through direct or indirect feedback. In this way, corporate car drivers’ intrinsic motivation must be targeted in order to encourage long-term sustainable driving behavior.

I.2 Objectives and Research Questions

These theoretical and practical motivations suggest a discrepancy surrounding the potential for feedback technologies to instill eco-friendly driving habits in their us-ers. While the approval of many in-vehicle feedback systems largely reflects a posi-tive general outlook for the technology, only a small number of current studies have assessed feedback delivered through external feedback technologies such as smartphones or fleet management information systems. There are no studies that investigate the use of these feedback technologies with corporate car drivers over a longer period of time. Corporate car drivers maintain their unique status as research subjects because their vehicle expenses, such as fuel and car maintenance costs, are covered by their employing organization. Monetary incentives and faster car depre-ciation have little authority to encourage them to behave sustainably. As these influ-encing factors are insignificant within this sample population, the extent to which feedback technologies stimulate eco-friendly driving can be directly evaluated.

Using eco-driving feedback technologies in the corporate context has great potential to support the success of corporate eco-driving that can reduce a company’s overall CO2 emissions and fuel costs. This is especially important for companies with large car fleets and is relevant at the continental European level, given the constant in-crease in the number of corporate cars (Dataforce, 2011). The growing acceptance of smartphone technologies within organizations and universal employee access to email present a cost-effective and time-efficient alternative to participating in eco-driving training. This research study has both theoretical and practical relevance and draws on identified gaps in the body of research appraised to formulate the follow-ing principal research question:

8 Introduction

How can eco-feedback technologies influence the average fuel efficiency of cor-porate car drivers when monetary incentives are not given?

To assist in answering the principal research question, the following sub-questions (SQ) must be addressed:

SQ1: Which eco-driving feedback technologies and feedback types are preferred by private and corporate car drivers?

SQ2: How does an eco-driving smartphone application influence corporate car drivers’ average fuel efficiency?

SQ3: Which feedback type, direct or indirect, do corporate car drivers prefer when using an eco-driving smartphone application?

SQ4: How does social-normative feedback about a driver’s fuel consumption via email influence corporate car drivers’ average fuel efficiency?

SQ5: What recommendations can help an organization increase corporate car driv-ers’ intrinsic motivation to improve their fuel efficiency?

I.3 Research Methodology in relation to Research Questions

I.3.1 Mixed-Methods Research Methodology3

This chapter offers a description of the epistemological viewpoint and research methodologies that prove most appropriate for the research. Epistemology refers to a system of assumptions about the nature of knowledge, its empirical material, and how it can be acquired (Flick, 2006). Positivistic, anti-positivistic, interpretivistic, and critical forms all represent different paradigms of research epistemologies (Orlikowski and Baroudi, 1991). These different approaches not only influence the way in which a researcher views reality but also have implications for the choice of research aims and methods (King et al., 1994). The positivistic/quantitative school of thought claims that all true knowledge is scientific and measurable (Black, 1999). Research methodologies in this field observe, prove, and measure phenomena by

3 Course work: Tulusan, J., MP-02 Qualitative Research Methods Lecture, Judge Business School, Cam-bridge University, England, October 2008.

Introduction 9

using statistical or mathematical techniques and testing theories according to hypo-thetical-deductive models, laboratory experiments, or surveys.

On the contrary, Merton (1968) argues that not everything is scientifically proven or possible to prove. Toulmin (1960, p.190) claims that a modern scientific researcher should “not only concentrate on abstract and universal questions but to treat again specific, concrete problems….” Researchers should thus demand the explanations behind the numbers and seek outcomes that are applicable to social reality (Lawson, 2008). The interpretivistic/qualitative school of thought is founded on the assump-tion that “systematic enquiry must occur in a natural setting rather than an artifi-cially constrained one such as an experiment” (Marshall and Rossman, 1989, p. 10-11). Interpretivistic approaches search for meaning based on the social phenomena and context in which the research takes place. An interpretivist tries to explore phe-nomena such as human actions by making sense of the meanings symbolized by those actions (Flick, 2006).

In this research, a mixed-methods research approach based on both positivistic/ quantitative and interpretivistic/ qualitative paradigms was chosen to: a) quantify the empirical data from the experimental research and b) obtain a better understand-ing of eco-feedback technologies from the point of view of the driver and in the context of broader social phenomena. Mixed-methods research approaches apply different research methodologies within the same research design, and the chosen methods should complement each other (Flick, 2006; Strauss and Corbin, 1998). The research methodologies detailed in Table 1 were therefore applied.

Table 1: Mixed-method Research Methodologies

Qualitative Methodologies Quantitative Methodologies Desk research to obtain an over-view of existing feedback tech-nologies and relevant theoretical insights.

Two experimental quantitative field tests to measure the impact of feedback technologies.

Case study to obtain a detailed long-term understanding of a company’s fleet management.

Pre- and post-experiment online surveys to support the findings with quantifiable results and proven scales.

Semi-structured interviews to ac-

10 Introduction

quire a better understanding of drivers’ opinions.

Pre- and post-experiment online surveys to support the findings with qualitative descriptive re-sults.

Desk Research

Desk research enables the researcher to attain an initial understanding of the re-search area by reviewing available secondary literature, including practical and the-oretical streams. The practical aspect focused on obtaining a better understanding of different types of feedback technologies available on the market and the technical capabilities required to support eco-driving concepts. The theoretical aspect drew on various concepts, schools of thought, and studies that highlight the potential impact of feedback technologies on promoting sustainable behavior. The keywords used to search for relevant articles in research databases such as EBSCOhost, Mendeley, and Google Scholar. These keyword searches identified articles in the fields of HCI, energy reduction through feedback technologies and green IS, eco-driving concepts, design of eco-feedback technologies, and sustainability-increasing behavioral change.

Case Study4

Case studies are applied to describe an interpretivistic approach to social research whereby a researcher can, over a period of time, observe, ask questions, and collect the data required to probe issues that are central to the research (Eisenhardt, 1989). Cornford and Smithson (1996) state that “the strength of the case study is in the richness of data that can be obtained” (Cornford and Smithson, 2006, p. 49). King et al. (1994) distinguish between the numbers of variables measured and total observations versus the number of case studies with which the researcher should be involved. These aspects need to reflect the purpose of the study and “the amount of

information a study brings to bear on a theoretical question”

(King et al., 1994, p. 52). Throughout this research the researcher was able to work 4 Course work: Tulusan, J., MP-02 Qualitative Research Methods Lecture, Judge Business School, Cam-bridge University, England, October 2008.

Introduction 11

with a global company with a large corporate fleet serving a total of 19,100 corpo-rate car drivers out of 60,000 employees. The company, which had the internal goal of reducing its overall energy consumption, supported the researcher by granting access to relevant data and providing management support. For this research, a smaller sample size of 450 corporate car drivers from the Swiss region was chosen for conducting the experiments. Several meetings with responsible employees, such as sustainability officers, fleet managers, industry experts, and fleet management software product owners took place. The researcher was also part of the company’s internal sustainability project team. Findings from the thesis were used to support the company’s internal sustainability strategy and are currently implemented as part of a fleet management software solution.

Semi-Structured Interviews5

Semi-structured interviews assist in gaining first-hand information by analyzing the rich responses of interviewees and enable the researcher to compare different view-points (Yin, 2008). It is important to prepare interview questions relevant to the field and test them in advance to ensure that the questions used during the interview yield credible responses (Burton-Jones and Gallivan, 2008). Semi-structured inter-views help to maintain a balanced flow of conversation between questions and an-swers as opposed to imposing a rigid question-and-answer dialogue or being too open (J. Hussey and R. Hussey, 1997). Follow-up semi-structured interviews were conducted with corporate car drivers who participated in the field tests in order to obtain a better firsthand understanding. Besides providing feedback, interviews helped define and reconfirm recommendations and guidelines for instituting organi-zational eco-driving initiatives supported by feedback technologies. These were then reconfirmed through workshops and interviews with relevant company stake-holders.

5 Course work: Tulusan, J., MP-02 Qualitative Research Methods Lecture, Judge Business School, Cam-bridge University, England, October 2008.

12 Introduction

Online Surveys

Surveys enable the collection of opinions using a standardized approach (such as an online questionnaire using proven scales related to environmental behavior) and provide answers that can be evaluated using descriptive or quantitative methods (J. Hussey and R. Hussey, 1997). A pre-survey was conducted with both private and corporate car drivers to get a better understanding of their preferences for feedback technologies that support sustainable driving and the types of feedback offered by these technologies. A post-experiment survey evaluated the opinions of corporate car drivers who had participated in the experimental field tests, including whether direct or indirect feedback was preferred for promotion of sustainable driving. Ad-ditionally, these surveys were used to control for drivers’ attitudes towards envi-ronmental behavior and their technological affinity in order to strengthen the gener-alizability of the findings. For measuring these behavioral attributes, the surveys used constructs and scales that had been applied in the prior literature on energy reduction and behavioral economics.

Experiments

The core research methodology centered on two experiments conducted on corpo-rate car drivers working within the case study company. As the researcher is work-ing for the company, it was possible to run these experiments while guaranteeing the confidentiality of any sensitive data collected during them. The company man-agement made the decision to pilot these experiments within a smaller group of 450 corporate car drivers to establish whether the technological approaches to eco-feedback had an impact on their employees’ overall fuel efficiency. Two experi-ments were conducted between the third quarter of 2011 and the first quarter of 2012. The first opt-in experiment included fifty corporate car drivers, of which twenty-five used a smartphone eco-driving application for eight weeks. In the se-cond opt-out experiment (n = 240), corporate car drivers received a monthly email that detailed each driver’s individual fuel efficiency on a discrete or continuous feedback scale, illustrating how ‘good’ or ‘bad’ their fuel efficiency was in compar-ison to their colleagues (social comparison). The second experiment ran for three months and enhanced the researcher’s understanding of the extent to which email feedback on fuel consumption influenced drivers’ fuel efficiency.

Introduction 13

I.3.2 Relationship to Research Questions

A summary of the overall research process is illustrated in Figure 2, helping to re-flect the relationships between the research questions, chosen methodologies, and findings. Findings obtained through this mixed-methods research approach made it possible to address more than one sub-research question using the same data set; they have been highlighted in the relevant chapters.

Figure 2: Research Process and Relation of Research Questions to Methods

14 Introduction


Initial research focused on secondary data collection/ desk research to acquire a thorough understanding of feedback technologies currently on the market, as well as obtaining a theoretical grounding in this topic area. This information was used to formulate two online surveys to investigate what drivers expected from feedback devices and which feedback types they preferred or found most relevant to them. The first survey was answered by private car drivers (n = 139), the second by corpo-rate car drivers (n = 131). This allowed for comparing preferences across the differ-ent user groups.


The preliminary research into SQ1 assisted in designing the first empirical study. This experiment probed how an eco-driving smartphone application influences the fuel efficiency of corporate car drivers. Fifty drivers participated in the opt-in exper-iment, of which twenty-five used the smartphone application for eight weeks, com-pleted an online post-experiment survey, and participated in semi-structured inter-views. The quantitative evaluation of driver’s fuel efficiency together with the qual-itative approach of the survey and interviews enabled exploring how the use of an eco-driving smartphone application changes over time. These findings support the argument that an eco-driving smartphone application can have a positive impact on the fuel efficiency of corporate car drivers who have not taken part in eco-driving training.


The Feedback Intervention Theory of Kluger and DeNisi (1996) proposed that vary-ing the features of the feedback delivered could manipulate the extent to which in-dividuals would change their behavior. As highlighted by Kluger and DeNisi (1996) “a device is required that gives the driver immediate and accurate fuel consumption

Introduction 15

information, yet is not a distraction from safe driving.” Meschtscherjakov et al. (2009) highlighted that the distraction of such a technology can induce inflated risk levels during driving. Immediate and context-related feedback is integral to promot-ing drivers’ fuel efficiency as they drive. Additionally, the ability to control the fre-quency of the feedback received has been proven to motivate users to change their behavior. This is known as “feedback control” and allows individuals to control the amount and timing of feedback with which they are provided (Meschtscherjakov et al., 2009). The eco-driving smartphone application allowed corporate car drivers to choose between direct or indirect feedback types and recorded how long each feed-back type was used during each journey. These data were used to statistically evalu-ate which feedback type was most popular; this was validated in the semi-structured interviews.

SQ4: How does social-normative feedback about driver’s fuel consumption via email influence corporate car drivers’ average fuel efficiency?

Receiving feedback from a smartphone application depends on certain prerequisites, such as owning a smartphone and activating the application during driving. These hurdles can reduce the acceptance of such feedback technologies. An alternative solution that is available to every employee in a corporation is providing feedback via email, known as indirect/ offline feedback. This type of feedback is easily ac-cessed by a large number of company drivers, as most of them have access to an email account. Corporate car drivers have limited knowledge of their monthly fuel consumption unless noting it autonomously, but monthly emails make this infor-mation readily available to them and assist with improving their awareness of fuel consumption and how to improve their overall fuel efficiency. Hence, the second experiment provided greater insight into how drivers’ fuel efficiency changes as a result of receiving an email detailing their fuel consumption and comparing this to that of their colleagues (social comparison). The data needed to calculate the fuel efficiency is automatically collected from the company’s petrol credit cards during the process of paying for each tank filling.

16 Introduction


Findings from the previous research questions must be considered within the correct context to guarantee the successful application of eco-feedback technologies in an organizational setting. Management must understand which factors, intrinsic or ex-trinsic, motivate their employees(Beswick, 2007). Extrinsic motivation refers to the use of external rewards (e.g., financial incentives or extra days off) and/ or recogni-tion (e.g., participation in important meetings) to stimulate change (Zahorsky, 2010). Intrinsic motivation stems from within the person; satisfaction thus derives from completing the task itself, thereby improving the driver’s self-esteem (Zahorsky, 2010). Because external motivations such as monetary incentives to im-prove fuel efficiency are irrelevant for this sample group, greater understanding is needed to establish how to motivate corporate car drivers. Their intrinsic motivation must be drawn upon so that they feel that sustainable driving is partly their respon-sibility. How this can be achieved will be explored and defined with clear recom-mendations in this chapter; this is done using information collected via the semi-structured interviews, alignment meetings, and the online survey.

I.4 Scope of the Thesis

The goal of this doctoral thesis is to attain a better understanding of how eco-feedback technologies can be used to educate corporate car drivers about their fuel efficiency. Data on tank refills and the number of kilometers driven monthly were evaluated to determine the impact of the eco-friendly smartphone application or monthly email on drivers’ fuel efficiency. These findings and the resulting recom-mendations should help corporations ascertain whether or not emissions reduction concepts need to be supported by these feedback technologies to maximize their impact. If so, corporations will be able to reduce their overall CO2 emissions and fuel costs with minimal investment.

Introduction 17

490.0 kTons

I.4.1 Project Context6

These research activities were part of the Bits to Energy Lab (B2E), a collaboration between the University of St. Gallen (HSG), Swiss Federal Institute of Technology (ETH Zurich), and SAP Research. The B2E Lab took the lead on the academic re-search activities, receiving some management support from the case study partner, and the researcher had access to the primary data generated during the field tests. This was possible because the researcher is working in cooperation with both organ-izations, the B2E Lab and SAP Research.

The case-study company has 60,000 employees and a total of 19,100 corporate car drivers (Q4, 2011). The number increased by 10% in 2011 and is predicted to rise again in 2012. The total CO2 emissions in 2011 were 490 kTons, of which 23% were attributed to corporate car drivers and 38% to business flights (see Figure 3). Electricity accounted for 10%, and commuting and data centers from the case-study company were each responsible for 8%.

Figure 3: 2011 SAP AG’s global distribution of CO2 Emissions7

The high CO2 emissions and energy consumption of the 19,100 corporate cars indi-cate an urgent need to determine effective and sustainable mobility solutions. The first approach was to reduce the guideline CO2 emissions cap for new corporate cars

6 Published paper: Tulusan, J., Steggers, H., Staake, T., Fleisch, E., Supporting eco-driving with eco-feedback technologies: Recommendations targeted at improving corporate car drivers’ intrinsic motivation to drive more sustainable, Energy Informatics 2012 (EI 2012), Atlanta, Georgia, United States, October 2012. 7 www.sustainabilityreport.com

18 Introduction

from 220 gCO2/ km down to 190 gCO2/ km. However, this guideline is only appli-cable to newly purchased cars; as corporate cars are driven for up to five years or 150,000 kilometers, implementation is gradual over time. A more immediate and cost-efficient approach is thus required to modify driving behavior. Past initiatives have focused on classroom training strategies such as eco-driving training, which require drivers to invest time and the company to invest money.

I.4.2 Financial Saving Potential

Based on the potential fuel saving figures for corporate car drivers from the German Federal Environmental Agency (2009), the case study company could save approx-imately four million Euros per year if their total fuel efficiency improved by one liter per 100 km. This would represent a savings of nearly 10% of the total 2011 petrol expenditure of 42.5 million Euros. With a moderate average travel distance of 15,000 kilometers per year, the savings would be 1,250 Euros per corporate car8. Since corporate car drivers drive an average of 35,000 kilometers per year, the sav-ings could be much higher; nevertheless, a conservative estimate of potential fuel savings was chosen.

I.4.3 Example of Data Set

The dataset used for calculating the fuel efficiency of each corporate car driver, an example of which is shown in Table 2, includes each corporate car’s ID, the driver’s employee ID, monthly total kilometers driven, monthly total liters filled, total num-ber of kilometers driven, and how long the driver has owned the vehicle (in months).

8 Calculation is based on 19,100 corporate cars with a lifespan of six years and a constant petrol price of 1.40

Euro per liter.

Introduction 19

Table 2: SAP Corporate Car Drivers’ Data Set Example

Car ID

Employee ID

Monthly total km

Monthly total li-

ters filled

Monthly fuel effi-ciency in l/ 100km

Total km

Total ownership in months

1 1 3398 260 7,7 26290 11

2 2 2132 195 9,1 19545 7

3 3 1433 121 8,4 24245 10

4 5 2368 146 6,2 33390 13

These data were automatically collected from the drivers’ petrol credit cards during the process of paying for tank refilling and made available to the researcher. It was then possible to verify monthly changes in the overall fuel efficiency of each driver participating in the field tests. The fuel efficiency is calculated from these data by dividing the total amount of liters filled per month by the total distance driven per month and multiplying by 100 (see Figure 4).

100

kilometers

litersFE

Figure 4: Fuel Efficiency Calculation

Further data were collected via the online pre- and post-experiment surveys using open-ended questions and likert scales ranging from most to least preferable. To extend the understanding gleaned from these purely numerical values, semi-structured interviews were conducted to gain deeper insights into why users pre-ferred certain feedback technologies over others. This enabled the researcher to gain a better understanding of why eco-feedback technologies influence the driving be-havior of corporate car drivers.

I.4.4 Research Context

Focusing on the technical rationale of this topic would provide a limited under-standing of why drivers adapted their driving after receiving feedback from eco-feedback technologies. Difficulties would have arisen when applying these findings to a practical business setting. As such, this research builds on two research streams

20 Introduction

to analyze changes in fuel efficiency from both social/ behavioral and technological perspectives (see Figure 5). This is defined in the literature as the socio-technical perspective (Bostrom and Heinen, 1977); emphasis is placed on evaluating the technical and social artifact (the corporate car driver) to gather a better understand-ing of why new technologies have the potential to change organizational processes (Keen, 1981).

Figure 5: Research Perspectives

The technological perspective builds on the literature in the field of feedback infor-mation systems, specifically persuasive mobile eco-feedback technologies. Fogg (2002) classified persuasive technologies as “any interactive computing systems

designed to change people’s attitudes or behaviors.” These technologies can bridge the gap between one’s lack of environmental awareness and everyday behavior;; this has been termed the ‘environmental literacy gap’ by Froehlich et al. (2010). Pub-lished research has claimed statistically significant reductions in household energy use (Geller et al., 1982; Graml et al., 2011; Loock, Graml, et al., 2011). Interaction and usage behavior have also been explored within the Human Computer Interac-tion (HCI) research domain. Authors such as Froehlich et al. (2010) and Lee et al. (2010) are particularly interested in how humans interact with these devices (e.g., how corporate car drivers use the eco-feedback technologies over time) and how the technology in these devices should be developed, designed, and applied.

The behavioral perspective, with its origin in social-psychological concepts (Cialdini et al., 1991), assesses the effectiveness of these devices in stimulating be-havioral changes, such as changing fuel efficiency by evoking more ecological driv-ing behaviors. Past research has evaluated why people speed and how to encourage seat belt use to improve safety (Geller et al., 1982). Determining how (whether via numbers, symbols, or scales) and when (direct/ real-time vs. indirect/ after driving) eco-feedback should be presented is important for reducing a driver’s cognitive

Introduction 21

workload and level of distraction when using a feedback device (Donmez, 2007). The motivations for employees within an organization to adapt their behavior on work-related tasks has been appraised using the concepts of extrinsic and intrinsic motivational factors (De Young, 1985; Zahorsky, 2010).

The business value perspective builds on the social/ behavioral technological find-ings to suggest recommendations for industry stakeholders and present the first practical design for supporting eco-driving via feedback technologies. These rec-ommendations and concepts can be used by organizations to promote energy reduc-tions or to define new CO2 emissions reduction policies within an organization. As an example that directly reflects this research, relevant findings are currently being implemented as part of the overall CO2 reduction strategy of the case study compa-ny. Specifically, the company is offering the eco-driving smartphone application to all interested drivers and is providing information to all drivers via a monthly feed-back email. The latter compares the details of an individual driver’s fuel efficiency and CO2 emissions to that of their colleagues and offers eco-driving tips. These ser-vices will also be added to a fleet management software solution and could be of-fered to other corporations. Further details are outlined in Chapter VI.

I.4.5 Target Audience

The thesis is targeted at both practitioners and members of the academic community who want to gain a better understanding of how feedback technologies enable cor-porate car drivers to improve their overall fuel efficiency. For practitioners (i.e. sus-tainability managers in organizations, corporate fleet managers, or consultants) the-se findings could extend and support a sustainability strategy aimed at reducing their company’s overall energy consumption. Those who develop products such as smartphone applications or fleet management software solutions also have a vested interest in how feedback technologies should be enhanced to improve services to their clients. Findings are particularly relevant for corporations with a large corpo-rate car fleet.

Academics in the fields of green IS, HCI, and behavioral economics will find that the outcomes of this research contribute to the technical research domain and en-hance knowledge of eco-feedback technology usage and user interaction. The be-

22 Introduction

havioral economics research focus aims to understand why corporate car drivers’ behavior changes even if monetary incentives are not relevant. Given the present importance of energy reduction, this research will be of strong interest to the re-search community.

I.5 Structure of the Thesis

Figure 6: Structure of the Thesis

The last section of this chapter will briefly explain the content of each chapter (see Figure 6).

The introduction outlines the practical and theoretical relevance of the thesis, after demonstrating the overall significance of the topic. The research questions, sub-questions, and mixed-methods research methodology are outlined, indicating the direction of the ensuing thesis.

The second chapter reviews ‘state of the art’ concepts, the theoretical foundations of eco-driving and eco-feedback technologies and discuss studies that have indicated how these technologies enable drivers to adopt an eco-friendly driving style.

Introduction 23

Chapter third, a preliminary study, evaluates drivers’ preferences for feedback tech-nologies based on an explorative investigation, indicating which feedback types are preferable for influencing the driving behavior of both private and corporate car drivers.

Chapter fourth explains the eco-driving smartphone application experiment, in which corporate car drivers used an eco-driving application for eight weeks. Statis-tical analysis is used to analyze usage patterns of the eco-driving application over time and which of the feedback meters shown on the application were preferred.

Chapter fifth describes the second experiment, which evaluated the difference in fuel efficiency between groups of drivers’ using two different feedback scales. This information was delivered to the participants via monthly emails that highlighted their fuel consumption figures. This field test involved 240 corporate car drivers.

Chapter sixth outlines insights and recommendations for increasing intrinsic moti-vations of drivers to improve their fuel efficiency. The five basic arguments from the FIT from Kluger and DeNisi (1996) have been applied for a structured analysis. Guidelines were given on the findings of the post-survey and validated through semi-structured interviews.

Chapter seven provides an outlook of a prototype fleet management solution, data structure and the system´s architecture by illustrating current mock-ups of the solu-tion. Drawbacks are also highlighted.

Finally, the last chapter highlights key findings and summarizes the theoretical, practical, and technological implications. Furthermore, provides recommendations and the first conceptual model of how to apply eco-driving programs in organiza-tions with the support of eco-feedback technologies. Limitations and future research are also considered and an outlook of a prototype fleet management solution illus-trated.

Eco-Driving and Eco-Feedback Technologies Related Work 25

II Eco-driving and Eco-Feedback Technologies Related Work

This chapter provides an overview of eco-driving and related techniques that can improve drivers’ fuel efficiency. The potential fuel savings achievable through eco-driving are also illustrated with examples from previous research. The difference between private drivers and the user group focused on in this research, corporate car drivers, is also discussed.

II.1 Overview of Eco-driving9

Different driving styles result in fluctuations in fuel efficiency; for example, aggres-sive drivers consume 30% more fuel than conservative drivers (Romm and Frank, 2006). Aggressive drivers who express bodily or psychological aggression towards drivers, passengers, or pedestrians also demonstrate risky behavior (Tasca, 2000). Risky driving behavior is chiefly characterized by speeding, changing lanes fre-quently, and maneuvering without signaling (Dula and Geller, 2003). In contrast, economic driving behavior, referred to as hypermiling, has the ultimate goal of con-suming the least possible volume of fuel, even if this means compromising on safe-ty to a degree (Chapnick, 2007). Another behavior is eco-driving; this does not compromise on road safety and can lead to average reductions in fuel consumption of up to 15%, lower greenhouse gas emissions, improved road safety, and reduced accident rates (GreenRoad, 2008). Eco-driving depends on individuals’ behavior;; its principles thus aim to alter and discourage driving practices that do not conform to eco-driving techniques.

EcoDrive (2012) defines several practices that can alter driving behavior:

Anticipating traffic flow and looking ahead in order to be able to react quick-ly and avoid sudden starts or stops.

9 Published paper: Tulusan, J., Soi, L., Paefgen, J., Staake, T., Eco-efficient feedback technologies: Which eco-feedback types prefer drivers most?. World of Wireless, Mobile and Multimedia Networks IEEE Confer-ence, Lucca, Italy, June 2011.

26 Eco-Driving and Eco-Feedback Technologies Related Work

Shifting into a higher gear as soon as possible once the threshold of 2500 or 2000 revolutions per minute (rpm) for petrol and diesel cars, respectively, is reached.

Using the highest gear possible at a low RPM and maintaining a steady speed.

Smoothing deceleration by releasing the accelerator in time and leaving the car in gear.

Using fuel-saving devices onboard vehicles, such as cruise control or a navi-gation system’s ecoRoute setting.

Considering switching off instruments that consume extra energy, such as air-conditioning.

Checking tire pressure at least once a month.

Advantages and disadvantages of such eco-driving are listed in Table 3.

Table 3: Advantages and Disadvantages of Eco-Driving

Advantages Disadvantages

Vehicle type unimportant Driver has to be willing to change driving behavior

Easy to learn Eco-driving training re-quired

Raises awareness of eco-logical driving practices

Few training possibilities available

Low effort (one day train-ing)

Might not have a long-term effect without further interventions

Low investment (from 100 Euros)

Loss of time due to long-er journey time

Advantages of eco-driving are its independence from the vehicle type and the ease of learning the driving techniques with low financial and time investment (a one day training starts from 100 Euros). Among the studies investigating the efficiency of eco-driving, however, one demonstrated that the majority of drivers who improved their fuel economy were already motivated to do so before training; this motivation appeared to have an substantial impact on reducing fuel consumption (Johansson et


al., 2003). During and shortly after training, drivers achieved maximum improve-ments in fuel efficiency of up to 15%; these gains declined in the long term if ongo-ing prompts were not given (EcoDrive, 2012; Onoda, 2009). This gradual deteriora-tion in maintaining fuel efficiency indicates the necessity for substantial ongoing measures to modify activities that are performed automatically. Potential interven-tions include ongoing feedback, regulatory actions, economic incentives, and feed-back technologies that inform drivers after they have participated in the classroom training.

II.1.1 Eco-driving Training Practices10

Campaigns, such as the pan-European ECODRIVEN campaign that ran from 2006 to 2008, are one approach to introducing eco-driving. That campaign helped raise awareness of eco-driving by offering lessons during regular one-day classroom training courses that included theoretical and practical sessions (EcoDrive, 2012). In the theoretical classroom session, eco-driving guidelines and recommendations were taught. In the practical session, drivers were able to use a driving simulator and/ or their own vehicle in a designated driving area to practice the eco-driving techniques. Even though further classroom education is welcomed, it is insufficient to effectively alter driving habits that have become entrenched in individuals through years of practice, as drivers tend to return to previous driving habits soon after training is complete. Therefore, classroom training should be supported by internet-based training material and/ or reoccurring practical training sessions that explain the eco-driving principles again or remind participants of the most relevant and up-to-date driving practices. Ideally, changes in driving style that promote eco-driving behavior should be maintained on a long-term basis.



II.1.2 Eco-driving Fuel Saving Potentials11

The transport sector report produced by the International Energy Agency (2009) analyzed the short- and medium-term impact of eco-driving initiatives. Immediately after eco-driving training, average fuel efficiency improved by up to 15 % (Onoda, 2009). In the long-term (i.e. more than one year later), an improvement of approxi-mately 5% was maintained without further feedback and an improvement of 10% was achieved with continuous post-training feedback (i.e. tips and driving style analysis) (EcoDrive, 2012). A five-percent improvement in fuel efficiency can be sustained by using in-car eco-driving equipment or an onboard computer that pro-vides feedback on fuel consumption, correct gear changes, or cruise-control func-tions. Corporate studies conducted by EcoDrive (2012) with 350 service car drivers from the Canon company showed an overall reduction in fuel consumption of 6.1%, with 35% fewer car accidents. Corporate car drivers working for the German com-pany Hamburger Wasserwerke achieved a 6% decline in fuel consumption and a 25% decrease in the number of vehicular collisions. The Austrian bus company NIGGBus was able to improve its fuel efficiency by 5% in 2000 and 7% in 2001. These figures show that eco-driving training can have a positive impact on the driv-ing behavior of corporate car drivers. The following sub-chapter introduces the dif-ferences between two driver groups: private and corporate car drivers.

II.2 Differences between Private and Corporate Car Drivers12

Corporate car drivers are defined as those who drive a corporate car, largely for work-related travel but also for private reasons. The nature of their occupations re-quires daily travel to customer meetings (i.e. to pitch sales or provide consulting services) and is integral to the company’s generation of leads and income.

11 Published paper: Tulusan, J., Staake, T., Fleisch, E., Direct or indirect sensor enabled eco-driving feed-back: Which preference do corporate car drivers have?, Internet of Things 2012 – Third International Confer-ence on the Internet of Things (IoT 2012), Wuxi, P.R. China, October 2012. 12 Published paper: Tulusan, J., Steggers, H., Staake, T., Fleisch, E., Supporting eco-driving with eco-feedback technologies: Recommendations targeted at improving corporate car drivers’ intrinsic motivation to drive more sustainable, Energy Informatics 2012 (EI 2012), Atlanta, Georgia, United States, October 2012.


Table 4: Differences between Private and Corporate Car Drivers

Private Car Drivers Corporate Car Drivers Travel Reasons Private Work and/ or private Average travel distance (in Europe) 13,500 km/ year 35,000 km/ year

Payment of petrol costs Self-payment Paid by the company Payment of car mainte-nance Self-payment Paid by the company

Incentives to drive ecologically Save petrol costs No incentives for the

driver Passenger car registra-tions in 2012 (Germany) 2,250,000 750,000

Passenger car registra-tions in 2011 (Switzer-land)

213,000 106,000

In Europe, corporate car drivers drive an average of 35,000 kilometers per year (see Table 4). In comparison, private car drivers drive an average of less than half this distance, 13,500 km per year (DTLR, 2012). Private car drivers must pay their own petrol and car maintenance costs, whereas corporate car drivers pay a monthly fee (a leasing fee or a fee for the company) for their car of between 300 and 600 Euros. This fee covers petrol costs for business-related mileage and maintenance and ser-vice costs, depending on the car type and country. The car’s CO2 emissions have an impact on the size of the monthly fee; the higher the emissions, the greater the fee. In some countries, such as Germany, drivers have to pay an additional monthly tax to the government of 1% of the total purchase price of the corporate car. Petrol is paid through corporate credit cards, and the total number of kilometers driven must be typed into the credit card terminal during the payment process at the petrol sta-tion. The petrol companies provide the company with the tank filling data each month, but the corporate car drivers do not receive a monthly overview of these costs or of the volume of petrol consumed. Hence, drivers are unaware of these monthly costs unless they maintain a personal record.

For private car drivers, incentives to drive more ecologically include direct savings in fuel costs, whereas corporate car drivers have their fuel and vehicle maintenance costs paid by the company. As such, economic and eco-friendly driving concepts do not directly lead to financial rewards for these individuals. An improvement in their


fuel efficiency would, however, result in financial savings for their employing com-pany.

The total number of corporate cars in Germany indicates that a quarter (750,000 of 3,000,000) of cars newly registered in 2012 were corporate cars (Dataforce, 2011). For Switzerland in 2011, the fraction was one third (106,000 of 319,000) (Bahnmüller, 2012; Ullmenstein, 2012). Bearing this in mind, it is important to un-derstand how corporate car drivers can be motivated to adopt more sustainable driv-ing behavior.

II.3 Eco-Feedback Technologies

II.3.1 Eco-Feedback Technologies in Cars13

GreenRoad (2008, p.12) suggested “a device is required that gives the driver im-mediate and accurate fuel consumption information, yet is not a distraction from safe driving”. A limited number of in-car eco-feedback devices have been pio-neered for vehicles that use petrol. In spite of ongoing advancements in technology, a large percentage of feedback devices are utilized for hybrid or electrical vehicles, as the most effective concept of ecological driving behavior extends across this range of vehicle types (Romm and Frank, 2006).

Eco-feedback technologies in cars aim to generate the driver’s awareness of their driving style and consumption of petrol (Formosa, 2009). The driver needs to re-ceive information visually through a feedback technology, whereas the technology needs to know what the driver is doing. This interplay (see Figure 7) of the infor-mation received and the response of the driver could influence the driving style (Formosa, 2009).

13 Published paper: Tulusan, J., Brogle, M., Staake, T., Fleisch, E., Becoming a Sustainable Driver: The Impact of Mobile Feedback Devices. 4th ERCIM eMobility WWIC Conference, Lulea, Sweden, June 2010


Figure 7: Feedback Technologies and Response of Driver14

Car manufacturers focus on enhancements of on-board communication instruments, which support the driver during the journey. These functions focus more on telematic solutions or car2car communications, such as calculating the correct dis-tance to a car in front and automatically slowing down the car to keep a safe dis-tance. Ford, with his SmartGauge concept, worked with the service design company IDEO to define a concept for an on-board instrument to make the driver aware of eco-efficient driving. The notion that design can influence behavior underlines the importance of developing systems that are relevant to the driver (Formosa, 2010). The SmartGauge concept connects the instrument with the controller area network (CAN) bus of the car to store relevant data, such as average speed, acceleration, braking, liters used, etc., from the car. From this information the system is able to understand the driving behavior and displays real-time feedback by displaying leaves of the fuel usage (see Figure 8 right side).

Figure 8: Ford’s SmartGauge in-car Interface14

14 Formosa, D. (2009), “Ford SmartGauge: Designing an Extra 9 MPG?,” Smart Design. Retrieved from

www.smartdesignworldwide.com/pdf/SmartDesign_SmartGauge.pdf


II.3.2 Mobile Eco-Feedback Technologies15

Through the advancements of smartphone technologies by companies, such as Ap-ple, Google, HTC, and Samsung, the latest eco-feedback technologies are eco-driving applications that run on mobile devices. Eco-driving applications use the GPS sensor and the accelerometer technologies of the smartphone and empower the driver by analyzing their driving style to provide recommendations about how they can drive in a more eco-friendly manner.

In 2009, only one application was available on the market (see Figure 9). In 2012, 29 offer the iOS and/ or Android operating systems. Since it was not possible to compare the iOS ratings with the Android application and user ratings were only in relation to the latest version of the application, it was not feasible to provide a valid ranking for the least and most favored applications. The overall trend indicates that more and more applications are available on the market, but the frequency use dur-ing driving and the impact that these applications have on driving behavior remain unclear.

Figure 9: Eco-driving Applications Market Share 15 Published paper: Tulusan, J., Brogle, M., Staake, T., Fleisch, E., Becoming a Sustainable Driver: The Impact of Mobile Feedback Devices. 4th ERCIM eMobility WWIC Conference, Lulea, Sweden, June 2010


One example of a current application is the Bliss Trek (see Figure 10) eco-driving application, which tries to persuade the driver to modify their driving in a playful way by using speed and acceleration data. Real-time feedback with respect to eco-logical driving behavior is illustrated using a green meter represented by leaves and a driving score. The goal is to reach 1000 points to reach the next level. When the driving score deteriorates clouds appear on the screen and the sky darkens (Stoppani, 2009).

Figure 10: Bliss Trek Application Interface16

Another example is the greeenMeter application (see Figure 11), which is more straightforward and displays the average speed in an eco-driving efficiency gauge. Depending on the average speed and roads driven the software displays the area for eco-efficient driving. Besides this gauge, further feedback meters that show the CO2 emissions, fuel efficiency, and fuel costs are able to be selected.

16 Stoppani, S. (2009), “Bliss TrekTM,” Overview of the BlissTrek Application. Retrieved August 5, 2012,

from http://blisstrek.com/


Figure 11: greenMeter Application Interface17

This interface changes colors according to your driving behavior; it turns dark blue when you consume more fuel and bright green when you drive more efficiently.

Figure 12: Color changing Feedback18

While you drive efficiently the silver ball remains into the green area. If you start consuming more fuel or accelerate aggressively, the ball moves to the yellow and

17 Hunter, C. (2009), “greenMeter: iPhone/iPod Eco-driving App.” Retrieved September 5, 2012, from

http://hunter.pairsite.com/greenmeter/ 18 Source: The Game (Changer), motive, 2009


red areas. In addition to this you get a driving score, marking your driving behavior at each trip.

Figure 13: Green Driving App19

A selection of further applications currently on the market that have similar eco-driving functions are shown in Table 5:

Table 5: Overview of Eco-driving Applications

Logo App. Name

Com-pany

Release Date

Target Market

Oper. System

goDriveGreen

ProLogic, Inc.

Jan. 12th, 2011 USA iOS

EcoDrive

Appli-cations Co. Ltd.

May 19th, 2012

China, Japan

iOS/ Android

Eco Bccard

Ltd. Sep. 4th, 2012 Korea Android

DriveGain

DriveGain Ltd.

Jul. 20th, 2010

UK, Nether-lands

iOS

Tokyo Smart Driver

maysun soft Co. Ltd.

Oct. 27th, 2011 Japan iOS

19Source: GreenGasSaver, The Apple Seed Story, 2009


II.4 Smartphones Market Share

Looking at the smartphone market share penetration (see Figure 14), Frost & Sulli-van (2009) estimated a market share of smartphones of the total mobile phones by 2014 of 37% and Gartner (2010) of 40%. In 2008, the smartphone market share was only 12% (Frost & Sullivan, 2009) or 11% (Gartner, 2010). The total mobile phones’ sales rate in 2014 is predicted to be 1,516 million (Frost & Sullivan, 2009) or 1,908 million (Gartner, 2010) units. These estimates support the importance eco-driving smartphones application could have in the upcoming years.

Total mobile phones

Smart phones (F&S)20

Market Share (F&S)

Total mobile phones

Smart phones (Gart.)21

Market Share (Gart.)

2008 1178 139 12% 1222 139 11%

2009 1223 190 16% 1211 172 14%

2010 1267 251 20% 1400 260 19%

2011 1313 321 24% 1500 374 25%

2012 1375 398 29% 1600 484 30%

2013 1451 478 33% 1750 621 35%

2014 1516 555 37% 1908 770 40%

20 Frost & Sullivan. (2009), 2010 Outlook & Forecast: Mobile & Wireless Communications. 21 Gartner. (2010), Forecast: Mobile Devices, Worldwide 2003-2014, 2Q10 Update


Figure 14: Smartphone Market Share Outlook

II.5 Impact of Eco-Feedback Technologies on the Driving Behavior

The absence of a significant body of literature in the field of mobile eco-efficient feedback technologies does not allow for one to confidently speculate the potential success of similar systems. At the discretion of one’s knowledge, currently no re-search exists, besides studies in which an eco-driving navigation or on board system have been used that directly measure the influence Green IS or smartphone technol-ogy have on driving behavior. However, feedback technologies “seems to be a

promising approach to be researched in automotive environments, when it comes to fostering a more fuel-efficient driving behavior” (Meschtscherjakov et al., 2009, p.82). Froehlich et al. (2009) operated this type of mobile technology when record-ing the daily transportation behavior of participants, i.e. if the person used the bus or bike, which presented a positive outcome for a small sample group (n=13) and field test duration of only three weeks.

Monetary, environmental, and technological factors were found to be prominent in inspiring drivers to modify their driving style, however, caution must be taken when

12% 16%

20%

24% 29%

33% 37%

11% 14%

19%

25% 30%

35% 40%

0% 5% 10% 15% 20% 25% 30% 35% 40% 45%

0 100 200 300 400 500 600 700 800 900

2008 2009 2010 2011 2012 2013 2014

Smar

tpho

ne m

arke

t sha

re

Tota

l sm

artp

hone

s (m

illio

ns)

Total smartphones (Frost&Sullivan) Total smartphones (Gartner) Smartphone market share (Frost&Sullivan) Smartphone market share (Gartner)


generalizing these finding as many of the studies analyzed how to influence drivers to use other modes of transport or were limited to a laboratory setting. Foxx and Hake (1977) acquired a 20% reduction in students’ mileage to campus by offering cash bonuses as a consequence of encouraging them to use their cars less and travel on alternative modes of transport, i.e. carpooling, bike or bus, for an experimental duration of one month. Deslauriers and Everett (1977) persuaded people to use the bus by paying passengers $0.10 for each journey; however, this was only successful for the period of time when reimbursement was given. Van der Voort et al. (2001) conducted a laboratory study dedicated to improving drivers’ fuel efficiency. Re-sults showed a reduction of 7% in fuel consumption when participants tested a fuel-efficiency support tool in a driving simulator. The support tool provided drivers with eco-driving information in response to their driving style and driving condi-tions. However, this device has yet to be trialed in the field.

Studies by Boriboonsomsin et al. (2010), Lee et al. (2010), Meschtscherjakov et al. (2009), Pace et al. (2007) and Tester et al. (2000), among others, have evaluated eco-driving on-board feedback systems. An improvement in fuel efficiency by 1% on highways and 6% on city streets was shown in a field test with 20 drivers in Boriboonsomsin et al. (2010) study. Participants participated in an eco-driving training program and had access to a vehicle with an on-board feedback device, which provided them with instantaneous fuel consumption information (Boriboonsomsin et al., 2010). Siero et al. (1989) applied feedback interventions to urge postal-workers in the Dutch Postal Service to be more eco-friendly when driv-ing by providing them with instant feedback via an installed tachometer in their de-livery trucks and agreeing on a goal of reducing fuel consumption by 5%. Results revealed a 7.3% reduction which was also maintained six months thereafter (Siero et al., 1989). Situational and personality variables such as self-esteem, altruism, en-vironmental believes, anxiety, attitude towards driving, and positive or negative expectation of achievement are additional variables that could have had an impact on behavioral change. For this reason, one cannot claim which feedback proposed had the strongest impact on reducing fuel consumption. Siero et al. (1989, p. 417) recognized this short-coming in their research and suggested that “…it would be

desirable to split up such an experiment into a number of sub experiments to exam-ine the effect of each individual element”.


II.6 Feedback Intervention Theory

Although there is evidence to suggest that it is possible to reduce energy consump-tion, “there does not appear to be substantial research on how specific behavioral changes lead to measurable reductions in GHG emissions” (Doppelt et al., 2009, p. 7). Many studies describe social psychological and structural factors when deter-mining which transport mode to take, but struggle to make the link between a spe-cific behavioral change and one’s attitude. Difficulty persists when attempts are made to relate behavioral modification with feedback interventions; quite often the results have been contradictory and feedback has not necessarily improved perfor-mance (Balcazar et al., 1986). Further research is necessary to get a better under-standing of the relationship between behavioral modification and feedback interven-tions.

Van Velsor et al. (1997, p.36) define feedback as “information about a person’s

performance or behavior, or the impact of performance or behavior, that is inten-tionally delivered to that person in order to facilitate change or improvement”. Kluger and DeNisi (1996) applied the Feedback Intervention Theory (FIT) to ap-praise how feedback interventions provoked driving motivation and behavior, de-termined by comparing feedback to goals, standards, or norms. They defined feed-back interventions (FIs) as “actions taken by (an) external agent(s) to provide in-formation regarding some aspect(s) of one’s task performance” (Kluger and DeNisi, 1996, p.255). Henceforth, FIs are applied to influence individual’s task per-formance and motivate them to change their actual behavior; this is referred to as ‘knowledge of results'. Interventions may encompass a broad spectrum of tasks that are elicited by intent from an external agent. Kluger and DeNisi (1996) proclaimed five basic arguments (see Table 6) for the FIT:

Table 6: Five Arguments Feedback Intervention Theory22

Behavior is regulated by comparisons made from feedback obtained by goals or standards.

Goals and standards are organized in a hierarchy.

22 Kluger, A.N. and DeNisi, A. (1996a), “The effects of feedback interventions on performance: A historical

review, a meta-analysis, and a preliminary feedback intervention theory.,” Psychological bulletin, Vol. 119 No. 2, p. 254.


Attention is limited therefore, only feedback-standard gaps that receive attention actively participate in behavior regulation.

Attention is limited, and therefore only feedback-standard gaps that re-ceive attention actively contribute to behavior regulation.

FIs change the focus of attention and consequently, have an impact on behavior.

Individuals use feedback to verify their performance in relation to their goals or standards (Froehlich et al., 2009). Goals and standards are organized in a hierarchy; the difference between the feedback obtained by the FI and the goal or standard set is defined as the ‘feedback-standard gap’. The feedback-standard gap allows indi-vidual’s to reflect on their actual (positive or negative) performance with regards to the specified goal or standard (Kluger and DeNisi, 1996). Individuals may respond to the feedback-standard gap in one of four ways: behavioral change, rejecting the feedback, amending the goal/ standard set, or rejecting the goal/ standard set alto-gether. Standards can include multiple facets, such as norms, prior or past perfor-mance levels, and performance of other groups. Hence, multiple standards can in-fluence the performance of one’s response to the FIs in variable ways;; for this rea-son, it is not as straightforward as claiming that behavioral change is solely a result of the goals or standards appointed.

II.6.1 FIT and Computer Aided Feedback

Alder (2007) applied the FIT by Kluger & DeNisi (1996) to appraise how feed-back, provided through a computer performance monitoring (CPM) IS, influences the performance of employees within the company. Alder (2007) argued that com-puter mediated feedback allows one to focus on the tasks and results in improved performance. Constructive feedback given with regards to performance was only effective in promoting behavioral modification when the CPM was not perceived as a control device by employees. Feedback from the CPM device in conjunction with face-to-face feedback delivered the most benefits (2007). The value of CPM to or-ganizations is undeniable as it produces performance data at an efficient and fre-quent rate, regularly drawing using awareness to their personal performance figures and enabling them to modify their behavior when necessary (Grant and Higgins, 1989). Electronic monitoring provides employers and employees following benefits:


it is objective (Worsnop, 1993), it may help employees to avoid stress associated with evaluation uncertainty (Alder and Tompkins, 1997), and it improves perfor-mance feedback and evaluation (Angel, 1998; Henriques, 1996). On the downside, CPM may be seen as an invasion of employees’ privacy, contribute to decrease job satisfaction and increase stress (Greengard, 1996). CPM has to be seen as one com-ponent of a complete performance management system, which is influenced by oth-er elements that are also part of the process (Alder, 2007). In this research the as-sumption is made that “FIT predicts that computer-mediated feedback will lead to a focus on the task” (Alder, 2007, p.163) and could improve the driving behavior.

II.7 Research Gap23, 24

Greater insight is needed into how eco-feedback technologies, such as CPM and/ or mobile feedback technologies (e.g., smartphones), can be used to influence drivers to adopt a sustainable driving style without classroom training. Feedback technolo-gies can inform the driver about eco-driving techniques and provide information about their fuel consumption. Improving drivers’ understanding of their fuel con-sumption could stimulate their motivation to drive more sustainably. The feedback intervention type must be displayed so that obstruction is minimal, and that users are able to decipher relevant information easily using their peripheral vision without much effort or increased cognitive workload (Meschtscherjakov et al., 2009). Moti-vating participants to improve their driving behavior will play a significant role in promoting fuel efficiency. When deliberated within the correct context, eco-feedback must correlate with drivers’ personal beliefs and culture to influence them in modifying their behavior (Foxx and Schaeffer, 1981).

The FIT of Kluger and DeNisi (1996) will be applied as a theoretical foundation to evaluate changes in task performance (i.e. change of driving behavior) in response to eco-feedback technologies. Feedback through ICTs is recognized as part of the concept of CPM evaluated by Alder (2007) and Nussbaum and duRivage (1986).

23 Published paper: Tulusan, J., Staake, T., Fleisch, E., Direct or indirect sensor enabled eco-driving feed-back: Which preference do corporate car drivers have?, Internet of Things 2012 – Third International Confer-ence on the Internet of Things (IoT 2012), Wuxi, P.R. China, October 2012. 24 Published paper: Tulusan, J., Steggers, H., Staake, T., Fleisch, E., Supporting eco-driving with eco-feedback technologies: Recommendations targeted at improving corporate car drivers’ intrinsic motivation to drive more sustainable, Energy Informatics 2012 (EI 2012), Atlanta, Georgia, United States, October 2012.


CPM enables organizations to provide their employees with information about their personal performance by collecting relevant data when working. This improves their awareness of areas for development and enables them to modify their behavior (Grant and Higgins, 1989). How these concepts were applied to this research and evaluated via two field tests with corporate car drivers is explained in the following chapters.

Private and Corporate Car Drivers’ Preferences for Eco-feedback Types 43

III Private and Corporate Car Drivers’ Preferences for

Eco-feedback Types

III.1 Overview and Research Question

Based on the eco-driving fundamentals and technologies introduced in Chapters 0, this chapter evaluates the preferences of private and corporate car drivers for feed-back types presented by eco-feedback technologies. Studies by Graham et al. (2011) have indicated that feedback detailing financial savings (i.e. reduced fuel costs) and/ or CO2 savings influence the driving behavior of private car drivers. Feedback relat-ed to financial savings had a stronger influence on reducing car usage than that re-lated to CO2 savings, and drivers achieved greater fuel reductions when both types were given, as opposed to only one type (Graham et al., 2011). However, how feed-back given by eco-feedback technologies influences corporate car drivers has not yet been explored. The following research sub-question was therefore identified:


Findings from this study provide insights into private and corporate car drivers’ preferences concerning eco-feedback types. Outcomes also confirm Graham et al.’s (2011) findings that eco-feedback must be easily understood and that feedback types illustrating financial savings have the strongest impact on private car drivers. Corporate car drivers preferred receiving feedback that detailed their fuel efficiency and included eco-driving tips, as this enhanced their understanding of how their driving could impact fuel efficiency. As expected, financial savings were not rele-vant. Feedback on sustainable driving techniques was received well by both driver groups but especially by drivers who had no understanding of eco-driving concepts or techniques.

III.2 Research Design and Research Methodology

The research design centered on two online surveys: one for private car drivers and one for corporate car drivers (see Figure 15). Findings were compared across these two driver groups, and results were used as a foundation for designing the experi-ments conducted with corporate car drivers (see Chapter IV and Chapter V).

44 Private and Corporate Car Drivers’ Preferences for Eco-feedback Types

Figure 15: Research Design

The first online survey ran for four weeks in August 2010 and was completed by private car drivers (n = 139). The goal was to obtain a better understanding of how feedback technologies function and which eco-feedback types are best for stimulat-ing eco-driving behavior. The second online survey (n = 131) ran for three weeks in August 2011 and was completed by corporate car drivers prior to the field test. This survey was used to acquire a better understanding of which eco-feedback types could best encourage this sample population to adapt their driving habits to drive more sustainably. Questions enquired about participants’ preferences for the feed-back types they would find helpful to see on an eco-driving feedback device. The complete surveys are shown in Appendixes A and B.

III.3 Data Evaluation and Findings

III.3.1 Data Evaluation

The survey included three different question types: five-point likert scale, multiple-choice, and open-ended. The five-point likert scale questions required participants to rate each item’s favorability between one and five, with one being the least fa-


vorable and five being the most. A rating higher than three (the scale mean) was interpreted as favoring the selected item. This enabled the researcher to make a comparison between the specified items and ascertain the most popular or well-received item in each category. For each item, a dependent two-tailed t-test was used to determine whether the deviation from the scale mean was statistically sig-nificant.

Multiple-choice questions were analyzed by calculating the overall percentage of respondents choosing each item (i.e. by summing the number of times each item was chosen and dividing this sum by the maximum number of times the item could have been selected). Open-ended questions gave participants the opportunity to ex-press themselves more uniquely than allowed via the predefined five-point likert scales or multiple-choice answers. The unstructured answers received from the open-ended questions were evaluated to identify similar opinions and ideas (i.e. fac-tors), which were then grouped into categories (see Figure 16).

Open-ended Questions

Evaluation and Reduction

Factors FactorsFactors Factors

Categories Categories

Ranked Factors and Categories

Figure 16: Categorization Process of Open-ended Questions

Cronbach’s alpha was calculated for the five-point likert scale questions, revealing a

value = 0.73 for the first survey and = 0.79 for the second. This suggests decent

consistency between the overall answers (Bernstein and Nunnally, 1994): as a


measure of the meaningfulness of a set of data, Bernstein and Nunnally (1994) rec-ommend a Cronbach’s alpha value above 0.7. The normality of the likert scale re-sponses was tested using the Shapiro-Wilk normality test. The result was non-significant, confirming that it was valid to compare answers between the two driver groups. The next two paragraphs describe the findings from the two online surveys; differences between private and corporate car drivers are highlighted.

III.3.2 Findings from Survey Conducted with Private Car Drivers

Of the 139 participants who completed the full survey, 64% were male and 36% female; 66% were between 25 and 34 years old. The number of years for which par-ticipants had been driving varied, with 43% of drivers having five to nine years ex-perience followed by 28% having zero to four years experience. Fifty-eight percent of participants drove on a daily basis, and sixty-eight percent declared that they were aware of feedback devices that targeted improvements in driving behavior. As shown in Figure 17, navigation systems were the most widely known devices, with 35% of respondents being familiar with them, followed by trip computers (25%), on-board gauges (23%), and smartphone applications (only 13%). This demon-strates that only a minority of drivers is aware that there are smartphone applica-tions available to help them improve their driving style. Other in-car technologies known to the participants were flat/ short shift and cruise control.

Figure 17: Awareness of Feedback Devices

89 (35%)

64 (25%) 57 (23%)

33 (13%)

9 (4%)

0 10 20 30 40 50 60 70 80 90

100

Navigation System

Trip Computer

On-board Gauge

Smartphone Application

Other (please state)

n (%); N=252


Fifty-one percent of the drivers had used feedback devices while driving, and 68% of these drivers reported that their driving style had changed as a result. Changes in driving behavior were explored in detail via the open-ended questions. Analysis of this data identified four categories (see Figure 18): a) more sustainable driving, b) car handling during driving, c) perception of the interface, and d) improved driving experience.

Figure 18: Changes while using a Feedback Device

A strong indication of more sustainable driving behavior was a motivation to reduce fuel consumption or fuel costs, which aligns with the literature on why private car drivers improve their driving style (Johansson et al., 2003). An additional contrib-uting factor was car handling: drivers aimed to drive their car more smoothly and improved gear changes to up-shift the gears more rapidly. This supports the goal of possibly slowing down the car’s eventual deterioration (EcoDrive, 2012). How par-ticipants perceived the interface used for the feedback device was an influential fac-tor; an interface that was too much of a distraction from driving was considered un-favorable from a safety point of view. The ease of interpreting the information pro-vided by the feedback device was thus important; this was also found by Meschtscherjakov et al. (2009). The final category defined was the impact of GPS technology on driving experience. Factors relevant to this category were improved driving experience and finding destinations easier because a GPS-enabled system provided routing options.

Que

stio

n: C

ould

you

des

crib

e th

e ch

ange

s in

the

way

you

dro

ve w

hile

usi

ng a

feed

back

dev

ice? Sustainable

drivingReduce fuel consumption as a motivation (24)

Driving more smoothly (13)

Switching gears more often (5)

More comfortable with routing decisions (4)

Distraction (5)

No difficulty finding destination (3)

Gaming aspect (1)

Time saving (2)

Environmentally friendly as a motivation (2)

Car handling

Perception of the

interface

Driving experience GPS usage


Having introduced the topic area, the next survey section queried participants’ pref-erences regarding specific feedback types. These items were graded using a five-point likert scale, and there was an overall tendency to favor most feedback types (see items with p < 0.05 in Figure 19).

Items Mode Mean Std. Dev.

Deviation from Scale

Mean

Sig. (2-tailed)

AVG fuel efficiency (monthly) 5 4.47 0.86 1.47 .000*** AVG fuel savings (monthly) 5 4.02 1.12 1.02 .000*** Personalized eco-driving tips 5 3.86 1.16 0.86 .000*** Alternate. eco-efficient route 5 3.79 1.19 0.79 .000*** Goal setting 3 3.46 1.14 0.46 .000*** AVG CO2 emission (monthly) 3 3.39 1.24 0.39 .000*** Social comparison 3 3.16 1.24 0.16 .092 Rewards 3 3.03 1.30 0.03 .775

Note: * p < 0.05; ** p < 0.01; *** p < 0.001

Figure 19: Preferences for Feedback Types

Statistically insignificant items were: a) fuel consumption feedback in comparison to other users who drive the same distance (item - social comparison) and b) a re-ward when the driver has achieved an efficient driving behavior (item - rewards).

17%

10%

11%

5%

6%

6%

6%

2%

18%

23%

11%

15%

8%

5%

3%

1%

27%

25%

29%

32%

21%

22%

13%

11%

24%

25%

27%

26%

30%

28%

35%

22%

15%

17%

22%

22%

35%

38%

41%

65%

0% 20% 40% 60% 80% 100%

Rewards

Social comparison

AVG CO2 emission (monthly)

Goal setting

Altern. eco-efficient route

Personalized eco-driving tips

AVG fuel savings (monthly)

AVG fuel efficiency (monthly)

1 Strongly Disagree 2 3 4 5 Strongly Agree


Since the results are not statistically significant, no clear preference for these two feedback types can be concluded from the answers of this survey.

All other feedback types were relevant for private car drivers, with a statistically significant deviation from the scale mean (p < 0.05). Participants most preferred to see their own average fuel efficiency (4.47 out of 5), followed by the potential sav-ings they could have achieved by driving sustainably (4.02 out of 5). Receiving feedback consisting of personalized eco-driving tips, an alternative eco-efficient route, goal setting for fuel efficiency improvements, or monthly average CO2 emis-sions was not as important. In sum, private car drivers had the strongest preference for feedback that detailed their fuel consumption (87% agreement), followed by that which noted potential financial savings (76%).

When asked how the feedback should be provided, 52% of the respondents desired visual feedback (i.e. shown by a symbol), 43% preferred both visual and auditory (i.e. through a voice), and 6% auditory only. This shows that feedback should be presented visually or visually and audibly. Multiple-choice questions were used to probe participants’ awareness of the impact of CO2 emissions on the environment and their views of the importance of reducing these emissions (see Table 7). When asked if they understood what was meant by ‘100 grams CO2 emission per km,’ 54% were unable to translate this into a meaningful context. This clearly indicates drivers’ lack of understanding of the impact of CO2 emissions on the environment. Of the 54%, 71% felt that improving their understanding of CO2 emissions would better motivate them to adapt their driving style to be more eco-friendly. Sixty-seven percent of participants agreed that knowing their CO2 emissions would be necessary for them to improve their driving style.

Table 7: Private Car Driver’s Awareness of CO2 Emissions

Yes No Understand ‘100 grams CO2 emission per km’ 46% 54%

Motivated to improve FE if un-derstanding of CO2 improved 71% 29%

Relevance providing CO2 emis-sions feedback to improve FE 67% 23%


Section three of the questionnaire asked participants to order different types of feedback technologies from most to least preferred (see Figure 20). The technolo-gies to rank were three on-board devices, two smartphone applications, and one ap-plication running via the internet. The objective of this line of questioning was to obtain a better understanding of which feedback technologies were preferred.

Items Mode Mean Std. Dev. Deviation from Scale

Mean

Sig. (2-tailed)

Color-changing screen 6 4.40 1.52 0.90 .000***

Leaves 5 3.74 1.67 0.24 .085 Navigation Sys-tem 6 3.71 1.77 0.21 .136

Silver ball iPh-one app. 2 3.41 1.56 -0.09 .500

CO2 emissions iPhone app. 3 3.24 1.54 -0.26 .040

PC application 1 2.60 1.64 -0.90 .000*** Note: * p < 0.05; ** p < 0.01; *** p < 0.001

Figure 20: Preferences for Feedback Technologies

The most favored feedback system was a color-changing dashboard (mean rating: 4.4 out of 6) that showed the fuel consumption according to the driving style; the

37%

16%

11%

16%

13%

5%

19%

19%

23%

15%

14%

10%

16%

22%

20%

12%

16%

14%

11%

19%

16%

21%

18%

16%

9%

16%

18%

12%

21%

24%

8%

8%

11%

24%

18%

32%

0% 20% 40% 60% 80% 100%

PC application

CO2 emissions iPhone app.

Silver ball iPhone app.

Navigation system

Leaves

Color-changing screen

Least preferable Most preferable


deviation from the scale mean showed statistical significance (p < 0.05), with 72% of respondents preferring it (four or more scale points). The least important was found to be the PC Application (mean rating: 2.6 out of 6), with 72% giving it lower preference (three or fewer scale points). This feedback technology provided indirect feedback about an individual’s CO2 emissions after the driver manually entered their fuel consumption into their PC. Other feedback technologies did not show a statistically significant outcome (p > 0.05), thus not eliciting a clear preference from the participants.

Open-ended questions were utilized to explore participants’ reasoning behind their preferences for feedback systems. Once the coding process was completed and rele-vant influential factors were determined, it became clear that a color-changing dashboard was the most favored because this enhanced drivers’ ease of interpreting the feedback they received (see Figure 21). Using three different colors to illustrate their driving style was considered to be the most straightforward and non-distracting method for presenting the necessary information to drivers.

No distraction (11)

Real-time feedback (2)

Easy to understand (34)

Easy to track (8)

Attractive design (6)

Relation to driving behavior (4)

Most

Lack of information (3)

Too basic (3)Least

Que

stio

n: W

hat w

ere

the

reas

ons y

ou li

ked

the

Col

our C

hang

ing

Das

hboa

rd th

e m

ost/l

east

?

Figure 21: Reasons for Favoring the Color-changing Feedback System

A PC application was considered to be least favored because of the timing of the feedback (after the event); the application was also found to be unattractive and too complicated in design (see Figure 22). Drivers preferred real-time feedback, as they immediately received direct feedback regarding their driving behavior.


Attractive design (2)

Easy to understand (4)

Feedback type (2)

Relation to CO2 savings (2)

User friendly (2)

Most

No real-time feedback (18)

No useful feedback (3)

Inappropriate concept (2)

Unattractive design (17)

Complicated (12)

Distraction (4)

Not user friendly (2)

Least

Que

stio

n: W

hat w

ere

the

reas

ons y

ou li

ked

the

PC

App

licat

ion

the

mos

t/lea

st?

Figure 22: Reasons for Not Favoring the PC Application Feedback System

The next open-ended question asked participants’ about their views on which feed-back types would be beneficial to display on an eco-feedback device in order to as-sist them to drive more sustainably. Since this question was not mandatory, fewer answers were given than to other questions. Unstructured answers were coded, de-fined into factors, and grouped into categories reflecting similar content or meaning. Several themes were apparent (see Figure 23). Appearance was found to be im-portant, as user-friendly feedback should not distract the driver during driving. Ad-ditional information that participants found to be useful was more detailed material, such as long-term data on fuel efficiency. This information would help participants acknowledge the financial benefits of sustainable driving in addition to its environ-mental impact.


Que

stio

n: D

o yo

u ha

ve a

ny fu

rther

idea

s to

disp

lay

info

rmat

ion

abou

t eco

-eff

icie

nt d

rivin

g?Appearance

Support

Visualization

No distraction (9)

More information (6)

Environmental related (7)

Focus on fuel efficiency (4)

Long-term data acquisition (2)

Gear shift recommendation (2)

Financial related (8)

Information provisioning

Visual support (2)

User friendly (3)More efficient information display (2)

Figure 23: Suggestions for Additional Eco-efficient Feedback

III.3.3 Findings from Survey Conducted with Corporate Car Drivers

The second survey was completed by 131 (29%) of the 450 corporate car drivers working for the case study company. The survey ran in August 2011 for three weeks, prior to the field tests. The demographic questions revealed that 77% of re-spondents used their car every day or at least three to four times per week. Seven percent did not know their average monthly distance driven, 27% drove between 2,001 and 3,000 kilometers, and 16% above 3000 kilometers per month (see Figure 24), which is consistent with the average yearly total distance driven by corporate car drivers in Europe (DTLR, 2001).


Figure 24: Average Kilometers driven per Month

As corporate car drivers have their fuel costs reimbursed by their company, 26% of drivers did not know their monthly fuel costs, while 37% estimated this cost to be between 201 and 400 CHF and 25% estimated it to be between 0 and 200 CHF.

Respondents rated the importance of receiving a certain type of feedback for im-proving fuel efficiency using a five-point likert scale, and results are shown in Fig-ure 25.

6%

14%

18%

23%

25%

7%

5%

13%

14%

18%

22%

12%

25%

16%

21%

28%

27%

28%

31%

18%

37%

42%

16%

17%

19%

0% 20% 40% 60% 80% 100%

Personalized eco-driving tips

AVG fuel efficiency (monthly)

AVG CO2 emission (monthly)

AVG km driven (monthly)

AVG fuel savings (monthly)

1 Strongly Disagree 2 3 4 5 Strongly Agree


Figure 25: Preference of Feedback Types to improve Fuel Efficiency

Drivers most preferred receiving tips to improve their driving habits, with a mean value of 3.83 out of 5, 65% agreement (four or more scale points), and a statistically significant deviation from the scale mean (p < 0.05). Receiving information about their average fuel efficiency was a close second (3.77 out of 5), with 69% agree-ment. As fuel costs were reimbursed, drivers considered feedback regarding their average monthly fuel costs to be least useful (2.89 out of 5), though this result is not statistically significant.

When asked if they understood what was meant by ‘100 grams of CO2 per km,’ 74% of respondents answered ‘yes,’ but 37% were unable to translate the value into a feasible context (see Figure 26). Twenty-seven percent did not understand this value at all, and 60% of these drivers said they would have a better understanding if a vivid comparison was given (e.g., ‘100 grams of CO2 per km is similar to using a computer for 4.5 hours.’) Finally, only 53% believed that knowing their CO2 emis-sions figures was influential in improving their average fuel efficiency.

Scenario Mode Mean Std. Dev. Deviation from Scale

Mean

Sig. (2-tailed)

Personalized eco-driving tips 5 3.83 1.18 0.83 .000***

AVG fuel efficiency (monthly) 5 3.77 1.41 0.77 .000***

AVG CO2 emission (monthly) 4 3.11 1.32 0.11 .319

AVG km driven (monthly) 4 3.06 1.43 0.06 .603

AVG fuel savings (monthly) 1 2.89 1.45 -0.11 .364

Note: * p < 0.05; ** p < 0.01; *** p < 0.001


Figure 26: Understanding of the Meaning of 100 grams of CO2 per km

III.4 Discussion

Only 13% of the private car drivers participating in the study had knowledge of the emerging eco-feedback smartphone applications available on the market. This aligns with the fact that smartphone device penetration was only 19% in 2010, though this is expected to rise to 40% by 2014 (Gartner, 2011). Better-known feed-back technologies included navigation systems (35%), trip computers (25%), and on-board gauges (23%).

A strong preference for undisruptive feedback technologies was found, which con-firms the findings of Meschtscherjakov et al. (2009). The on-board unit that changes color from blue to green according to the driving style was most popular, as it was efficient and effective at indicating drivers driving style in a real-time setting. The PC application was least favored due to the offline nature of its feedback and its complicated design format, implying that drivers like receiving direct feedback when driving.

As shown in Figure 27, private car drivers preferred receiving information about their personal fuel consumption (4.47 out of 5) and average fuel savings (4.02 out of 5) to receiving personalized tips for a more ecological driving style (3.86 out of 5). In comparison, corporate car drivers preferred to receive tips that helped them im-prove their driving style to become more eco-friendly (3.83 out of 5), closely fol-lowed by information about the average fuel efficiency (3.77 out of 5). The least interesting information for this user group was, not surprisingly, the average month-


ly total cost of petrol (2.89 out of 5). This is in line with the existing literature: fi-nancial savings are important reasons for private car drivers to change their driving behavior (Graham et al., 2011), whereas corporate car drivers are reimbursed by their company and therefore consider financial savings less important.

Items Mode Mean Std.

Dev.

Deviation from Scale

Mean

Sig. (2-tailed)

Private Car

Drivers


AVG fuel savings (monthly) 5 4.02 1.77 1.02 .000***


Corp. Car

Drivers


AVG fuel savings (monthly) 1 2.89 1.45 -0.11 .364


Figure 27: Preferences for Feedback Types, Private vs. Corporate Car Drivers

Both groups agreed that information about their CO2 emissions was less important; private and corporate car drivers ranked feedback about CO2 emissions at 3.39 out of 5 and 3.11 out of 5, respectively. Limited understanding of the meaning of CO2

4,47 4,02 3,86 3.77

2,89

3,83

0,00

1,00

2,00

3,00

4,00

5,00

Personal fuel consumption

Financial savings Personalized eco-driving tips

Private Car Drivers Corpor. Car Drivers


emission figures may be one reason why participants did not value this information. This is supported by the fact that only 46% of private car drivers understood the meaning of the CO2 emissions figures given, compared to 74% of corporate car drivers. Of the latter group, though, only 37% were able to relate it to their own fuel efficiency.

III.5 Conclusion

This chapter’s findings provided a first insight into the potential impact feedback technologies could have on driving behavior. They also identified the ideal feed-back types to be given using feedback technologies and to be tested during the field tests. Both driver groups preferred to receive information about their actual fuel consumption. Details related to financial savings were relevant for private car driv-ers, but CO2 emission figures were not as important due to drivers’ difficulties in relating this information to their driving styles. Consistent with findings from Gra-ham et al. (2011), feedback about financial savings was more relevant for private car drivers than was information about CO2 savings. Further education is required to raise drivers’ understanding of CO2 emissions as they relate to fuel consumption figures, enabling drivers to relate this meaningfully to energy efficiency. For exam-ple, CO2 emissions can be compared to those resulting from the electricity con-sumption of a laptop. Based on these findings, it can be argued that eco-feedback technologies have the potential to change driving behavior, particularly when the feedback is real-time, unobtrusive, and easy to understand. Offline feedback (i.e. after driving) has the advantage of providing detailed data about driving behavior but must be presented in an easy-to understand way.

III.6 Limitations and Future Research

The first survey with private car drivers was conducted in 2010, whereas the survey with corporate car drivers was conducted in 2011. To make an ideal comparison of the driver groups, it would have been better to conduct the surveys at the same time. However, these preliminary findings were still valuable and could be incorporated into the design of the field tests. Graham et al. (2011) found that showing CO2 sav-ing figures to private car drivers had a positive impact on reducing their car usage. The preferred feedback types identified in these surveys will be used in the field


tests to attempt to alter corporate car drivers’ driving behavior, as explained in the following chapters. Whether corporate car drivers improve their fuel efficiency needs to be further evaluated; this is especially important as the use of smartphone technologies is continuously rising (Gartner, 2010; Morgan Stanley, 2010).

The next chapter, Chapter IV, evaluates how direct/ real-time feedback, provided by an eco-driving smartphone application, influences the fuel efficiency of corporate car drivers. Chapter V then explains how indirect feedback, provided via a monthly fuel consumption email, impacts corporate car drivers’ fuel consumption.

Direct Feedback by an Eco-driving Smartphone Application 61

IV Direct Feedback by an Eco-driving Smartphone Applica-tion25, 26

IV.1 Overview

The personal transport sector constitutes an important target of energy conservation and emission reduction programs. In this context, eco-feedback technologies that provide information on the driving behavior have shown to be an effective means to stimulate changes in driving in favor of both, reduced costs and environmental im-pact. This field test extends the literature on eco-feedback technologies as it demon-strates that a smartphone application can improve fuel efficiency even under condi-tions where monetary incentives are not given, i.e. where the drivers do not pay for fuel.

The field test, which took place with 50 corporate car drivers, demonstrates an im-provement in the overall fuel efficiency by 3.23%. The theoretical contribution un-derlines the assumption that direct/ real-time feedback provided by an eco-driving smartphone application can favorably influence behavior even without direct finan-cial benefits for the agent. Receiving real-time feedback was also preferred com-pared to indirect feedback (accumulated over time or offline feedback). Given the large share of corporate cars, findings are also of high practical importance and mo-tivate future research on eco-driving feedback technologies.

25 Published paper: Tulusan, J., Staake, T., Fleisch, E., Providing eco-driving feedback to corporate car driv-ers: what impact does a smartphone application have on their fuel efficiency, 14th ACM International Con-ference on Ubiquitous Computing (UbiComp), Pittsburgh, Pennsylvania, United States, September 2012. 26 Published paper: Tulusan, J., Staake, T., Fleisch, E., Direct or indirect sensor enabled eco-driving feed-back: Which preference do corporate car drivers have?, Internet of Things 2012 – Third International Confer-ence on the Internet of Things (IoT 2012), Wuxi, P.R. China, October 2012.

62 Direct Feedback by an Eco-driving Smartphone Application

IV.2 State of the Art and Related Work

IV.2.1 Feedback Types27

As mentioned in Chapter 0, feedback geared towards promoting sustainable driving can be presented using various mediums, such as: email, brochures, classroom train-ing, driving simulator, videos, online portals, and mobile devices. It is important to consider when and what type of feedback is required to ensure the success of an eco-driving intervention. Direct feedback should not endanger the driver’s safety (Donmez, 2007). Distraction is a crucial factor which must be considered because this can lead to increased risk levels and even fuel efficiency (Meschtscherjakov et al., 2009). Indirect feedback, on the other hand, can be proclaimed to be ‘richer’. For instance, feedback presenting electricity consumption for households includes data reflecting the actual consumption shown on tables, graphs, charts, or scales (Graml et al., 2010).

In the following section, a list of feedback types relevant for this research will be introduced and should instill a fundamental understanding about how and when eco-driving feedback can be utilized to motivate drivers to reduce energy consumption.

Context-sensitive feedback is defined by Dey et al. (2001, p. 108) as “any infor-mation that can be used to characterize the situation of entities (e.g. person, place, or object) that are considered relevant to the interaction between a user and an ap-plication, including the user and the application themselves”. Arroyo et al. (2006) used context-sensitive feedback with a sensor enabled logging device to remind in-dividuals to appropriate driving techniques to reduce their stress and improve safe-ty. Studies by Meschtscherjakov et al. (2009), Hutton et al. (2001), and Smiley et al. (1989) have demonstrated an impact on drivers’ performance when providing feed-back about their driving ability through a support system. Consequently, they found that the driving performance can be enhanced by providing feedback in relation to the specific context.



Direct or real-time feedback also known in the literature as ‘just-in-time’ feedback provides immediate feedback within a certain situation (Fogg, 2002), and has been applied in various fields to motivate behavioral change, e.g. encourage seat belt use (Geller et al., 1982) and to reduce electricity consumption (Graml et al., 2010; Win-kler and Winett, 1982). With the emergence of sensor-enabled mobile devices and their capabilities of collecting data through GPS and acceleration sensors, it is now possible to provide context-sensitive just-in-time feedback to drivers (Froehlich et al., 2010). It highlights various driving habits, e.g. aggressive or smooth accelera-tion and braking, which is immediately illustrated through screen color changes or by an fuel/ CO2 emissions consumption scores (Froehlich et al., 2010). One must consider the vital notion of ensuring that just-in-time feedback does not compromise the safety of its user by increasing their cognitive workload and distracting their attention from the act of driving (Donmez, 2007). As driving conditions have the tendency to change rapidly, feedback can impede on the driver’s ability to respond to the task demands and impair their task performance (Arroyo et al., 2006). This must be taken into consideration when providing eco-driving real-time feedback to corporate car drivers.

Information providing indirect feedback or not in real-time is collected over a dura-tion of time, which may range from several minutes to one or more driving cycles, affording drivers an insight into their general driving style and how this may evolve over time. It can be shown in different forms, e.g. in an accumulated form or offline (Graml et al., 2010). Accumulated feedback can be imparted using either fuel con-sumption figures or using schematics, e.g. leaves growing (Formosa, 2009). The idea of rewards is conceptualized by accumulated feedback, as the more efficient the driver evolves the more advanced the figures or schemas become. Examples of systems that can be attributed to this category are Ford’s SmartGauge (Formosa, 2009) and Honda’s EcoGuide (Froehlich et al., 2010). A prototype sensor enabled mobile tool, evaluated by (Froehlich et al., 2009), illustrated that users were capable of positively perceiving accumulated feedback regarding green transportation prac-tices. Acknowledgment can be given to the rewarding nature of advanced technolo-gies that inspire users to adapt their behavior and continue to reduce their energy consumption (Froehlich et al., 2010). Offline feedback is a decontextualized type of feedback;; e.g. a detailed historical breakdown of driver’s fuel consumption, CO2


emissions, acceleration, braking, and gear shifting patterns is formulated once their general driving behavior is monitored and analyzed. Evidently, feedback is not pro-vided during the act of driving and so can be combined with social networks, where challenges from within the online community could potentially improve behavior. This form of feedback is known as social normative feedback, as defined by Cialdini (2003), and applied to change environmental behavior contributing to litter-ing, towel reuse (Goldstein et al., 2008), and energy consumption (Siero et al., 1996). Fiat’s eco:Drive is an example of where this type of feedback is used. Driv-ers are able to compare their own fuel efficiency with their peers on Fiat’s eco:Drive online community.

Indirect feedback can also be provided prior to driving, as this category includes systems aimed at informing the driver about the planned route. This information is largely delivered in the form of navigation systems or internet websites through which multiple routes are presented to drivers regarding distance, travel time, or fuel optimization. This enables drivers to make an informed decision about the best possible route to meets their situational demands.

IV.2.2 DriveGain Eco-Driving Smartphone Application

An eco-driving smartphone application called DriveGain, a product of DriveGain Ltd. in UK (DriveGain Ltd., 2012), was used for the purpose of the first field test. After extensive evaluation of the aforementioned products and the feedback re-ceived from the user study introduced in Chapter III, this application was chosen because of the quality of eco-driving feedback and access to the data collected by the application from the drivers participating in the field test. Besides a very com-prehensive car model database of more than 25,000 vehicle types, the application provides different feedback types related to the eco-driving concepts, such as cor-rect gear change during acceleration and braking, and most efficient average speed depending for each vehicle type. Similar feedback types were also tested by van der Voort et al. (2001), they demonstrated that drivers were able to reduce their fuel consumption by 16% when using the support tool. It is important to acknowledge that this test was confined to a driving simulator. Feedback types identified to re-duce fuel consumption were correct gear changing during acceleration and smooth acceleration. These feedback types were also implemented within the DriveGain


application. The DriveGain application runs on the Apple operating system (iOS), therefore, only iPhone models 3G, 3Gs, 4, 4s and 5 are compatible. These models have a GPS and acceleration sensor that enable the DriveGain application to obtain data from these components.

Figure 28: Overview of DriveGain Application Interface28

The top third of the interface screen (see Figure 28) illustrates: optimal gear change (recommended gear), journey score, and type of vehicle. Feedback meters are locat-ed in the central third of the screen, and below this, functions to activate music, re-set the journey scores, upgrade with new feedback meters and settings of the appli-cation. Once the application has been started, the car type chosen (see Figure 29 and Figure 30) and a GPS signal received, it will log and monitor the driving behavior for this specific route. The application provides several feedback meter types which provide individual feedback towards acceleration, braking, CO2 emissions, or costs (see Appendix C). To focus on the eco-driving feedback concepts and to evaluate the difference direct or indirect feedback can have on a driving behavior, only two feedback meter types were activated for this experiment:

28 DriveGain Ltd. (2012), “DriveGain.” Retrieved September 19, 2012, from http://drivegain.com

Recommended Gear

Journey Score

Feedback Meter Types

Application Settings

Type of Vehicle


Figure 29: Settings Screen

Figure 30: Select Vehicle Screen

Figure 31: Advanced Savings Feedback Meter (direct feedback)

Figure 32: Fuel Savings Feedback Meter (indirect feedback)29



The first known as ‘Advanced Savings’ (see Figure 31) provides direct/ real-time feedback that combines acceleration, braking and the current vehicle speed. Aggres-siveness or smooth acceleration together with braking is shown in an efficiency score which is calculated in a percentage value (100% being best). Additionally, the current speed is shown on a scale. The second meter, known as ‘Fuel Savings’ (see Figure 32) provides indirect accumulated feedback in a three-minute interval for the average figures for acceleration, braking, and vehicle speed. Acceleration, braking, and vehicle speed is measured by the GPS receiver. All three feedback measure-ments are displayed on a scale categorized red to green (green being most ecologi-cal), as well as a numerical score from 0 (least) to 100 (representing most ecologi-cal).

The journey score visible on the top right corner of the screen, 0 (least) and 100 (most ecological), is calculated from data collected regarding acceleration, braking, and speed values with respect to each car type. The recommended gear feature prompts drivers with a manual gearbox when to change gear. Only two gears, ‘P’ for parking and ‘N’ for driving were represented for cars with automatic gearboxes. Relevant data for this study was collected from this application: 1. Journey start and end location (measured by the GPS receiver); 2. When and duration application was used in seconds; 3. Distance travelled in meters per journey; 4. Duration each type of feedback meter was used in seconds; 5. Values for acceleration, braking and ve-hicle speed.

Besides the eco-driving information displayed on the application, additional data sets were recorded by the application and automatically transmitted to the compa-ny's server after each journey. This data were accessible for each driver on an online driving portal (ODP). This information was ‘richer’ in detail, as the raw data were evaluated using an information system and made visible to each individual user on their ODP account. An overview cockpit summarized recent journeys driven, fuel usage, fuel efficiency, high speeds, and CO2 emissions on the first page (see Figure 33). Further details were broken down on a dedicated sub-page for each section. This allowed appraisal of offline feedback during the experiment, as every driver had to register as a user to the ODP where they were then able to review their in-formation after any driving episode.


Figure 33: Detailed Journey View shown in Online Driving Portal30

IV.3 Research Questions and Hypothesis

The literature indicates a discrepancy about the impact an eco-feedback technology has on driving behavior. Generally speaking, the approval of numerous in-vehicle on-board feedback systems is largely reflected on in a positive manner (Boriboon-somsin et al., 2010; Meschtscherjakov et al., 2009; H. Lee et al., 2010; S. Siero et al., 1989). Meschtscherjakov et al. (2009) highlighted that inflated risk levels can be induced by distraction through inappropriate feedback. This was further illustrated by Lee et al. (2010) who found that drivers, as a result of amplified cognitive work-load and stress levels, ignored feedback provided by a non-user-friendly design or a poorly positioned system on their dashboard during a field test.



To understand if an eco-driving smartphone application has an impact on the fuel efficiency, the following hypothesis was defined to answer sub-research question 2.


Hypothesis 1: Corporate car drivers who use an eco-driving smartphone application improve their average overall fuel efficiency, with µ1 representing the liters of pet-rol per 100 kilometer without and µ2 with eco-driving feedback.

(H1) H0 = µ1 ≤ µ2;; H1 = µ1 > µ2.

Real-time and context related feedback is integral when promoting fuel efficiency of individuals when driving. The ability to control the frequency of the feedback received has been proven to motivate users to change their behavior. This is known as “feedback control,” e.g. individuals are able to control the amount and timing of the feedback they receive (Corbett and Anderson, 2001). The DriveGain application allows corporate car drivers to choose between different feedback types, such as ‘Advanced Savings’ with direct feedback that provides real-time feedback about one’s actual driving style and ‘Fuel Savings’ with a three-minute interval feedback that provides the feedback in an accumulated format. The second hypothesis was defined to answer sub-research question 3.


Hypothesis 2: Corporate car drivers who use an eco-driving smartphone application use real-time feedback for longer periods of time when compared to accumulated feedback, with µ1 representing average time in seconds regarding real-time feed-back and µ2 representing average time in seconds of which accumulated feedback was used.

(H2) H0 = µ1 ≤ µ2;; H1 = µ1 > µ2.

Drivers were also able to review their eco-driving data offline, known as indirect feedback, and summarized in the ODP. Literature in the field of energy savings in


households reflect that feedback related to energy consumption should also be pro-vided immediately using a non-obstructive form (Van Houwelingen and Van Raaij, 1989; Mountain, 2008). This enables individuals to relate feedback directly to a specific behavior and prompts an immediate response (Darby, 2006; Graml et al., 2010). The importance of indirect feedback in relation to eco-driving is not clear and so hypothesis three was defined to get a better understanding to appraise SQ3.

Hypothesis 3: Corporate car drivers who use an eco-driving smartphone application prefer to receive feedback during driving in comparison to offline feedback provid-ed via a ODP, with µ1 representing the preference to receive feedback during driv-ing through the smartphone application and µ2 representing the preference of re-ceiving the feedback via a ODP.

(H3) H0 = µ1 ≤ µ2;; H1 = µ1 > µ2.

IV.4 Research Design and Research Methodology

IV.4.1 Research Design

It was deemed appropriate to use surveys and a field experiment for this experiment in order to review drivers’ interaction with technology. This research design was founded on a sequential multi-method data collection approach that incorporated an assortment of approaches within a sequential order, permitting data from one level to be utilized in the next (Mingers, 2001). Before and after the field test, drivers were requested to complete a survey that allows monitoring and control of factors that relate to individual differences, such as: driving habits, environmental believes, and preference of feedback types. The opt-in experiment was announced in August 2011 by email by one board member of the case study company to 450 corporate car drivers based in Switzerland (see Figure 34).


Figure 34: Between-subjects Experimental Design

Sixty-two employees responded to the announcement email. The criterion to partic-ipate was ownership of an iPhone model 3G, 3Gs, 4 or 4s. Seven drivers with other phones had to be removed from the sample. Tank-refill data of the remaining 50 drivers were available prior to and during the introduction of the field test for each driver. This made it possible to calculate the drivers’ average baseline fuel efficien-cy from January to October 2011 (prior to implementing the experimental condi-tion). Once the average fuel efficiency (= baseline) was calculated, a random selec-tion of 25 drivers for the Control Group (CG) and 25 for the Treatment Group (TG)


took place. The normality of the distribution was met, as shown by the not signifi-cant outcome of the Shapiro-Wilk test for normality (with P > .05, skeweness = -.579 and kurtosis = 1.739). Participants in the TG used the eco-driving application for a duration of eight weeks from 24th of October to 16th of December 2011. As it would be in a non-study environment, it was left to the drivers in TG to decide how often and when the application would be used; however, they were informed that the application should be used on a regular weekly basis within the designated eight weeks.

The drivers could choose between the two different feedback types, ‘Advanced Sav-ings’ with direct/ real-time feedback and ‘Fuel Savings’ with indirect/ three-minute interval feedback. At the end of the experiment the recorded data from the applica-tion were evaluated (see Appendix D). Data were transmitted through the GSM network directly from the smartphone to the DriveGain Ltd. server after each rec-orded route and provided to the researcher after the experiment (see Appendix E). Participant’s agreement to automatically transmit the collected data for each route was mandatory to participate in this field test. This it made it possible to verify, without self-reporting errors, the frequency-use of the application together with the duration each type of feedback meter was used by each driver.

IV.4.2 Research Methodology

The pre-survey (n = 131) conducted in August 2011, gathered details about the gen-eral understanding of drivers’ fuel efficiency, CO2 emissions, and transportation modes (see Appendix B). The post-survey (n = 24) conducted in March 2012, in-quired about the environmental attitude of the driver, eco-driving feedback useful-ness of the eco-driving application, feedback types, usage patterns, and recommen-dations for improvement (see Appendix F). To get a richer understanding, semi-structured interviews with 15 drivers were conducted from March to April 2012 (see Appendix G), which investigated details about changes in usage patterns, i.e. how often and how long the DriveGain application was utilized by each driver throughout the eight weeks. Details about the relevance of feedback types, e.g. real-time vs. accumulated vs. offline feedback and the combination of a smartphone ap-plication with an ODP were also explored in-depth. Each interview lasted 20 minutes.


Besides the qualitative approach, quantitative findings were based on the data col-lected from the DriveGain application and fuel efficiency data collected from the petrol cards of each corporate car driver. Drivers’ monthly tank-refill details were collected from petrol credit cards provided by the company. For each tank-refill, the total number of kilometers driven and volume of petrol were recorded during the payment process as a normal prerequisite to get permission to refuel on the compa-ny account (a measure to prevent drivers from refueling other cars at the company’s expense). Tank-refill data were sent to the company through a mobility operator. This allowed the researcher to calculate and compare the fuel efficiency per tank-refill more objectively, as drivers only had to record the total number of kilometers shown on their in-vehicle system at the time of payment. The total number of liters required to calculate the fuel efficiency were automatically reported by the system from the petrol station, thus, eliminating the need for further self-reported input from drivers. At the end of the field test more than 7000 fuel fillings from the Con-trol (n = 25) and Treatment Group (n = 25) were recorded and more than 800 jour-ney data sets collected from the TG’s application usage analyzed.

The between subjects statistical evaluation for comparing the means of fuel effi-ciency between the CG (without use of smartphone application) and TG (with use of smartphone application) was an independent samples t-test. A bivariate Pearson correlation coefficient was computed to identify the relationship between the total duration of application usage and duration of direct or indirect feedback type (’Ad-vanced Savings’ or ‘Fuel Savings’).

IV.5 Data Evaluation and Findings

IV.5.1 Descriptive Evaluation

The descriptive evaluation of the survey (see Table 8) shows that 75% of the partic-ipants were male and 25% female. Twenty nine percent were aged between 45 and 49 years, 25% between 40 and 44, and 17% between 30 and 34, which shows a bias towards people above 40 years. Seventy five percent of participants used their car every day, 21% five to six times per week, and 92% drove on average more than


24,000 kilometers per year, which conforms with the corporate car drivers criteria (Department for Transport, 2011).

All drivers used the eco-driving application on a weekly basis; 37% used it seven to eight times, 25% three to four times, and 17% five to six times. When asking which routes and travel purpose the application was used for, of the 119 answers (multiple selections were possible) 95% used the application on motorways and business re-lated travel and 92% for distances longer than 40 kilometers. Fifty percent used the application for private travel and 46% for short distances.

Table 8: Sample Demographics

Variable Category Frequency In % Mode

Gender Female Male

6 18

25 75

Male

Age

Under 30 30-34 40-44 45-49 50-54 55-59

1 4 6 7 3 3

4 17 25 29 13 13

45-49

Years of Experience

0-9 10-14 15-19 20-24 25-30 More than 30 years

0 5 1 6 8 4

0 21 4 25 33 17

25-30

Car Usage

Less than once a month 1-2x per month Every week 1-2x per week 3-4x per week 5-6x per week Every day

0 0 0 0 1 5

18

0 0 0 0 4 21 75

Every day


Variable Category Frequency In % Mode

Total kil-ometers per year

Less than 6000 km Between 6000 – 12.000 km Between 12.000 – 18.000 km Between 18.000 – 24.000 km Between 24.000 – 30.000 km Between 30.000 – 36.000 km More than 36.000 km per year

0 0 0 2 4

10 8

0 0 0 8 17 42 33

Between 30.000 –

36.000 km

Usage of application per week

Never 1-2x 3-4x 5-6x 7-8x 9-10x More than 10x

0 2 6 4 9 2 1

0 8 25 17 38 8 4

7-8x

Routes/ Travel Purposes

Business related travel Motorway For long distances (more than 40 km) Country / rural road Private related travel For short distances (less than 40 km) Inner cities

23 23 22

18 12 11

10

96% 96% 92%

75% 50% 46%

42%

Business related trav-el; Motor-way

The next question asked participants about the influential aspects of the eco-driving smartphone application regarding their driving behavior on a 7-point likert scale, one representing least and seven most favorable (see Figure 35).


Question: The feedback given by the application influenced me in…


Mean

Sig. (2-tailed)

…shifting gears earlier. 7 5.45 1.75 1.45 .020* …a smoother acceleration. 7 5.13 1.90 1.13 .008** …smoother braking. 6 4.83 1.52 0.83 .013* …driving in the highest gear. 6 4.36 1.69 0.36 .492

...reducing average speed on motorways. 2 2.83 1.55 -1.17 .001**

Note: * p < 0.05; ** p < 0.01; *** p < 0.001

Figure 35: Influence on Driving Behavior

Sixty percent of drivers with a manual gearshift felt that the feedback given by the application influenced them to shift the gear earlier (5.45 out of 7). Seventy one percent of all drivers accelerated smoother (5.13) and 75% decelerated smoother (4.83). All items were also statistically significant (p < 0.05). The least influential factor was to reduce the average speed on motorways (2.83) with 72% not in favor (three or less scale points), which could relate to the absence of speeds limit on German motorways.

4%

4%

17%

10%

13%

8%

27%

42%

4%

8%

13%

30%

8%

4%

18%

4%

10%

17%

29%

18%

21%

10%

25%

46%

36%

4%

40%

29%

0 0,2 0,4 0,6 0,8 1

…shifting gears earlier.

…a smoother acceleration.

…smoother braking.

…driving in the highest gear.

...reducing average speed on motorways.

1 Strongly Disagree 2 3 4 5 6 7 Strongly Agree


IV.5.2 Statistical Evaluation

The initial findings of the demographical evaluation provided an overview of the sample size and an understanding about how the eco-driving application promoted eco-driving techniques. The following chapter answers the sub-questions SQ2 and SQ3 with a statistical evaluation using the software packages STATA and SPSS.

The first step included the identification of outliers. Five drivers were identified as outliers using the Grubb’s test due to very low or excessively high baseline con-sumption and so were removed from the sample. As each participant’s data reflect-ing each tank filling from the beginning of the year were available, and to control for seasonal effects, the fuel efficiency of each driver was calculated for each tank filling from the start of January 2011 until the treatment started, and during the treatment phase. Once the average fuel efficiency was calculated for these two timeframes and changes in percentage calculated, a paired samples t-test compari-son between participants’ (CG vs. TG) changes in fuel efficiency was made.

(H1) Before the t-test was conducted three outliers in the CG and two outliers in the TG in terms of atypical changes of consumption due to false entries of kilometers driven during the experiment phase were identified using Grubb’s outlier procedure. Three additional data sets were removed in the TG to control for periods where the application was not used. A total of CG = 22 and TG = 20 were used for the t-test. The outcome showed that all participants’ fuel efficiency changed, be it positively or negatively. The maximum improvement in the CG was -4.94%, and -12.15% in the TG. The minimum change in the CG was 8.40% and 1.92% in the TG. On aver-age the corporate car drivers in the TG had improved their overall fuel efficiency more so by the end of the treatment phase (M = -2.2, SE = 0.87) compared with the CG (M = 1.03, SE = 0.70). This difference between the means of CG and TG1 re-sult in a 3.23% fuel efficiency improvement (see Table 9) and was statistically sig-nificant p < 0.01 with t(40) = 3.23, p = 0.006 (see Table 10). As the Levene’s test was not significant (p = 0.459) (see Table 11), the equality of variances can be as-sumed and the H0 rejected in favor of the H1.


Table 9: Group Statistics

Id_trtmt_grp N Mean Std. Devia-

tion Std. Error

Mean

diff_trtmt_avg_fe_y 0 1

22 20

1.034 -2.202

3.291 3.875

.702

.867

Table 10: Independent Samples Test (t-test for Equality of Means)

Table 11: Independent Samples Test (Levene’s Test for Equality of Variances)

(H2) The analysis of driver data sets, inclusive of 800 recorded journeys, were used to compare the variations in duration-use of direct versus indirect feedback. It was evident that the application was used mainly during the week and at daytime. Fur-thermore, an average distance per route of 33 kilometers was travelled with an aver-age usage of the application at 28 minutes per route. The total average journey score rating of 75 indicates a good ecological driving behavior.

The total time the application was used by all participants during the field test was 416 hours. Data collected revealed that the real-time feedback meter was used 85.11% (354 hours) of the time compared to the accumulated feedback with only 2.65% (11 hours). For the remaining 51 hours (12.24%) the application was sus-

t-test for Equality of Means

t df Sig. (2-tailed)

Mean Difference

Std. Er-ror Dif-ference

Diff_trtmt_avg_fe_y Equal variances assumed Equal variances not assumed

2.926 2.902

40

37.495

.006 .006

3.236 3.236

1.106 1.115

Levene’s Test for Equality of Variance

F Sig.

Diff_trtmt_avg_fe_y Equal variances assumed .558 .459


pended due to no GPS signal, e.g. driving through a tunnel, which is quite common in Switzerland.

A bivariate Pearson correlation coefficient was computed to identify the relation-ship between the total duration of application usage and duration of direct or indi-rect feedback type used. Findings showed a strong positive correlation with statisti-cal significance between the total duration of application usage and usage of direct/ real-time feedback (Adv._Sav._Meter), r = .878, n = 759, p < .000. When compar-ing the total duration of application usage and usage of indirect/ accumulated feed-back (Fuel_Sav._Meter), a weak correlation with no statistical significance (p > 0.05) was shown with r = .190, n = 95, p = .065 (see Table 12).

Table 12: Correlation of Direct or Indirect Feedback vs. Duration

score distance total_duration Direct Feedback (Adv._Sav._Meter)

Pearson Correlation Sig. (2-tailed) N

-.178 .000

759

.800

.000

759

.878

.000

759 Indirect Feedback (Fuel_Sav._Meter)

Pearson Correlation Sig. (2-tailed) N

.205

.046

95

.205

.046

95

.190

.065

95

Qualitative interpretation of data collected via the online post-survey (24 answered from the total of 25 drivers) were evaluated in percentage (1% least preference to 100% highest preference) or on a scale from one to seven. An additional 15 inter-views were analyzed following an axial coding process to be able to compare the rich answers.

The post-survey supports the findings from the bivariate Pearson correlation, as 88% of the participants preferred to use real-time (“Advanced savings”) feedback compared to only 19% who opted to use accumulated (“Fuel savings”, every three minutes) feedback (see Figure 36).


Question: Which type of feedback did you choose to receive?


Mean Sig. (2-tailed)

“Advanced savings” (real-time feedback) 7 6.06 1.60 2.06 .000***

Real-time feedback was useful. 6 5.06 1.86 1.06 .028*

“Fuel Savings” (every three minutes) 2 2.41 1.94 -1.59 .004*

Used both feedback meters evenly. 1 2.22 1.48 -1.78 .000***

Three minutes inter-val feedback was use-ful.

1 2.00 1.73 -2.00 .000***

Note: * p < 0.05; ** p < 0.01; *** p < 0.001

Figure 36: Direct – vs. Indirect Feedback

When considering the usefulness of the application in stimulating eco-friendly driv-ing behaviors, real-time feedback ranked 5.06 and accumulated feedback (three minutes interval feedback) ranked only 2.0. All values have a statistical significance of p < 0.05.

The post-experimental interviews were valuable in attaining an insight into why participants favored real-time feedback; the immediate nature of real-time feedback

6%

38%

44%

63%

12%

11%

44%

22%

13%

11%

13%

17%

17%

11%

35%

33%

13%

6%

13%

53%

22%

6%

0 0,2 0,4 0,6 0,8 1

“Advanced savings” (real-time feedback)

Real-time feedback was useful.

“Fuel Savings” (every three minutes)

Used both feedback meters evenly.

Three minutes interval …



enabled participants to directly link eco-driving principles to their driving practices. Accumulated feedback lacked any significance with regards to usefulness when driving on motorways, as very similar feedback was given by the application to drivers with a constant average speed and when there was not much traffic.

With respect to these findings one is able to reject the H0 in favor of the H1; to con-clude, real-time feedback was more eminent and favorable compared to accumulat-ed feedback.

(H3) H2 was appraised using results from both the post-survey and interviews. The post-survey indicates that 71% of participants reviewed their journey scores on the ODP once or twice per week and only 12.5% three to four times (see Figure 37).

Figure 37: Frequency of Journey Score reviewed in the ODP

In fact, they preferred to receive real-time feedback during a driving episode (88%) or in a weekly summary (63%) (see Figure 38). Multiple selections were possible for this question. The multiple selections were further evaluated using a 7-point likert scale question format, which asked drivers how thy preferred the feedback to be provided (see Figure 39).

17 (71%)

3 (13%) 2 (8%)

1 (4%) 1 (4%) 0 (0%) 0 (0%)

0 2 4 6 8

10 12 14 16 18

1-2x 3-4x Never 5-6x 7-8x 9-10x More than 10x

N = 24; n (% of total N)


Figure 38: Preference of when Feedback should be received

Visualization (6.00) was most favorable with 88% agreement followed by weekly emails (4.88) with 61% agreement. Both items showed a statistical significant value of p < 0.05. Feedback through an online portal had the lowest mean with 3.21 and 63% disagreement, but, the statistical significance was above 0.05.

Question: I would like to receive feedback from the application about how I drive…


Mean

Sig. (2-tailed)

21 (88%)

15 (63%) 12 (50%) 11 (46%)

3 (13%) 3 (13%) 2 (8%)

0

5

10

15

20

25

Real-time during

Once a week

After driving

Once a month

Once every 3 months

Once a year

Once a day

n (%); N = 67

9%

25%

33%

29%

4%

4%

25%

8%

21%

4%

4%

13%

13%

8%

22%

17%

13%

8%

4%

13%

8%

8%

4%

46%

39%

8%

21%

17%

38%

9%

13%

4%

8%

0 0,2 0,4 0,6 0,8 1

…by visualisation

…by a weekly email

…by a computerized voice and …

…by a computerized voice

…by an online portal



…by visualization 6 6.00 1.22 2.00 .000*** …by a weekly email 6 4.88 1.75 0.88 .023* …by a computerized voice and visualization 2 3.33 2.14 -0.67 .141 …by a computerized voice 1 3.33 2.12 -0.67 .137 …by an online portal 1 3.21 2.15 -0.79 .084

Note: * p < 0.05; ** p < 0.01; *** p < 0.001

Figure 39: Preference of how Feedback should be provided

The interviews clearly identified that participants felt the ODP was beneficial in getting a summary of the journeys driven, supporting an improved understanding of driving behavior/ patterns, but did not play an influential role in modifying their actual driving style. As the data collected during the post-survey can only be de-scriptive statistically validated and the interviews interpreted, this hypothesis cannot be accepted or rejected. However, a trend towards receiving direct feedback through visualization compared to the offline feedback through an ODP is reflected.

Besides these findings related to hypotheses (1) to (3), another question queried about the preference to which feedback items the driving style should be compared with (see Figure 40).

Question: My driving style should be compared with…

29%

13%

17%

4%

8%

29%

38%

4%

4%

21%

13%

8%

4%

8%

8%

8%

13%

13%

13%

8%

21%

50%

63%

38%

13%

4%

29%

13%

8%

0 0,2 0,4 0,6 0,8 1

…with the same car model.

...who drive a similar amount of km.

...with the same job role in organization.

…with same driving style.

…from my organization.




Mean

Sig. (2-tailed)

…with the same car model. 6 6.00 0.88 2.00 .000*** ...who drive a similar amount of km. 6 5.63 1.17 1.63 .000***

...with the same job role in organization. 6 3.83 2.24 -0.17 .719

…with same driving style. 2 3.42 1.91 -0.58 .148 …from my organization. 2 2.92 1.56 -1.08 .002*

Note: * p < 0.05; ** p < 0.01; *** p < 0.001

Figure 40: Preference of Social Comparisons

A preference from drivers towards a comparison with the same car model (6.00) with 92% agreement and similar number of kilometers (5.63) with 89% agreement was identified. Both means were statistical significance (p < 0.05). Other feedback items for a comparison with other drivers, such as the same job role in the organiza-tion or the same driving style, were not as important and did not show a statistical significance (p > 0.05). The least important comparison item was drivers from the same organization with 68% disagreement and a mean of 2.92 (p < 0.05).

IV.6 Discussion

Comparing the improvement of corporate car drivers’ fuel efficiency of 3.23% with existing studies referred to in this thesis strengthens the underlying assumption that smartphone technologies can have a positive impact in promoting the eco-driving behavior of corporate car drivers. Studies with greater improvements in fuel effi-ciency where either conducted in a driving simulator (Van der Voort et al., 2001) or with a company where the management set a fuel reduction target of 5% (Siero et al, 1989). Existing studies did not evaluate improvements in corporate car drivers’ fuel efficiency where monetary incentives are not relevant. The data recorded by this application provides important insights into driving behavior and, in line with the body of knowledge, showed that a strong acceleration and shifting gears too late contributes strongly to fuel consumption, which supports the findings of the exper-iment conducted in the driving simulator by Van der Voort et al. (2001).

A strong positive correlation between total application and direct/ real-time feed-back usage underlines the assumption that real-time feedback was used more than


indirect/ accumulated feedback. Findings are in accordance with literature in the field of energy savings, which also identified direct/ real-time feedback as being favorable to participants (Darby, 2006; Graml et al., 2010). Explorative findings from the interviews provides additional support, as influencing driving styles to in-corporate more eco-friendly behaviors required immediate prompts when directly involved in the act of driving. This was especially important for the more experi-enced drivers, e.g. corporate car drivers, as their driving habits are that much more internalized (Stenner, 2009).

Existing findings, from the energy savings in household literature (Graml et al., 2010), revealed that providing energy consumption feedback through an online por-tal made it possible to convince stakeholders to change their electricity consumption habits, particularly when accompanied by social psychological concepts (Mountain, 2008). When considering sustainable driving, findings from the post-survey and interviews indicate that the ODP has a role in promoting participants awareness and understanding of eco-friendly driving concepts (e.g. details about acceleration/ braking per journey, total kilometers driven, or average speed), but do not strongly influence behavioral modification to encompass eco-friendly driving styles.

It is apparent that the feedback source together with the frequency of feedback un-derpins the relevance sensor enabled services can have in shaping corporate car drivers’ behavioral change towards eco-driving practices. When considering partic-ipants’ control over the level of feedback they received, e.g. frequently through real-time feedback or intermittently via accumulated/ offline feedback, existing findings in organizations recommend that by bestowing this ownership on participants en-hance their desire to modify their behavior at work (Corbett and Anderson, 2001). Furthermore, if feedback from mobile feedback technology sources were given in conjunction with face-to-face feedback, e.g. by management, it can be predicted that a greater improvement in fuel efficiency may be found. This assumption is corrobo-rated by Siero et al. (1989), who provided feedback to their postal lorry drivers via a mobile device as well as face-to-face and achieved a long-term fuel efficiency im-provement of 7.3%.


IV.7 Conclusion

The findings highlight how mobile feedback technologies, which provide context related eco-driving feedback during driving, improve corporate car drivers’ fuel efficiency. An existing smartphone application from the company DriveGain Ltd. (DriveGain Ltd., 2012) with eco-driving functionalities was selected and tested in an experimental setting with one company’s pool of drivers. The improvement in fuel efficiency of 3.23% with t(40) = 3.23, p = 0.006 demonstrates that sensor ena-bled mobile feedback technologies can play an important role in reducing a compa-ny’s overall CO2 emissions and petrol costs. Findings also strengthen existing litera-ture, eco-feedback technologies support the notion that feedback provided through sensor enabled mobile devices can empower their users to drive sustainably (Boriboonsomsin et al., 2010; Siero et al., 1989). Our findings extend the current research area, as participants in this study were not motivated by financial rewards.

This study appraised which type of feedback, direct or indirect feedback, was pre-ferred by the users. The application provided direct/ real-time or indirect/ three-minute accumulated interval eco-driving feedback to participants. Additionally, par-ticipants were able to review further driving related data in greater detail using an ODP. It is justified to conclude that direct/ real-time feedback provided during a driving episode was largely preferred over indirect feedback delivered via accumu-lated or offline feedback. The importance of an ODP within this eco-driving context appears to be limited, as participants indicated a greater preference to attaining di-rect feedback when driving; simply reviewing a summary of driving related figures on the ODP was not enough to warrant a change in their driving style. The main function of the ODP appeared to be in its summary of the driving routes available.

IV.8 Limitations and Future Research

Generalization of these findings may be limited due to the sample size of n = 50 and short field study duration. However, other studies in this research domain have had even smaller sample sizes (n < 20) (Froehlich et al., 2009) or the treatment period was shorter (H. Lee et al., 2010) or based on a driving simulator (H. Lee et al., 2010; van der Voort et al., 2001). Findings which entail a lager sample size were mainly conducted with truck drivers and did not evaluate an eco-driving smartphone


application with different feedback meters in combination with ODP, rather a fixed installed on-board system that constantly monitored the driving behavior. Therefore, this field study provides a preliminary insight into how and when feedback should be provided using an eco-driving application with sensor enabled feedback technol-ogy, within a corporate context to promote improvement of fuel efficiency. Future studies with a larger sample size and longer duration should evaluate long-term ef-fects how both feedback types combined, direct and indirect have an impact on driving behavior compared to each individual approach.

As findings from qualitative research methods are more difficult to generalize, find-ings from the interviews, where n = 15, should be met with caution. Hence, the va-lidity of H2 requires further investigation by means of an additional experiment, which specifically evaluates the impact of indirect feedback provided via an online portal vs. direct feedback provided during driving.

Participants for this study were selected from one company’s pool of drivers. As it was an opt-in field test, participants who chose to take part in the experiment may have already had a pro-environmental attitude or were technological affine, expos-ing a degree of bias in the sample selection. It was possible to control this aspect by analyzing their environmental beliefs and technological affinity using validated scales from energy saving literature in the post-survey. Findings indicated that driv-ers only had moderate (= partly agree) pro-environmental attitude (5.02 out of 7.0) as well technological affinity (5.1 out of 7.0); this reflects that results of this study can be related to other corporate car drivers with a similar environmental attitude and technological affinity.

Indirect Feedback by a Mobility Information System 89

V Indirect Feedback by a Mobility Information System31

V.1 Overview

The Green information systems (IS) discipline has the potential to support environ-mental sustainability in corporations by drawing employees’ attention to their ener-gy consumption. A sample of 240 corporate car drivers from one company were selected for the second field test; over a duration of three months their fuel efficien-cy was calculated using a mobility information system from which additional quali-tative data was used to provide a detailed monthly email reflecting customized feedback for each participant. The emails summarized driving related figures such as total kilometers driven per month, CO2 emissions, and tips to incorporate eco-friendly driving practices. A comparison of their average monthly fuel efficiency rating against that of their colleagues was made and illustrated using either a cate-gorical or continual scale. This provided them with social normative feedback with the desire to explore if this had a greater impact on reducing participants’ fuel con-sumption.

Results indicate that descriptive social normative feedback had a significant impact in motivating participants to improve their fuel efficiency. However, a difference in the overall fuel consumption savings between the groups who received the feedback by a categorical - or continual feedback scale was not apparent. These findings are vital in endorsing the importance of maximizing the potential use of feedback in-formation systems in combination with social normative feedback to promote eco-friendly driving within the corporate driving world. Ultimately, the second experi-ment supports previous findings and the capacity for corporations to reduce their CO2 footprint and fuel costs by improving their employees’ (i.e. car drivers) aware-ness of fuel consumption using a mobility information system to collect relevant data, which can then be appraised against ecological concepts, and customized feedback sent to each driver via a monthly email. Hence, this optimizes the attrib-utes of Green IS technologies whilst keeping the need for corporate investment to a minimal.

31 Published paper: Tulusan, J., Staake, T., Fleisch, E., Providing eco-driving feedback to corporate car driv-ers: what impact does a smartphone application have on their fuel efficiency, 14th ACM International Con-ference on Ubiquitous Computing (UbiComp), Pittsburgh, Pennsylvania, United States, September 2012.

90 Indirect Feedback by a Mobility Information System

V.2 State of the Art and Related Work

V.2.1 Green IS

Green IS proposes a medium through which environmental sustainability can be advocated, whether this is addressed by improving individuals’ awareness of eco-logical issues or modifying current practices and patterns of human behavior. A re-cent swift away from Green IT towards Green IS has been made; this movement involves reinforcing the business process that promotes ecological concepts (i.e. supporting a sustainable supply chain), and providing an evidence-based decision support system which facilitates organizational strategies to meet their CO2 reduc-tion goals (Thambusamy and Salam, 2010; Watson et al., 2010). Many studies eval-uate the conceptual and organization levels, but few appraise the application of Green IS in enhancing employees’ motivation to be progressively energy efficient (Jenkin et al., 2011; Loock, T Staake et al., 2011). Studies by Froehlich et al. (2009), Loock et al. (2011), Holmes (2007), and Mankoff et al. (2007), amongst others, generated promising results that suggest Green IS can have an influential role in modifying individuals’ behavior to reduce energy consumption.

Fogg (2002) defined the term ‘Captology’, which refers to information communica-tion technologies (ICT), such as websites and information systems combined with behavioral changing concepts. He classified persuasive technologies as “any inter-active computing system designed to change people’s attitudes or behaviors” (Fogg, 2002, p. 1). Froehlich et al. (2010, p.1999) expanded on this concept and characterized eco-feedback technologies as “technology that provides feedback on individual or group behaviors with a goal of reducing environmental impact”. Eco-feedback technologies have the potential to bridge the gap between individuals’ lack of environmental awareness and how their everyday behavior, such as driving to work, impacts the environment;; this is termed ‘environmental literacy gap’ (Froeh-lich et al., 2010). When considering energy saving in households, Geller et al. (1982) has shown that eco-feedback technologies have been effective strategies in reducing energy consumption.

Despite the substantial body of literature in areas, such as energy savings in house-holds using smart meters, and recently, online IS portals to motivate users to reduce energy consumption (Graml et al., 2011; Loock, T Staake, et al., 2011), limited re-search proclaims the impact of IS in improving driving habits. Hence, one can only


speculate the potential success IS can have on modifying human behavior within the energy saving context. Graham et al. (2011) adopted an approach to research in the field by applying concepts from social psychology with Green IS. This provided them with a strong theoretical platform from which to assess the effectiveness of an on-line intervention in encouraging 128 college students to reduce their daily car usage. Using a 2x2 experimental design they discovered that a combination of both feedback types, representing drivers’ monetary savings and reduction in CO2 emis-sions, had the greatest bearing in reducing daily car usage, followed by feedback regarding monetary savings alone. Environmental feedback in isolation had the least impact. Participants were required to report their car usage details and savings in miles, when using alternative modes of transport, on an online portal. This data en-abled the researcher to provide personalized ecological feedback to the participants using Green IS to stimulate behavioral modification.

V.2.2 Cognitive Intentions and Behavioral Change in Transportation

On reviewing the existing literature concerning behavioral change in transportation, attempts have been made to apply theories from social psychology research to in-fluence driving behavior, but with variable success. Social psychology theories, such as the Rational Choice Theory (Simon, 1955), the Theory of Reasoned Action (Ajzen and Fishbein, 1977), and the Theory of Planned Behavior (Ajzen, 1991), were applied to evaluate travel mode choices of individuals (Bamberg et al., 2003; Gardner and Abraham, 2010) or in analyzing interventions to reduce speeding (Goldenbeld et al., 2008; Stead et al., 2005). One critique of the Rational Choice Theory is that it assumes that individuals are able to comprehend their preferences and therefore, able to make an informed decision having considered all their op-tions. This assumption has been scorned by several authors, such as Baron (2008), Tversky and Kahneman (1986), Schwartz et al. (2002). For instance, Schwartz et al. (2002) argued “the assumption of complete information that characterizes rational choice theory is implausible” (Schwartz et al., 2002, p.1178).

The Theory of Planned Behavior (TPB) on the other hand, has been applied in sev-eral transportation research papers and is claimed to be “the most widely applied

model of modifiable cognitive antecedents of travel mode choice” (Gardner and Abraham, 2010, p. 832). TPB claims that behavior is triggered by intent and per-


ceived behavioral control (PBC). Intention is defined by attitudes (i.e. perceived outcomes of the behavior in question) and subjective norms (i.e. perceived social approval from others of the behavioral act performed) (Ajzen, 1991). Intention therefore motivates people to behave in a specific manner and is equivalent to a fa-vored behavioral act (Fujii and Gärling, 2003). Perceived behavioral control is de-fined as the professed ease with which a new behavior can be implemented (Ajzen, 1991), i.e. how easy it could be to drive sustainably. Newnam et al. (2004), Goldenbeld et al. (2008), and Stead et al. (2005) used the TPB theory to direct their interventions at reducing speeding, whereas Bamberg et al. (2003) and Forward (2004) utilized it to encourage drivers to travel by bus through means of providing them with prepaid bus tickets.

Additionally, one’s environmental beliefs reinforce attitude and intention. Environ-mental beliefs elicit moral obligations and consequently influence intentions, which “have been shown to contribute variance in travel-mode choice over and above TPB cognitions” (Gardner and Abraham, 2010, p. 833). Transportation research appraised the importance of reducing journey times and financial savings; however, with energy reduction being a priority, recent studies have acknowledged environ-mental beliefs too. For further information on these types of studies refer to Bam-berg and Schmidt (2003), Steg and Sievers (2000) and Polk (2003). It is evident that additional research investigating the relationship between sustainable driving and one’s environmental beliefs, personal moral norms, and attitudes is required. Be-sides cognitive intentions, motivational factors of employees, i.e. why employees are motivated to act more sustainable, needs to be further understood.

V.2.3 Descriptive Social Normative Feedback and Boomerang Effect32

Social norms, defined as conventional rules or behaviors imposed by society (Cialdini et al., 1991), have elicited a strong impression on both attitudes and be-havior, specifically in the context of social responsibility. This form of feedback was successfully embraced when evaluating energy conservation in households (Loock et al., 2011; Schultz et al., 2007), fuel consumption (Siero et al., 1989), re-

32 Published paper: Tulusan, J., Staake, T., Fleisch, E., Direct or indirect sensor enabled eco-driving feed-back: Which preference do corporate car drivers have?, Internet of Things 2012 – Third International Confer-ence on the Internet of Things (IoT 2012), Wuxi, P.R. China, October 2012.


ducing alcohol consumption among college students (Agostinelli et al., 1995), or towel reuse in hotels (Goldstein et al., 2008). All studies recognized that social normative feedback had a strong influence on participants’ behavior, and hence, could also play an influential role in motivating corporate car drivers.

Cialdini et al. (1991) and Reno et al. (1993) extended the concept of social norms in their focused theory of normative conduct by differentiating between descriptive norms (individual’s perception of what is commonly executed in a certain situation) and injunctive norms (individual’s perception of what is commonly approved or disapproved by society), and identifying which circumstances these norms affect behavior. The norm must be a focal point for the individual, i.e. command their at-tention when in a certain situation at that present time to influence their behavior; this is defined as normative focus (Cialdini et al., 1991).

Feedback provided by social norms does not always have a positive impact. Schultz et al. (2007) defined a crucial example of this called the ‘boomerang effect’. In their eight-week study of 290 households in California, household energy conservation was addressed by providing descriptive normative feedback. Each household’s av-erage energy consumption was presented together with social normative feedback indicating the average energy consumption of the neighborhood. While households with above-average figures decreased their consumption, some individuals with be-low-average energy consumption increased their consumption towards the average. This reflects that individuals with energy consumption below the average for their neighborhood adjust their consumption accordingly, resulting in a negative change.

The boomerang effect was offset by adding injunctive feedback, which showed a smiling face to support the below-average group as consuming less energy than the average households shown in the descriptive normative feedback (Schultz et al., 2007). After adding this injunctive message to the descriptive normative feedback, the energy consumption of the below-average households stayed at the desired low-er level. Evidently, social-normative feedback without injunctive feedback can pro-duce mixed outcomes; this creates a need for further investigation into the potential impact of descriptive social normative feedback on corporate car drivers’ fuel effi-ciency.


V.2.4 Layout of Rating Scales

Recent studies in the field of energy savings analyzed the impact of social norma-tive feedback on household energy consumption by, for example, providing descrip-tive normative feedback or injunctive and descriptive normative feedback (Loock et al., 2011; Schultz et al., 2007). Little attention (at least not in the scientific litera-ture) has been paid to the design and layout of feedback scales through which social normative feedback can be given. Some relevant insights can be found in earlier research in psychology and psychophysics from Freyd (1923) and Froberg and Kane (1989), who assessed the design and layout of rating scales. Their research focused on the statistical validity and reliability of rating scales in surveys (Svensson, 2000). The definitions and layouts of these scales provide an overview how feedback scales, in combination with social normative feedback, could be de-signed and tested.

Verbal Descriptor Scales (VDS) display items in “ordered lists of verbally described and/ or numerically labeled response categories” (Belz and Kow, 2011, p. 231). The number of descriptive items correspond to the values of the ordered list and can vary from 2 to 11 items (Svensson, 2000). Ordered lists include distinct rating cate-gories in either verbal (e.g., Poor, Fair, Good, etc.) or numerical values. An example of a VDS scale with a verbal ordered list with five items is shown in Figure 41:

Poor Fair Good Very Good Excellent

Figure 41: Verbal Descriptor Scale (VDS)

Visual Analogue Scales (VAS) show the responses on a continuous line, typically on a horizontal line but in some cases on a vertical one (Scott and Huskisson, 1977). The endpoints on the line represent the extreme values of the variables; for exam-ple, ‘poor’ and ‘excellent’ (Belz and Kow, 2011). These formats are used to provide feedback via a scale ranging from 0 to 100 in a continuous form without distinct rating categories; other ranges (e.g., from 0 to 200) have also been applied (Aitken, 1969). Figure 42shows an example of a VAS scale:


Figure 42: Visual Analogue Scale (VAS)

The combination of VDS and VAS scales is called a Graphic Rating Scale (GRS) and combines the verbal or numerical items with a continuous line. The GRS dis-plays the categorical values, either verbally (e.g., from poor to excellent) or in units ranging from 0 to 100 on a continuous line, with pre-defined categories marked, as shown in Figure 43 and Figure 44.

Figure 43: Verbal Graphic Rating Scale (GRS)

Figure 44: Verbal and Numerical Graphic Rating Scale (GRS)

Comparative studies between VDS and continuous VAS or GRS scales have report-ed conflicting outcomes, with no firm evidence of which type of scale provides a better outcome in terms of validity and reliability (Svensson, 2000). Svensson (2000) found that the VDS and GRS have a stronger impact than the VAS format, as the VAS format shows a continuous line and does not provide any categorical or pre-defined line items. However, Aitken (1969) describes the VAS scale with more items (e.g., any value between 0 and 100) as more sensitive than discrete scales. This has been disproved by Svensson (2000), who showed that an increased number of possible responses does not lead to a more valid and reliable outcome.

V.2.5 Layout of Social Normative Feedback Scales

The discussion in the previous chapter illustrates the lack of clarity in the design of scales, adding to the gap in research on how feedback scales should be designed to provide social normative feedback. Considering the types of scale used in the ener-gy-savings research domain, social normative feedback could be presented in vari-ous formats. Common formats include scales in which the average value for a popu-lation is displayed in comparison to an individual’s own consumption (social nor-mative feedback). For instance, in the field of energy savings, low, average, and


high values of electricity consumption in a neighborhood are used as benchmarks to inform households in the same area with similar demographic structures about their electricity consumption in comparison to the benchmark values (Baeriswyl et al., 2011). This type of scale is similar to a VAS format, and Figure 45 illustrates these factors via a linear scale with categorical benchmark items for low, average, and high energy consumption and showing the consumption of the individual household.

Figure 45: Normative Feedback to improve Energy Consumption (Baeriswyl et al., 2011)

Another example of a standardized rating scale, which includes elements of a VDS, is defined in the EU directive 2002/91/EC, as shown in Figure 46:

Figure 46: Energy Efficiency Rating Scale (EU directive 2002/91/EC)

The CO2 emissions energy efficiency rating scale is used to provide an index of how much CO2 a household is currently emitting and compare this value via an or-dered categorical list, shown as a vertically oriented GRS (from category ‘A’, very low CO2 emissions, to category ‘G’, very high CO2 emissions). Each category is further defined in numerical ranges (i.e. 100-92, 91-81, etc.). The CO2 emissions efficiency is also color coded, with the lowest CO2 emission category dark green and the highest dark red. Finally, the feedback scale provides an index for potential


reductions from the current CO2 emissions. A scale with the same layout but differ-ent color coding, from light blue to dark blue, was defined by the EU Commission for rating electricity consumption of households.

These examples demonstrate that rating scales in energy-reduction research are sim-ilar to the rating scales from psychology and psychophysics that were evaluated by Freyd (1923) and Froberg and Kane (1989). No study in this research domain eval-uates VAS and GRS scales in relation to social normative feedback and energy re-duction. Research from Cialdini et al. (1991), Loock et al. (2011), and Schultz et al. (2007) supports the importance of providing social normative feedback in combina-tion with injunctive feedback for improving energy reduction. The latter helped to prevent the energy consumption of below-average households from rising above the average once the average consumption of similar households was known (Schultz et al., 2007). In terms of the design of scales, VAS have many items on a continuous linear scale, whereas VDS and GRS have pre-defined categorical/ discrete items. Whether a certain type of scale has a stronger influence on environmental behavior needs to be further evaluated.

For the purpose of this study, two different feedback scales that follow the GRS and VAS design principles and provide descriptive normative feedback were applied. The feedback displays the driver’s own fuel consumption in comparison to the av-erage fuel consumption of other drivers from the same company. The GRS (Figure 47, left) provides feedback in discrete/ categorical areas (defined in three fuel con-sumption areas: low, middle, and high), whereas the VAS (on the right of the fig-ure) shows the feedback in a continuous format, with the lowest fuel consumption on the left endpoint and the highest consumption on the right endpoint. Color cod-ing emphasizes the fuel consumption on both types of scales, with dark green denot-ing low fuel consumption and dark red denoting high fuel consumption.

Figure 47: Categorical and Continual Feedback Scale Formats


V.3 Research Questions and Hypothesis

Studies have identified an improvement in participants’ fuel efficiency when mone-tary rewards were awarded, i.e. financial savings as a measure of driving infrequent-ly or reimbursement when another transportation mode was used (Deslauriers and Everett, 1977; Graham et al., 2011; Thogersen and Moller, 2008). Further clarity is required to ascertain how feedback interventions that are not driven by financial rewards can motivate corporate car drivers to adopt sustainable driving habits, and hence, improve their company’s carbon footprint and reduces fuel costs. Reflecting on the literature currently available, it is apparent that there is limited research, be-sides truck or van drivers, that investigates this client group. Furthermore, the role Green IS reinforced by social normative feedback in promoting eco-friendly driving habits is a relevant gap in research that will be investigated.

This study focuses on feedback delivered by a mobility IS (MIS) to improve the fuel efficiency of corporate car drivers. As employees working for this corporation had regular access to their emails and a corporate mobile phone, it was feasible to extend the parameters of the investigation from the first field test to maximize the capacity of the existing MIS, which stored relevant data from each driver, by send-ing customized feedback via email. Sending feedback by email enabled the re-searcher to access a larger sample of the research corporations driving fleet, more than 50% of their drivers. Concepts of Green IS were employed to focus on the technical aspect of delivering feedback to participants, and findings from environ-mental social psychology used to appraise how the intervention and which type of feedback scale best influenced drivers. The fourth sub-question from the thesis is stated as follows:


To assist in answering this sub-question, three hypotheses were defined:

The first hypothesis evaluates the question, how does providing eco-driving tech-niques/ tips by email, without any eco-driving training or use of an eco-driving smartphone application, effect driving behavior (see Table 13).


Table 13: Illustration of Hypothesis 1

Corporate Car Drivers’ FE

Eco-driving techniques feedback by email ↑ ↓ =

The first hypothesis is defined using µ1 to represent fuel efficiency without any feedback and µ2 to denote receiving eco-driving tips by email:

H1: Providing details about ecological driving techniques by email improves corpo-rate car drivers’ fuel efficiency.

(H1) H0 = µ1 ≤ µ2; H1 = µ1 > µ2.

Having acquired a better understanding of how eco-driving tips by email impacts behavior, it was also important to understand to what extent individual feedback about driver’s fuel consumption and social normative feedback influenced behav-ioral change. Findings from Graham et al. (2011) have shown that increasing the environmental awareness of individuals by providing feedback about their reduction in CO2 emissions reduced their car use, and therefore their overall fuel consump-tion. When considering the social psychological theory perspective, it was evident that the Theory of Planned Behavior by Ajzen (1991) had been extensively applied to review travel mode choice, i.e. taking the bus instead of the car. For this study, the application of the Feedback Intervention Theory (FIT) by Kluger & DeNisi (1996) introduced in Chapter II provided a foundation from which to reflect on how feedback interventions influence individuals’ task performance (i.e. driving style). The feedback intervention applied descriptive social normative feedback by presenting drivers’ fuel efficiency rating with a comparison against the fuel effi-ciency rating of colleagues from the same company (see Table 14).


Corporate Car Drivers’ FE

Social Normative Feedback by email ↑ ↓ =


No current findings exist that reflect on how social normative feedback format in-fluence corporate car drivers’ fuel efficiency, which raises the second hypothesis by µ1 representing the fuel efficiency without and µ2 with receiving social normative feedback by email.

H2: Providing social normative feedback about driver’s fuel consumption by email improves corporate car drivers’ fuel efficiency even when monetary incentives are not relevant.

(H2) H0 = µ1 ≤ µ2; H1 = µ1 > µ2.

It is evident that the setting and design of social-normative feedback encompasses several attributes and can generate mixed outcomes. One example, known as the boomerang effect, was tested by Schultz et al. (2007) and Loock et al. (2011). Freyd (1923) and Froberg and Kane (1989) evaluated the reliability and validity of cate-gorical versus continuous scale formats. Svensson (2000) identified that discrete/ categorical scales provide a better outcome compared to continuous scales. To this end, the third hypothesis aimed to appraise what difference the type of feedback scale (i.e. discrete/ categorical or continuous) has on participants’ response to the social normative feedback they received (see Table 15).


Scale Format FE rating Corporate Car Drivers’ FE

Low ↑ ↓ = Discrete Medium ↑ ↓ =

High ↑ ↓ =

1 ↑ ↓ = Continual to ↑ ↓ =

12 ↑ ↓ =

The third hypothesis is defined with µ1 representing the fuel efficiency shown on a discrete and µ2 representing a continual social normative feedback scale.


H3: Descriptive social normative feedback shown on a discrete/ categorical scale has a stronger impact on corporate car drivers’ fuel efficiency versus that shown on a continual scale.

(H3) H0 = µ1 ≥ µ2; H1 = µ1 < µ2.

V.4 Research Design and Research Methodology

The opt-out experiment (see Figure 48) was announced in August 2011 via email by management to 450 corporate car and car allowance drivers. Two hundred and forty participants from this sample were then randomly selected to take part in the exper-iment and assigned to three groups: Treatment Group 1 (TG1), Treatment Group 2 (TG2), and Treatment Group 3 (TG3).

Figure 48: Experimental Design

Evaluation of participants’ fuel efficiency was based on true fuel consumption fig-ures from sales data provided by the petrol credit cards of each corporate car driver. These data were transferred to the company’s MIS. Data included: total number of kilometers driven, liters filled, total cost for each tank filling, petrol type, and loca-tion of petrol station. The raw data were then used by the MIS to calculate the fuel


efficiencies of each driver after each tank filling by dividing the total number of liters of petrol filled by total number of kilometers driven since the last tank filling. To establish a baseline for each group, the average fuel efficiency prior to the inter-vention (01/01/2011 to 21/09/2011) was calculated, this allowed for a comparison to be made within subjects from each group (TG1, TG2, and TG3). The TG1 re-ceived a monthly email with only tips how to drive more ecological. Participants in the TG2 and TG3 received a monthly email (09/22/2011, 24/10/2011, 24/11/2011) detailing their: individual fuel efficiency data, difference from the previous month, average kilometers driven during the month, CO2 emissions, and eco-driving tips regarding how to improve their fuel efficiency (see Figure 49).

Figure 49: Monthly Feedback Email with Discrete or Continual Feedback

Two scales, discrete and continual, were used for to reflect how ‘good’ or ‘bad’ par-ticipant’s fuel efficiency was in comparison to their peers (descriptive social norma-


tive feedback). The discrete scale presented fuel efficiency in three categories: low, middle, and high (see Figure 49 left scale). The continual scale (see Figure 49 right scale) illustrated this information on a low to high continuum, which was divided into 12 sections and indicated the driver’s fuel efficiency with an arrow. The aver-age fuel efficiency percentage of the total sample (n = 240) prior to the initiation of the experiment was 13.57% above the stated fuel efficiency from the car manufac-turers, a value higher then claimed by the car manufacturer confirming their fuel efficiency figures were somewhat optimistic. This average fuel efficiency percent-age value from the total sample size was used to define the mean value as the base-line for each fuel efficiency rating scale (see Table 16).

Table 16: Fuel Efficiency Rating and Intervals

Type of Scale FE Rating Fuel Efficiency Intervals

in %

Discrete Low < 7.09% Middle Between 7.09% and 18.53% High > 18.53%

Continual

1 < -7.94% 2 -7.94% to -3.00% 3 -3.01% to 3.52% 4 3.53% to 7.08% 5 7.09% to 10.68% 6 10.69% to 13.91% 7 13.92% to 15.83% 8 15.84% to 18.52% 9 18.53% to 20.68% 10 20.69% to 23.54% 11 23.55% to 28.47% 12 > 28.47%

The individual fuel efficiency rating on the scales were based on the difference be-tween the actual fuel efficiency of the driver for the month and fuel efficiency of the car type as stated by the manufacturer. For instance, if a driver had an average monthly fuel efficiency of 6.7 liters per 100 kilometers (l/100 km) and the fuel effi-


ciency stated by the car manufacturer was 5.6 l/100 km, a difference of 1.1 l/100 km or 19.72% is evident. This percentage value of 19.72% was compared with the av-erage fuel efficiency value of 13.57% and the driver’s position on the feedback scale assigned. Furthermore, comparison between participants’ changes in monthly fuel efficiency and different car types with varying engine sizes, and against the rest of the total sample was possible.

V.5 Data Evaluation and Findings

V.5.1 Findings for Hypotheses H1, H2, and H3

Of the 240 participants selected for the experiment, 20 exercised their right to opt-out leaving in the TG1 = 72, TG2 = 75, and TG3 = 73. An additional three partici-pants were removed from the sample, one from the TG1 and two from TG3, as they were identified as outliers using the Grubb’s test due to very low or excessively high changes in their overall fuel efficiency. The paired t-test statistical method was used for each group to validate the comparison before and after the treatment phase.

(H1) The results of the paired t-test for TG1 (N = 71) indicated that the average fuel efficiency (M = 7.852, SE = 0.160) marginally increased during the treatment phase (M = 7.861, SE = 0.160). A difference of +0.009 l/100 km or +0.12% was calculated and found to be statistically insignificant (p > 0.05 with t (0.204) = 0.009, p = .839 (see Table 17).

Table 17: Paired Samples T-Test TG1

Paired Differences

Mean Std. Devia-

tion Std. Error

Mean

95% Confidence In-terval of Difference

Lower Upper

avg_fe_y – avg_fe_trtmt_t -.009 .374 0.444 -.098 .079


avg_fe_y - avg_fe_trtmt_t -.204 70 .839


As there was no statistically significant difference between the baseline and the treatment, we can conclude that providing feedback about ecological driving tech-niques by email does not change corporate car drivers’ fuel efficiency, and the Ho cannot be rejected.

(H2) In order to address the second hypothesis a paired t-test was applied to TG2 and TG3. Statistical significance (p < 0.05) was confirmed for both groups; the av-erage fuel efficiency for TG2 (N = 75) improved by -0.093 l/100 km or -1.24% with the mean value of M = 7.501, SE = .174 before the treatment and M = 7.408, SE = .186 during the treatment. Results found to be statistically significant (p < 0.05 with t (2.068) = -0.093, p = .042 (see Table 18).


Paired Differences

Mean Std. Devia-

tion Std. Error

Mean


Lower Upper

avg_fe_y – avg_fe_trtmt_t .093 .388 .045 .004 .182


avg_fe_y - avg_fe_trtmt_t 2.068 74 .042

The average fuel efficiency for TG3 (N = 70) improved by -0.134 l/100 km or -1.69% with the mean value of M = 7.907, SE = .166 before the treatment and M = 7.773, SE = .176 during the treatment. Results found to be statistically significant (p < 0.05 with t (2.214) = -0.134, p = .030 (see Table 19).



Paired Differences

Mean Std. Devia-

tion Std. Error

Mean


Lower Upper

avg_fe_y – avg_fe_trtmt_t .134 .505 .060 .013 .254


avg_fe_y - avg_fe_trtmt_t 2.214 69 .030

These outcomes indicate that provision of social normative feedback related to the participant’s fuel consumption by email improves corporate car drivers’ fuel effi-ciency even when monetary incentives are not relevant. Therefore, the Ho can be rejected in favor of the H1.

(H3) An independent t-test was used to compare the difference between the discrete scale (TG2) and continual scale (TG3). The result showed that, on average, the con-tinual feedback scale had a greater impact on participants’ fuel efficiency with a difference of 0.73%; however, this was statistically insignificant (p > 0.05 with t (144) = 0.73, p = .579). Therefore, it cannot be concluded that descriptive social normative feedback shown on a discrete scale has a stronger impact than a continual scale on corporate car drivers’ fuel efficiency, and the H0 cannot be rejected.

V.6 Discussion

The application of social normative feedback produced promising and statistically significant results when a within group comparison was made in groups TG2 and TG3. In both treatment groups an improvement in fuel efficiency was apparent; in TG2 -1.24% and in TG3 -1.69%. However, simply providing feedback about tech-niques endorsing sustainable driving habits by email does not motivate a change in participants’ task performance, i.e. driving style, to improve fuel efficiency. This assumption was also stated by Kluger & DeNisi (1996) and extended by Alder


(2007) in the application of the FIT, which identified that feedback interventions should be postulated in the correct context to trigger changes in task performance.

When comparing the discrete and continual feedback scales, no statistical signifi-cant difference between TG2 and TG3 was discovered. Therefore, we cannot con-clude that descriptive social normative feedback illustrated on a discrete scale gen-erates a stronger or weaker impact in improving fuel efficiency when compared to a continual scale. This echoes current literature findings that explain the design, relia-bility or validity of feedback scales, but do not disclose what impact the different scales have on an environmental behavior. Nevertheless, we can assume that indi-rect feedback by email and social normative feedback provided in the correct con-text can have an influence on driving behavior, which further confirms findings by Siero et al. (1989). In their study one short-coming was the application of direct and indirect feedback types; hence, they were not able to clearly conclude which of the feedback types had a stronger influence on the overall improvement in fuel efficien-cy (Siero et al., 1989).

V.7 Conclusion

To summarize, an opt-out field experiment was conducted over three months using drivers from one corporation. Access to participants’ MIS enabled rich data relevant to fuel consumption to be collected. Research within the fields of feedback technol-ogies and social psychology appraised how the potential use of Green IS can be maximized by using emails with social normative feedback. Feedback IS coupled with social psychological concepts have a valuable contribution to make in research approaches to modifying drivers’ behavior to stimulate improved fuel efficiency (Froehlich et al. 2010; Jenkin et al. 2011b; Meschtscherjakov et al. 2009).

Findings of the study imply that extended use of Green IS by sending emails that provide fuel consumption figures and social normative feedback had an influence in improving corporate car drivers’ fuel efficiency without any financial rewards. This affords support to recent literature that claims Green IS can stimulate behavioral modification to reduce energy consumption (Graml et al., 2010; Loock et al., 2011) and so bridging the ‘environmental literacy gap’ by enhancing individuals’ envi-ronmental awareness (Fogg, 2002; Froehlich et al., 2010). It is important to


acknowledge that providing social normative feedback on a discrete/ categorical or continual scale format did not show a statistically significant difference and that further research is required. In practice, organizations with large car fleets looking for cost effective strategies to reduce their CO2 emissions and petrol costs, Green IS supported by social normative feedback proposes an attractive solution.

V.8 Limitations and Future Research

As all groups received some level of feedback, with TG1 receiving the minimal of driving tips to promote sustainable driving, limited control over seasonal effects was possible. As weather conditions vary during the course of a year, it is honest to recognize that this can have an impact on fuel efficiency. A solution to address this would have been to assign a control group that did not receive any feedback, how-ever, the average fuel efficiency of TG1 prior to the field test was calculated from data collected over a duration of time (January to September) and only marginal fluctuations in fuel efficiency were detected in comparison to the treatment time (+0.12% per l/100 km). This suggests that seasonal effects had the slightest impact on fuel efficiency in 2011. During the period of the field test no extreme weather changes were apparent and so it is fair to claim that weather conditions had a mini-mal impact on fuel efficiency data collected for this study.

There were opportunities for further data analysis. Changes in fuel efficiency could have been evaluated on a monthly basis instead of accumulatively at the end of the field test. This would have provided information about how participants’ responded to feedback over time; i.e. was there a greater improvement in their fuel efficiency after the first email compared to the third email.

Recommendations to Increase Corporate Car Drivers’ Intrinsic Motivations 109

VI Recommendations to Increase Corp. Car Drivers’

Intrinsic Motivations33

VI.1 Overview and Research Questions

The eco-driving smartphone application provided direct and indirect feedback to the driver when driving and resulted in an average improvement in fuel efficiency of 3.23% (see Chapter IV). The MIS provided indirect feedback through a monthly fuel consumption email, which led to a fuel efficiency improvement of up to 1.69% (see Chapter V). In addition to analyzing whether feedback technologies have a pos-itive impact on corporate car drivers’ fuel efficiency, behavioral and motivational aspects that help corporate car drivers improve their fuel efficiency must be ad-dressed. The case study company currently has 19,100 corporate cars globally and is interested in the behavioral aspects of motivating employees without giving fi-nancial incentives as it seeks to encourage eco-driving via feedback technologies.

The main goal of this chapter is to derive recommendations by evaluating corporate car drivers’ motivations for fuel savings and opinions about implementing a fuel savings program supported by feedback technologies within their organization. Specifically, extrinsic and intrinsic motivations are evaluated to answer the follow-ing research question:


VI.2 Extrinsic and Intrinsic Motivation

Understanding what motivates an individual is a complex task. Management must understand which factors, intrinsic or extrinsic, motivate their employees (Beswick, 2007). Extrinsic motivation involves external rewards (i.e. financial incentives or extra days off) and/ or recognition (i.e. participating in important meetings) (Zahorsky, 2010). The disadvantage of this form of motivation is that employees focus on the reward, not on the action itself (Beswick, 2007). Intrinsic motivation 33 Published paper: Tulusan, J., Steggers, H., Staake, T., Fleisch, E., Supporting eco-driving with eco-feedback technologies: Recommendations targeted at improving corporate car drivers’ intrinsic motivation to drive more sustainable, Energy Informatics 2012 (EI 2012), Atlanta, Georgia, United States, October 2012.

110 Recommendations to Increase Corporate Car Drivers’ Intrinsic Motivations

stems from within the person, and satisfaction derives from completing the task it-self (Zahorsky, 2010), contributing to the driver’s self-esteem. When applied cor-rectly, extrinsic motivation can enhance intrinsic motivation. De Young (1985) sug-gested that intrinsic motivation encourages environmentally responsible behavior; individuals do not wait for rewards to be environmentally responsible but rather “seem to derive personal satisfaction from the very activities that others so often try to externally reinforce” (De Young, 1985, p. 289).

Studies on home energy savings discovered that people were more motivated to save energy and be environmentally friendly when they received personalized in-formation about their energy consumption (Shipworth, 2000). Users need to under-stand the information (Shipworth, 2000) and feel that they have freedom of choice and control (Rotter, 1966). It is necessary to support these individuals in developing their own intrinsic motivation (e.g., to drive sustainably), as recommended by De Young (1985). Siero et al. (1989) concluded that although it is important for man-agement to set feasible goals and provide information when making task assign-ments, they must also capture employees’ attention by reinforcing that energy-saving driving behavior is an integral part of their job. However, caution must be exercised to avoid external control (Amabile, 1993): the driver must feel a sense of autonomy, and information must be clear enough that they can understand and uti-lize it effectively. Implementing an organizational eco-driving initiative supported by eco-feedback technologies requires management to understand what motivates their employees and to meet their needs.

VI.3 Research Design and Research Methodology

In total, fifteen semi-structured post-experiment interviews (see Appendix G), each lasting 20 - 25 minutes, were conducted with field test participants. Before the in-terview, all interviewees filled out the post-experiment online survey (see Appendix F). This helped to obtain a better understanding of the interviewees before the inter-views took place. The semi-structured interviews were used to give interviewees a forum for expressing their experiences and opinions in a structured manner. Ques-tions for the semi-structured interviews were selected based on the interviewees’ answers in the post-experiment survey, completed directly after the field tests. This enabled the researcher to highlight specific post-experiment survey answers and


open questions that were in need of clarification during the interview. Together with obtaining feedback about the feedback technology, the post-experiment interviews’ purpose was to evaluate opinions and derive conceptual recommendations to be considered when implementing a fuel efficiency improvement or CO2 reduction initiative within an organization.

Structured analysis was used to evaluate the unstructured data collected in the fol-low-up interviews (see Figure 50). The process of structured analysis is defined as, “microanalysis, which consists of analyzing data word-by word” while determining “the meaning found in words or groups of words” (Strauss and Corbin, 1998, pp. 65–68). The interview data was transcribed and classified into topics to unveil emergent factors; this is known as the coding process. The subsequent process, axial coding, grouped similar factors into categories. This enabled a systematic analysis and constant comparison of factors and categories. Finally, factors were ranked ac-cording to the frequency of answers and assigned to their category. This process enabled a structured approach to interpreting the answers.

Figure 50: Interview Data Analysis


After evaluating and categorizing the interview data, the FIT (Kluger and DeNisi, 1996) was applied to obtain a better understanding how feedback interventions us-ing eco-feedback technologies enable corporate car drivers’ to modify behavior. In detail, five basic arguments drawn from the FIT and the literature on extrinsic and intrinsic motivation were compared with the categorized factors from the interviews and answers from the post-experiment survey.

VI.4 Data Evaluation and Findings

VI.4.1 Opinions and Concerns of Drivers

The structured analysis process of the interviews revealed several categories which are defined from the ranked factors. Categories with most ranked factors are sum-marized in Table 20 (all answers are shown in Appendix G) and further defined in two levels: Driver and Management Level. The Driver Level focuses on the opin-ions of the drivers about eco-feedback technologies, whereas the Management Lev-el on specific aspects a management should consider by rolling out an eco-driving concept supported with feedback technologies.

Table 20: Categories of Structured Analysis

Questions 1. Driver Level 2. Management Level 1. What do you think was the greatest bene-fit of the application?

Awareness: - Less about feedback, more about using the application (15) - Raise awareness (13) - Sensitize (11)

2. Besides the feed-back you have re-ceived, which addi-tional feedback criteri-on are relevant to you?

Feedback: - Real-time feedback (15) - Should not distract from driving (12) - Transparent feedback (12)


3. Do you think a monthly e-mail with a comparison of your FE against another driver would be useful? If yes, what kind of comparison would you prefer?

- Yes (15) Importance of valid comparison: - Same car model (15) - Same average total distance driven annually (14) - Transparent comparison (10)

4. Which other aspects would be relevant for you to use the applica-tion regularly in your corporate car?

Time importance: - Everyday on the road (15) - Important to be home fast (15) - Many customer visits (10) Incentives: - High influence of financial incentives (12)

Provide information: - Provide personalized information (14) - Expose fuel costs (13) Rewards: - Realistic/ fair rewards (12)

5. How should the management roll out a fuel improvement/ CO2 reduction initiative within the organiza-tion? i) What steps do you feel are crucial for im-plementation? ii) Where do you see problems?

Punishment/ control vs. awareness: - No punishment (14) - Unfair to punish drivers who need their car for daily business (13) - No control from management (12) Goal setting : - Goal setting necessary (12) - Realistic goals related to corporate car drivers (11)

All participants agreed that using the eco-driving smartphone application raised their awareness, encouraging them to think about actual driving habits in compari-son to a more sustainable driving style (question one). The application’s abilities to raise awareness and sensitize the driver to their personal driving style were consid-ered its biggest benefits (stated 13 and 11 of 15 times, respectively). Question two, which asked about which feedback criteria were important to the drivers, revealed that real-time and transparent feedback that did not distract from driving were fa-vored.

The feedback provided by a monthly email (question three) was also considered very important (stated 15 times). Respondents felt feedback should be given in the form of a social comparison with other drivers; participants specified that the com-


parison had to be valid, explicit, and transparent (stated 10 times). Drivers identified the most suitable social comparison at one between drivers who drove the same ve-hicle type model (stated 15 times) and approximately the same average distance per year (stated 14 times).

Greater focus was placed on cost savings (petrol-cost reduction or reduced deterio-ration of the vehicle) and time savings than on environmental benefits (question four). Employees working in sales or consulting are required to visit several cus-tomers per day to generate leads or offer consulting services. Drivers are thus una-ble to completely take control of their fuel efficiency, as this is partly dependent on where their customers are located and how many customers they have to visit on a given day. Depending on customers’ geographical locations, time pressure, and traf-fic conditions, employees adapted a different driving style regardless of the impact this had on their fuel efficiency. Participants stated that this was the main reason why management should restrain from penalizing drivers who were unable to meet fuel efficiency goals. It is necessary to consider the sales pressure and sales targets. Fifteen statements were made highlighting that when driving long distances daily, it is more important for the employee to reach home earlier than to improve their fuel efficiency.

In terms of the management’s role, they should provide each driver with personalized feedback that recognizes drivers’ progress and provides individual advice about how they can continue to improve. Drivers also favored knowing their month-ly total fuel costs and having realistic and fair rewards defined for improving their fuel efficiency. Providing the total sum of monthly fuel costs can also act as an indi-rect incentive for employees, especially when the company reimburses petrol costs for private journeys. This is the case for all the corporate car drivers employed by the case study company in Germany. The interviews revealed that the participants had a positive attitude towards implementing eco-driving but had concerns about management’s response to drivers who did not improve their fuel efficiency (ques-tion five). It is important that the management not punish or control the drivers, es-pecially if they need their car daily to perform their job. Aside from these concerns, participants had a positive attitude towards goal setting by management. They stated that setting realistic goals was necessary to enhance their motivation to drive more


sustainably (stated 12 times). In contrast, only three drivers felt that management intervening by defining fuel improvement goals was not desirable.

These answers provide valid insights into corporate car drivers’ opinions and valua-ble information for management. In order to better verify how feedback could trig-ger employees’ motivations, the five basic arguments of the FIT (Kluger & DeNisi, 1996) were applied and compared with the respondents’ answers in the next section.

VI.4.2 Analysis of Findings and Discussion

The FIT of Kluger & DeNisi (1996), introduced in Chapter II, is composed of five basic arguments that can be applied to explain behavioral modification using com-puter-mediated feedback interventions, e.g., through an eco-driving smartphone ap-plication or an MIS. The results of this research will be discussed in relation to the-se five basic arguments as well as employees’ extrinsic and intrinsic motivations, as defined in Chapter VI.2. These results are based on findings from the post-experiment survey and the semi-structured interviews that were conducted, coded, and evaluated after the field tests. A high-level summary of key findings is given in Table 21.

Table 21: Summary of Feedback Intervention Theory Arguments

Arguments Findings

Argument 1

- Existing driving behavior standards - Feedback by eco-feedback technologies compared with driving

standards - Enablement to adapt existing driving behavior - Set realistic goals without punishment or control from management

Argument 2 - Different personal goals and standards are defined hierarchically - Saving time is most important for corporate drivers - Sustainable driving is a subordinate factor

Argument 3

- Feedback-standard gap: Disparity between driving behavior and eco-driving behavior

- Smartphone application and MIS directs attention towards disparity - Increase awareness of eco-driving practices

Argument 4 - Driving eco-friendly is at the moderate level of drivers’ hierarchy - Set realistic goals that are attainable - Drivers must feel competent in order to be intrinsically motivated

Argument 5 - Feedback intervention changes drivers’ locus of attention - Deviation away from saving time towards driving more sustainably - Positive impact of more ecological driving behavior


Argument 1: Behavior is regulated by comparisons of feedback to goals or stand-ards. The first argument states that individuals have their own standards and might be reluctant to change their behavior without having external goals and receiving feed-back about their performance. This underpins the importance of influencing corpo-rate car drivers’ driving behavior through feedback interventions. The participants’ extensive driving experience has caused them to develop driving habits, which then constitute their standards. The feedback participants receive is generated by the feedback technology, which compares their driving behavior to relevant eco-driving practices. This helps raise their awareness by compelling them to consider accepting the feedback given and enables them to choose to adapt their driving behavior. Im-proved driver awareness was attained amongst the corporate car drivers. Together with improving their fuel efficiency, 79% discussed their driving behavior with friends and/ or colleagues during the treatment phase, as revealed in the post-experiment online survey.

The semi-structured interviews conducted with participants revealed that they were comfortable with management interventions in the form of goal setting, as long as there were no associated punishments. These findings align with the FIT, which suggests that employees can only be intrinsically motivated if they feel a degree of autonomy and competence in their task (Deci and Ryan, 1985). Employees should not feel controlled by external motivators (Amabile, 1993), and management must set realistic fuel efficiency improvement goals and provide constructive feedback. The study by Siero et al. (1989) supports these findings, concluding that it is im-portant for management to provide task-orientated assignments and sufficient con-trol (e.g., to set feasible goals) in order to successfully implement an eco-driving program.

Argument 2: Goals or standards are organized hierarchically. Individuals set their own goals and standards according to their intrinsic motivation in the context of extrinsic goals or standards. In addition to setting these goals, indi-viduals order them hierarchically. Corporate car drivers’ personal driving goals var-ied; for example, some aimed to save as much time as possible while driving, while others aimed to drive safely or to drive more efficiently (see Table 20). The goal given in the field test was to drive more sustainably, but consideration of partici-


pants’ personal goal hierarchies made it apparent that time took precedence for the-se drivers. One interviewee noted, “If you are on the road a lot, the limit is reached

very fast and time becomes the most important factor” (Corporate car driver 3, 2012). These different goals and standards need to be understood in order to define realistic goals that meet employees’ expectations.

Argument 3: Attention is limited, and therefore only feedback-standard gaps that receive attention actively contribute to behavior regulation. The feedback-standard gap in this research is the disparity between eco-driving hab-its and corporate car drivers’ driving behavior. For instance, the eco-driving smartphone application offered eco-driving feedback during driving. This interven-tion was an important factor in improving fuel efficiency, indicating the need to ac-tively direct drivers’ attention to the difference between their driving habits and eco-driving practices. Hence, increasing the driver’s awareness through eco-feedback technologies could shrink the gap between the actual driving style and a more ecological style. As emphasized by De Young (1985) in the household energy context, raising residents’ awareness of energy consumption by providing them with these details can stimulate their intrinsic motivation.

Drivers also expressed that the eco-driving smartphone application or monthly feedback email raised their awareness of their own driving behavior; they advised that, when considering future guidelines, it is important to increase awareness of the practices that support eco-friendly driving. The following preferences for receiving such feedback were highlighted in responses from the post-experiment survey: via visualization only (6 out of 7) and from a weekly email (4.88 out of 7). Providing feedback via e-mail may be a promising way to supplement and enhance the feed-back provided by an MIS.

Argument 4: Attention is normally directed to a moderate level of the hierarchy. The case study company’s corporate car drivers spend most of their time on the road, driving hundreds of kilometers almost every day. Ninety-six percent of partic-ipants used their car daily or 5-6 times per week, and 92% drove on average more than 24,000 kilometers per year; this conforms to the general criteria for corporate car drivers. They also had the added pressure of time, with a need to be punctual for customer meetings. This explains why time was the most important factor for these corporate car drivers, as highlighted in the semi-structured interviews (see Table


20). Participants had a strong desire to reach customers on time and to reach home as soon as possible at the end of their workday. Understandably, saving time is allo-cated the top spot in drivers’ goal hierarchy. The FIT states that attention is not al-ways actively directed to the highest goal but rather at a moderate level (Kluger and DeNisi, 1996b). Driving more eco-friendly was not the highest goal; it appeared at a moderate level of the hierarchy, and so attention could have been redirected towards reaching this goal. The post-experiment survey indicated that participants had mod-erate environmental attitudes (5.01 out of 7 on average), supporting the argument that driving more eco-friendly was at a moderate level of their hierarchy.

With respect to management interventions, Argument 4 provides the criteria for set-ting realistic goals. For instance, an improvement of 10% within two months would have been unrealistic as studies have shown that an improvement of up to 15% is only possible by attending eco-driving training (Onoda, 2009). Setting an unrealistic goal could have lowered drivers’ commitment to achieving it;; they instead may have placed it at the bottom of their hierarchy, where it would not have received any attention. This could have also lessened participants’ feelings of competence when they were unable to reach the goal (Amabile, 1993; Deci and Ryan, 1985). Conse-quently, the extrinsic motivation behind setting the goal is diminished, and the po-tential to positively influence the drivers’ intrinsic motivation is lost. This supports the notion that it is highly important to make an employee feel competent by setting realistic standards that intrinsically motivate them (Amabile, 1993).

Argument 5: FIs change the locus of attention and therefore affect behavior. Feedback provided by the eco-driving smartphone application and the MIS are CPM-mediated feedback interventions, which offer eco-driving feedback or fuel consumption details according to the car type. As explained, feedback directed driv-ers’ attention away from their preoccupation with time towards driving more sus-tainably;; hence, it could have changed the drivers’ locus of attention. By doing so, participants’ driving behavior was positively impacted, and they shifted gears earli-er (5.45 out of 7) and accelerated more smoothly (5.13 out of 7). These aspects of driving behavior are part of the eco-driving concept, supporting the goal of enabling participants to drive more sustainably. However, a reduction in the driving speed on motorways was not achieved (2.83 out of 7), as most drivers took advantage of the


unlimited speed limit on German motorways in order to reach their destination fast-er.

VI.5 Conclusion

This chapter introduced recommendations through which corporate car drivers can be motivated to reduce their fuel consumption with the support of eco-feedback technologies and without the benefit of any direct financial reward. Appraising the findings using the FIT of Kluger and DeNisi (1996) and an evaluation of extrinsic and intrinsic motivations revealed that it is imperative to stimulate drivers’ aware-ness of their fuel consumption. Feedback from the smartphone application and the MIS enhanced participants’ awareness of their personal performance and enabled them to modify their behavior when necessary. Actively directing drivers’ attention to the difference between their driving style and recommended sustainable driving habits was essential to the process and reduced the gap between actual driving hab-its and eco-driving practices (shrinking the feedback-standard gap defined in the FIT).

Intrinsic motivation was motivated by extrinsic motivations when self-determination and feelings of autonomy and competence were enhanced without making the participants feel controlled by the management. Research by Amabile (1993) supports these findings. Corporate car drivers were concerned that manage-rial monitoring would lead to loss of autonomy and punishment if their fuel effi-ciency did not improve. This highlights the need to ensure that an organizational roll-out of an eco-driving program is not associated with punishments, instead using positive motivation and developing drivers’ behavior via realistic goal setting and constructive feedback (e.g., providing fuel efficiency figures from the same vehicle model).

VI.6 Limitations and Future Research

Due to the relatively small sample size of the post-experiment survey (n = 24) and post-survey semi-structured interviews (n = 15), it may not be possible to generalize these findings to other companies. However, these findings can provide preliminary insight into the level of influence that eco-feedback technologies can have on corpo-


rate car drivers’ fuel efficiency. Furthermore, due to the long-term involvement of the researcher in this research domain, his interpretation of the interview data during the coding process may have been less objective than would be ideal. Findings from this study’s survey should be combined with theories and methods from the social-psychological research field in order to gain a more objective perspective by evalu-ating the data using statistically proven methodologies. For the purpose of further research, a field test with corporate car drivers from the same company in another region is planned. That region currently has 11,000 corporate car drivers, enabling access to a much larger sample size.

Development of an Eco-driving Dashboard System 121

VII Development of an Eco-driving Dashboard System Research findings were received well by the case study company’s different busi-ness areas. A follow-up project will focus on developing a prototype portal solution that informs each driver of their individual fuel consumption, and a follow-up field test with 11,000 corporate car drivers based in Germany is planned for end of 2013 or beginning of 2014. The goal is to offer the solution as a standard service in SAP´s fleet-management software and thus available to external customers. Busi-ness departments supporting this project include the fleet management department, the fleet-management software managers, the sustainability department, and the company’s workers’ council.

The prototype portal solution, which is currently in development, will utilize con-cepts and findings from the eco-driving field test presented in Chapter V and VI. The goal is to not only provide each corporate car driver with fuel consumption fig-ures (including fuel costs, fuel consumption, fuel efficiency, total distance driven, and average / total CO2 emissions) but also build a management dashboard enabling fleet managers to review the data via an online interface. As of now the data is be-ing received as a .csv flat file and evaluated and validated within MS Excel.

VII.1 System Architecture, Data Structure, and Data Processing The architecture comprises three back-end systems (see Figure 51): a) internal ser-vice provider system (ISP) in which each driver’s master data is stored, b) systems from SAP´s partner petrol companies (Aral, Esso, Total) that store fuel consumption data from each filling, and c) SAP Business Intelligence (BI). The BI system uses the SAP HANA database (DB), which stores fuel consumption data according to each driver´s master data, to calculate consumption figures. The front end is SAP´s Portal solution, which allows users to view the fuel consumption data in via an in-ternet browser.

Drivers´ master data (see Figure 52) include the driver´s ID, car type, and the DIN consumption figures for the car. The fuel consumption data details each petrol fill-ing, including the date, city, country, petrol type, total mileage, and total costs. The fuel consumption data are sent monthly as a flat file (.csv) to SAP´s fleet manage-ment department and figures needed for the fuel consumption calculations (such as total kilometers, total amount of liters filled per filling, and total costs) are stored in

122 Development of an Eco-driving Dashboard System

the HANA DB according to each driver´s ID, and consumption figures are calculat-ed by the BI system.

Figure 51: Eco-driving Dashboard Architecture

Figure 52: Driver´s Master Data Types

The HANA DB is an In-Memory database technology used by the BI system for primary data storage. The In-Memory technology allows faster data loading of


DataStore Objects (DSO), i.e. master data and fuel consumption data, and the BI models and calculates fuel consumption figures using defined mathematical equa-tions in InfoCubes. The BI system is connected to the Portal via an application pro-gramming interface, and the fuel consumption data is transferred via an .xml file to the front end (SAP´s corporate portal). The fuel consumption data are then dis-played to each driver under the ‘employee’ section of the portal and can also ac-cessed from a mobile device.

Figure 53: Eco-driving Dashboard Architecture

Since drivers have a unique single sign on token to SAP´s corporate portal, the data can only be accessed in the employee´s section from each individual driver. Through role management concepts assigned to each employee´s ID, the manage-ment cockpit can only be accessed from the car fleet management team and as-signed users.

VII.2 Mock-ups

Mock-ups were validated in the second quarter of 2013 using two user groups, cor-porate car drivers and fleet managers. Corporate car drivers receive individual fuel consumption figures, social comparison to other corporate car drivers with the same car type, information on used filling stations, historical fuel consumption, and eco-


driving tips (see Figure 54). A social-normative comparison with others’ data will be given if the corporate fleet has at least five cars of the same model. This thresh-old was defined by the company´s workers’ council and is necessary to maintain each driver’s privacy. A social comparison chart, color-coded from green to red, displays how fuel efficient the driver is when compared to other drivers. Fuel con-sumption is also shown on a timeline, allowing the driver to compare figures histor-ically. Finally, eco-driving tips are available through the portal, providing each driver with specific recommendations for improving fuel efficiency.

Figure 54: Corporate Car Driver´s Eco-driving Dashboard

The management´s cockpit (see Figure 55) provides an overview of the total corpo-rate car fleet, including total fuel costs, consumed fuel, total distance, and the petrol stations at which the most fuel was consumed. The social comparison shows aver-age fuel consumption by vehicle types across the corporate car fleet in meaningful comparisons across the same car brands and also car types. It is planned to develop further data useful for fleet management, such as vehicles’ maintenance cycles, in-dividual contract data, and drivers who are misusing their petrol cards (e.g., refilling the wrong petrol type or having average fuel consumption far above or below the average fuel consumption for their car type). Including a map displaying the most-used petrol stations has not been approved due to privacy issues.


Figure 55: Management´s Corporate Fleet Dashboard

VII.3 Mock-up incorporated in SAP´s Portal Figure 56 shows the latest version (as of June 2013) of the driver´s dashboard em-bedded in the Corporate Portal. This version shows all feedback types except for the filling stations. The interface is embedded in the portal with the mark-up language HTML5 and enables an interactive navigation experience by hovering over the con-sumption figures. By rolling-over a consumption figure (top left), the social com-parison figures are displayed in the top right. The ‘own consumption’ figure is shown in green, yellow, or red, depending on how ‘own comparison’ compares to drivers with the same car type. Historical monthly data are displayed in the timeline at the bottom of the screen. The historical data also show the actual fuel price in the market and compares the market price with the average price of the total number of fillings per month. This version for corporate car drivers is currently being tested with end-users to receive further feedback about its usability.


Figure 56: Management´s Corporate Fleet Dashboard

The fleet management version is still under development. It will apply the same concepts and technological features as for the driver´s dashboard and enables the management to get an understanding of the overall consumption data from the en-tire corporate fleet down to the driver´s and car types´ level.

VII.4 Drawbacks of the current eco-driving System

Findings from the research showed the impact real-time feedback has on sustainable behavior. However, fuel data has a one-month time lag, as data is transmitted from the petrol stations only once per month. Receiving the data more frequently from the petrol station companies would be beneficial for providing more up-to-date and frequent fuel consumption feedback to drivers.

The feedback is communicated via a monthly email to all drivers; afterwards, driv-ers can access the portal to review their fuel consumption data. This approach re-quires the users to proactively seek the data and might not be as effective for im-proving overall fuel consumption. Further feedback concepts, such as eco-driving training and videos, should be used in combination with eco-driving campaigns supported by the company´s management.


Another drawback involves the eco-driving tips. Findings showed that providing eco-driving tips solely by email has no effect. Therefore, the plan is to provide indi-vidualized eco-driving feedback by evaluating each corporate car driver’s historical data.

Conclusion, Recommendations, Outlook and Future Research 129

VIII Conclusion, Recommendations, Outlook and Future Research

This chapter summarizes key findings from the previous chapters for practitioners and academics, reflecting on the main research question:

How can eco-feedback technologies influence the average fuel efficiency of cor-porate car drivers when monetary incentives are not given?

Theoretical and practical motivations underpin the potential impact that eco-feedback technologies can have on promoting eco-friendly driving habits in corpo-rate car drivers. Within the research domain, only a limited number of studies focus on this driver group. This population of drivers remains unique, as their employing company reimburses their vehicle’s fuel and maintenance costs. This has allowed for a direct investigation of how use of eco-feedback technologies influenced driv-ers’ fuel efficiency in the absence of financial reward, the factor found to have the greatest impact in motivating private car drivers to modify their driving habits.

The next section, Key Findings, provides answers to the research sub-questions that were evaluated and summarized throughout this thesis. This demonstrates that theo-retical and practical contributions have been made by filling the research gaps that were outlined in the State of the Art and Related Work chapters. The Recommenda-tions section provides a novel conceptual framework, including guidelines that need to be considered when applying a fuel reduction program within an organization. Finally, the Outlook section discusses future research areas and offers insights into how a prototype eco-driving feedback information system is currently being devel-oped by the case study company.

VIII.1 Key Findings

Key findings are structured to align with the research sub-questions, highlighting important findings from each chapter.


130 Conclusion, Recommendations, Outlook and Future Research

Two online pre-experiment surveys completed by private and corporate car drivers compared each group’s preferences for receiving feedback from eco-feedback tech-nologies to improve their driving behavior. Key findings show that only 13% of private car drivers were aware of eco-driving smartphone applications and that navi-gation systems and in-vehicle on-board technologies were the best known. The lack of awareness of these smartphone technologies could have been due to the limited acceptance rate of smartphones in 2010. According to Gartner (2011), the smartphone penetration rate was only 19% in 2010 yet is expected to rise to 40% by 2014. A surge in popularity in eco-driving applications can be expected to directly correlate with this greater smartphone use.

Both driver groups had a strong preference for receiving undisruptive real-time feedback, as also identified by Meschtscherjakov et al. (2009). This real-time feed-back should be non-obtrusive and easy to use; offline-feedback, in comparison, can outline historical data about a person’s driving behavior. When comparing types of feedback information, such as personal fuel consumption, average fuel savings po-tential, and eco-driving tips, the feedback detailing the personal fuel consumption was the most important type for both driver groups. Following this, private car driv-ers preferred average fuel savings potential to personalized eco-driving tips while corporate car drivers preferred eco-driving tips to average fuel savings potential. This is not surprising, given that corporate drivers’ fuel costs are reimbursed.

Feedback about CO2 emissions was not as important, as 54% of the private car driv-ers and 27% of the corporate car drivers did not understand the meaning of the emissions unit, 100 grams of CO2 per kilometer. Further education and distinct comparisons need to be made to help users grasp this value by, for example, noting that 167 grams of CO2 per kilometer is equivalent to using a computer for nine hours daily for eleven working days.

These insights from the online surveys helped define which feedback types were given to corporate car drivers in the field tests.



The first field test provided direct/ real-time and indirect/ accumulated eco-driving feedback to twenty-five corporate car drivers who used the eco-driving smartphone application for eight weeks. In total more than 7,000 fuel fillings and 800 journeys by both treatment and control driver groups were recorded. An average improve-ment of 3.23% in fuel efficiency was found when comparing the treatment group to the control group, supporting the argument that mobile feedback technologies can positively impact corporate car drivers’ driving behavior. This conforms to the find-ings of Boriboonsomsin et al. (2010) and Siero et al. (1989), who provided real-time feedback using an on-board eco-driving system. Boriboonsomsin et al. (2010) showed an average fuel consumption improvement of 1% on highways and 6% on city streets with twenty participants, and Siero et al. (1989) demonstrated an im-provement of 7.3%. This large improvement in fuel efficiency was partly due to additional management guidelines and classroom-based training programs for the drivers.


The second part of the first field test evaluated driver’s preferences for direct/ real-time feedback as opposed to indirect/ accumulated feedback or offline feedback. Participants who used the eco-driving application for eight weeks were able to se-lect and change the feedback type when the application was in use. Results showed a strong preference for direct/ real-time feedback. These findings correlate with the literature on in-household energy savings, in which participants also favored real-time feedback (Darby, 2006; Graml et al., 2010). Besides the quantitative findings obtained from the data generated by the eco-driving smartphone application, quali-tative data was collected via the post-experiment survey and interviews. Evaluation of this data explained that real-time feedback was favored because it immediately notified the experienced corporate car drivers of when during their journey they could use specific eco-driving techniques. This was more effective than receiving feedback in an accumulated form or in three-minute intervals.


When considering offline feedback, an online driving portal (ODP) can play a role in increasing awareness of driving-related data (i.e. fuel usage, CO2 emissions, or routes driven) and eco-friendly driving concepts. However, showing a summary of driving-related figures in isolation does not initiate a change in driving behavior. The ODP must provide further feedback types, such as social comparisons, in order to motivate drivers. This was further evaluated to answer sub-question four.


The second field test provided 240 corporate car drivers, assigned into three treat-ment groups, with driving-related figures (e.g., fuel efficiency, CO2 emissions, total petrol costs, and total kilometers driven per month) via a monthly email generated by a mobility information system (MIS). The feedback about the average fuel con-sumption was also shown using social-normative feedback by displaying the driv-er’s fuel efficiency in comparison to the average fuel efficiency of all 240 drivers participating in the field test. At the end of the email, eco-driving and fuel reduction tips were provided.

Findings indicate that monthly feedback provided by an MIS email increased the average fuel efficiency of the first treatment group by 1.24% and that of the second group by 1.69%. The first and second groups received the social-normative feed-back displayed on categorical and continuous scales, respectively. These findings support literature in the fields of Green IS and social psychology that suggests that Green IS stimulates behavioral modifications to reduce energy consumption (Graml et al., 2010; Loock et al., 2011). However, providing eco-driving tips without any social-normative feedback figures did not have an impact on overall fuel efficiency.

Showing the feedback on discrete/ categorical in comparison to continual scale for-mats showed a difference of 0.73%; however, statistical validation did not provide a statistically significant difference in fuel efficiency between the two treatments groups. Hence, it cannot be concluded that showing feedback on a discrete/ categor-ical in comparison to a continual scale format has a stronger impact.

In summary, offline feedback presented in the form of a monthly fuel consumption email did have a positive impact on drivers’ fuel efficiency. It was important to rep-


resent fuel consumption figures in a social-normative feedback, only providing eco-driving recommendations did not have an impact.


The fifth sub-question derived guidelines and recommendations for how to stimu-late drivers’ intrinsic motivation according to the five basic arguments identified by the Feedback Intervention Theory (FIT) (Kluger & DeNisi, 1996). According to the FIT, eco-driving is at the moderate level of drivers’ hierarchy and therefore possible to direct drivers’ locus of attention to new driving habits. It is essential to actively direct drivers’ attention to eco-driving techniques in comparison to their existing driving behavior. Feedback technologies can be the medium through which to ad-dress this disparity and to raise drivers’ intrinsic motivations so that eco-driving becomes a natural driving habit. Drivers need to feel competent and autonomous in order to be intrinsically motivated; managerial control or punishment when fuel re-duction goals are not met can jeopardize this. Management should define realistic fuel reduction goals and provides constructive feedback (e.g., social-normative comparison with vehicles from the same car type).

VIII.2 Theoretical Contributions

VIII.2.1 Feedback Intervention Theory

This research contributes to the Feedback Intervention Theory (FIT) of Kluger and DeNisi (1996). Specifically, it builds on findings by Alder (2007), who applied the theory to motivate employees to improve their overall task performance by provid-ing feedback through a computer performance monitoring (CPM) system. Further-more, the research combines the FIT, originating in social psychology, with the or-ganizational behavioral research domain’s consideration of how employees’ extrin-sic and intrinsic motivational aspects can be triggered (Beswick, 2007). The FIT has never before been applied to appraise how feedback interventions influence the driving motivations and behavior of corporate car drivers.


According to the terms and definitions of the FIT, the external agents in this re-search were the eco-feedback technologies that provided different feedback types to motivate employees (corporate car drivers) to drive more sustainably. These eco-feedback technologies influenced corporate car drivers’ task performance and im-proved their fuel efficiency. Drivers compared the feedback provided to their goals, standards, and workplace norms. Through this evaluative process, drivers became more aware of the importance of sustainable driving. Applying the five arguments of the FIT created a structured approach to identifying motivations for corporate care drivers. Extrinsic and intrinsic motivational factors were evaluated to identify how drivers’ intrinsic motivations can be triggered in order to improve their fuel efficiency. These findings were evaluated in research sub-question five and thor-oughly discussed in Chapter VI.

VIII.2.2 Social Normative Feedback

These findings also contribute to the social-psychological theories developed by Cialdini et al. (1991), as this research evaluates how social-normative feedback in-fluences behavioral modifications. Existing findings by Domnez (2007) and Meschtscherjakov et al. (2009) showed that direct/ real-time feedback should not distract the driver; this was highlighted in experiments and interviews throughout this research. Indirect/ offline feedback needs to be presented in an enriched format to be influential. Providing historical data about fuel consumption alone did not have a strong influence on behavior. Instead, drivers requested a comparison be-tween themselves their cohort. Therefore, individual fuel efficiency figures were compared with the average fuel efficiency of all drivers participating in the field test. Drivers requested a valid comparison: most favored a comparison with drivers who had the same car model as theirs and drove similar average distances. Whether drivers shared the same job role, driving style, or worked for the same organizations was not relevant for the efficacy of social-normative feedback. These findings em-phasize the importance of providing the correct feedback type when applying so-cial-normative feedback to corporate car drivers.


VIII.2.3 Eco-Driving Concept

This thesis defined an initial concept for encouraging eco-driving within an organi-zation through eco-feedback technologies. Chapter VIII.4 defines the concept, which relates to relevant findings from the thesis. Clear recommendations were giv-en from the perspectives of both management and drivers. The concept defines real-istic fuel efficiency improvement goals, which feedback types should be used, and how feedback should be provided to best raise the awareness of corporate car driv-ers. Implementations of eco-driving in a corporate context should not only trigger short-term changes through extrinsic motivational factors like financial rewards. Long-term sustainable options for inspiring intrinsic motivations that make eco-driving a habit for employees are also highlighted. This concept can be used as a foundation for further research combining social-psychological, HCI, and infor-mation systems research streams.

VIII.3 Practical Implications

As stated in the introduction, the research context focused on three areas: the tech-nological, behavioral, and business value perspectives. Practical implications are outlined in each of these three areas.

VIII.3.1 Technological Perspective

Multiple technologies that communicate eco-feedback were tested. Outcomes of both field tests support Fogg’s (2002) argument that persuasive feedback technolo-gies can change an individual’s attitude or behavior. The applied eco-feedback technologies enabled corporate car drivers to improve their fuel efficiency, reducing the gap between each participant’s environmental awareness and their actual daily car usage. Froehlich et al. (2010) described this as closing the environmental litera-cy gap through feedback technologies. The smartphone eco-driving application pro-vided two different feedback types (direct and indirect) and revealed eco-driving techniques while an individual drove, whereas the MIS and the monthly fuel con-sumption email increased awareness of actual car usage by providing social-normative feedback. The fuel efficiency calculations and algorithms used to provide the social-normative feedback and findings from the second field test provided the


foundation for developing the fleet management system prototype. Development of the prototype is explained in detail in Chapter VII.

VIII.3.2 Behavioral Perspective

The pre-experiment surveys with private and corporate car drivers helped establish an understanding of both driver groups’ preferences for feedback technologies and types. They clearly identified that smartphone-based eco-driving technologies re-main underutilized; this is predicted to change in tandem with increasing smartphone sales figures (Gartner, 2011). Furthermore, both groups preferred to receive feedback in terms of fuel efficiency;; the meaning of ‘liters per kilometer’ was well understood. The meaning of CO2 emissions savings is unclear to most drivers, so further education is required to improve their level of understanding.

This thesis’ findings have also enhanced our understanding of corporate car drivers’ preferences and motivations for ecological driving. This research expanded upon Geller et al.’s (1982) examination of how to increase seat belt use and Graham et al.’s (2012) study into reducing car usage. It was found that it is important for cor-porate car drivers to maintain their autonomy and competence and not to be con-trolled or punished if they are unable to attain fuel reduction goals defined by man-agement. As Greengard (1996) stated, computer-mediated feedback can be seen as an invasion of employees’ privacy and lead to decreased job satisfaction;; it is im-portant to motivate drivers to improve their driving style themselves instead of forc-ing such changes upon them.

VIII.3.3 Business Value Perspective

The business value of this research includes financial savings for companies as well as environmental and conceptual rewards. Since the ratio of corporate to private cars in Europe is increasing annually (e.g., one quarter of cars in Germany and one third of those in Switzerland are corporate cars) potential fuel savings can be im-mense for companies with a large corporate fleet. Both field tests showed a poten-tial fuel savings as high as 3.23% without management following any further guide-lines. If management implements further recommendations, these saving could be even higher. Siero et al. (1989) demonstrated this, achieving long-term fuel savings of up to 7.3%. With total fuel expenditures of 42.5 million Euros in 2011, the case


study company could save up to 3.1 million Euros per year on fuel costs. Lower eventual deterioration of the car and fewer accidents due to improved driving styles are not included in this figure.

The environmental reward is a reduction in overall CO2 emissions without large investments by companies. Due to the ongoing environmental debate, fleet manag-ers have a strong interest in reducing overall fleet emissions as well as costs. Regu-lations and financial incentives that encourage employees to buy cars with lower CO2 emissions are already in place, but a change in vehicle alone will not lead to lower emissions. The driver must also adapt a more eco-friendly driving style, as up to 30% of fuel efficiency depends on driving style (Romm and Frank, 2006). In-creasing awareness and providing eco-driving tips via feedback technologies can sensitize drivers, as shown in both field tests.

Since the case study company’s core business is selling business software solutions, they have a strong interest in developing a software solution that can be affordable to a wide range of customers. Current fleet management software solutions offer functions used only by fleet managers, such as cars’ contracting details, car mainte-nance schedules, tank refill data, management of corporate fuel credit cards, and vehicle accident management. Providing this information to the individual driver is not current practice but can be implemented as an extra function within the fleet management software. The outcomes of the field tests were used as the foundation for designing and implementing a prototype solution to reduce overall CO2 emis-sions and fuel costs.

VIII.4 Recommendations and High Level Concept34

Analyzing the findings with respect to the FIT (Kluger and DeNisi, 1996b) and the concept of extrinsic and intrinsic motivations revealed important insights. These included ways to enable corporate car drivers to explore their intrinsic motivation to drive more sustainably by providing extrinsic motivation in the correct manner. The three core recommendations derived from the analysis are: a) to set realistic goals, b) to define feedback, and c) to provide feedback. The long-term process of enhanc- 34 Published paper: Tulusan, J., Steggers, H., Staake, T., Fleisch, E., Supporting eco-driving with eco-feedback technologies: Recommendations targeted at improving corporate car drivers’ intrinsic motivation to drive more sustainable, Energy Informatics 2012 (EI 2012), Atlanta, Georgia, United States, October 2012.


ing intrinsic motivation by applying these recommendations is illustrated in Table 22.

Table 22: Recommendations and High-Level Concept

Recommendation One: Set realistic goals

Behavioral change should be goal directed, as defining a realistic and context-specific goal is essential for stimulating change (Kluger and DeNisi, 1996b). Re-sults of the experiments showed an average fuel efficiency improvement of 3.23% when using the eco-driving smartphone application for the duration of eight weeks and 1.69% and 1.24% when providing an individual with a monthly fuel consump-tion email for three months. Siero et al (1989) found a 7.3% reduction in the fuel consumption of postal truck drivers when extrinsic incentives were given, realistic


goals defined, and management directions provided. Considering the results of these field tests and other studies, the case study company’s management is advised to set a realistic overall fuel efficiency improvement goal of 3-5% per driver. This could lead to potential fuel savings of up to 2.1 million Euros per year for the case study company.

Recommendation Two: Define feedback

The corporate car drivers from the case study company drive an average of 24,000 kilometers per year without any knowledge of their average fuel consumption and fuel costs. Fuel and other costs (e.g., car maintenance) are reimbursed by the com-pany and are not known by the drivers. By exposing drivers to these details, their fuel efficiency could be positively influenced, as their attention is directed to the feedback-standard gap (i.e. actual vs. expected fuel consumption) (Kluger and DeNisi, 1996b). It is recommended that the management provide drivers with these details in order to increase their overall awareness of their fuel consumption and petrol costs.

Once fuel consumption and cost figures are publicized and realistic goals set, driv-ers require constructive advice about how to improve fuel efficiency. For instance, eco-driving techniques should be promoted and training courses offered. Regular feedback should indicate drivers’ progress towards meeting their goal, enhancing the driver’s understanding and extrinsic motivation. The driver must be able to rec-ognize the positive impact they are having by reducing CO2 emissions. The man-agement should provide information on fuel costs, consumption, and efficiency, offer personalized eco-driving tips, and provide a valid comparison between drivers driving the same vehicle model for similar annual distances. Feedback leveraging such social comparisons has already been successfully applied in the literature on household energy reduction (Loock et al., 2011).

Recommendation Three: Provide feedback

How should the feedback be provided? The findings revealed that an eco-driving smartphone application is a cost-effective way to modify drivers’ driving behavior by providing positive real-time feedback. It is recommended that the application be incorporated into an overall organizational energy reduction plan that focuses on not


only feedback technologies but also governmental aspects. Providing details through a monthly fuel consumption e-mail or a mobility information system (e.g., an online portal) raises drivers’ awareness of their own fuel consumption.

Raising awareness about drivers’ fuel consumption is an important step towards enhancing corporate car drivers’ intrinsic motivation to drive in a more eco-friendly manner. Management should not punish or control drivers if they are unable to reach the fuel reduction goals given, as this would result in employees feeling con-trolled, with a loss of autonomy and a diminished sense of motivation. Efforts to implement these recommendations in the long term should support employees’ au-tonomy, especially in situations where drivers must use their car for their daily work.

VIII.5 Limitations35

One limitation is the generalizability of these findings to other organizations or re-gions, as the field tests were conducted with only one company within one region, Switzerland. In social science research, it is common to apply findings from a case study to a broader geography or to a wider social phenomenon (Flick, 2006; Wil-liams, 2000). In this particular case, employees of the Swiss organization frequently travel to Germany, and a follow-up field test in Germany is planned in 2013.

Participants in the eco-driving smartphone study were selected from one company’s pool of drivers. As it was an opt-in field test, participants who chose to take part in the experiment may have already had a pro-environmental attitude or a higher tech-nological affinity, creating a degree of bias in sample selection. It was possible to control this aspect by analyzing their environmental beliefs and technological affini-ty, as expressed in the post-experiment survey, using scales validated in the energy savings literature. Findings indicated that drivers only had moderate (i.e. ‘partly agree’) pro-environmental attitudes (5.02 out of 7.0) and technological affinity (5.1 out of 7.0); this suggests that the results of this study can be applied to other corpo-rate car drivers with similar environmental attitudes and technological affinity.

35 Published paper: Tulusan, J., Staake, T., Fleisch, E., Direct or indirect sensor enabled eco-driving feed-back: Which preference do corporate car drivers have?, Internet of Things 2012 – Third International Confer-ence on the Internet of Things (IoT 2012), Wuxi, P.R. China, October 2012.


The data obtained in the second field test (using monthly feedback provided by an MIS) could have been evaluated further. The rich datasets obtained from the MIS enable making comparisons based on many potential variables, including petrol and car type, gender, the employees’ position in the company, and the total monthly dis-tance driven.

The layout and design of the social-normative feedback provided via the monthly fuel consumption email needs to be evaluated further. The data showed no statisti-cally significant difference between the continual and categorical social-normative feedback scales, so it was unclear whether one of these feedback designs has a stronger influence when compared to the other. Nevertheless, providing feedback using social-normative comparisons had a positive influence and was preferred by participants. The follow-up field test with 11,000 corporate car drivers will focus on this aspect.

VIII.6 Future Research

Future academic field tests could evaluate if the social-normative feedback should be displayed in color or using grey bar charts and could consider which layout has a stronger impact on fuel efficiency. Whether the feedback scale displaying fuel effi-ciency in comparison to other drivers with the same car type should be continual or categorical needs to be further evaluated, as findings from this research did not show a statistically significant difference between the two feedback scales.

Furthermore, the impact of different feedback types (i.e. only fuel costs vs. CO2 emissions vs. both feedback types) could be evaluated. Findings from Graham et al. (2011), examining private car drivers, already identified that financial savings have a stronger influence in motivating individuals to drive less than does feedback stat-ing CO2 emissions savings. This information would help practitioners choose the feedback that has the strongest impact.

In addition to behavioral modification elements, further research should evaluate how conceptual and managerial recommendations can be further applied in a real-world setting. It is important to gain a better understanding of how and if the stated recommendations trigger corporate car drivers’ intrinsic motivation. If so, does sus-tainable driving become habitual even when the employees are not financially reim-


bursed? This question could not be answered fully through this research and re-quires long-term field studies to be able to measure the changes in fuel consumption over a longer period of time.

Appendix 143

Appendix LIST OF APPENDIXES Appendix A: Pre-survey Transportation Habits of Private Car Drivers ............ 144

Appendix B: Pre-survey Transportation Habits of Corporate Car Drivers ....... 152

Appendix C: Screenshots of EcoDriving Application (EcoDrive, 2012) .............. 154

Appendix D: Eco-Driving Smartphone Application Stata Output and............... 157

Appendix E: Data set Example from DriveGain Application Data Logger ........ 159

Appendix F: Eco-Driving Smartphone Application Post Survey ........................ 160

Appendix G: Eco-Driving Smartphone Application Post Interviews .................. 166

Appendix H: Monthly Personal Car Consumption Email with discrete Scale ... 169

Appendix I: Environmental Behavior Survey ...................................................... 169

144 Appendix

Appendix A: Pre-survey Transportation Habits of Private Car Drivers Conducted in August 2010 Welcome and thank you in advance for your time! This survey is part of a study about feedback systems, how they could influence your driving style and what kind of features you would expect from such a system. The survey will have 24 questions and will not take more than 10 minutes. Please try to answer all questions, as it is very important for the validity of results. However, there is not a right or wrong answer; what is more important is your opin-ion and true preferences. Should you have any questions or want to leave some feedback, please do not hesi-tate to contact me on [email protected]. Please note that your answers will be treated confidentially. Introduction Feedback systems can measure your driving style and provide relevant information. This can help you to understand e.g. how much fuel you consume or if you acceler-ate aggressively. By making you aware of these aspects, such systems can help you to achieve a more efficient driving style, with reduced fuel consumption and conse-quently less CO2 emissions. Here are some photos of systems that currently exist in the market. The first one is based on a navigation system, the second on an onboard gauge36 and the third on an iPhone application.

Figure 57: Audi navigation system (under development)

36 Gauge = any kind of instrument used for measurement. In cars, a gauge may show information

about fuel level, temperature, etc.

Appendix 145

This navigation system provides calculations for three different types of routes, namely optimal fuel efficiency, time and emissions. Drivers can also see the amount of different types of emissions they save if they choose the one with lowest emis-sions, compared to the rest of the routes37.

Figure 58: Ford Gauge

This gauge, placed in the car's dashboard, provides different levels of information to choose from and displays either instant or long-time feedback regarding fuel con-sumption, depending on what the driver is most interested in finding out.38

Figure 59: greenMeter iPhone application

This iPhone application displays fuel usage and carbon emissions amount, in order to improve efficiency by providing instantaneous feedback.39

37 Source: Audi ""Clean Air, a Viable Planet"" Initiative, Eco-Driving Systems: Now Your Car Can

Gently Nag You Into Being More Fuel-Wise, Green Car Advisor, 2009 38 Source: Ford, Eco-Driving Systems: Now Your Car Can Gently Nag You Into Being More Fuel-

Wise, Green Car Advisor, 2009 39 Source: greenMeter Application, Hunter Research & Technology, 2009

146 Appendix

I. Introductory questions 1. Do you know any feedback devices that aim to improve one's driving behav-ior? Yes (135; 68%) No (63; 32%) 2. If yes, what kind of feedback devices do you know? Multiple answers are

possible Navigation System (89; 35%) Smartphone Application (33; 13%) Trip Computer (64; 25%) Other (9; 4%) On-board Gauge ( 57; 23%) 3. Have you ever used such a device while driving? Yes (103; 51%) No (99; 49%) 4. If yes, did you notice any change in the way you drove? Yes (65; 68%) No (30; 32%) 5. Could you describe these changes briefly? Answers see Chapter IV. 6. If no, do you think that it could affect your driving behavior towards a more

efficient driving? Yes (83; 85%) No (15; 15%) II. Attribute Importance 1. What type of information would you prefer to see in a feedback device? 1 Most preferable … 5 Least preferable AVG fuel efficiency (monthly) = 4.47 AVG fuel savings (monthly) = 4.02 Personalized eco-driving tips = 3.86 Alternative eco-efficient route = 3.79 Goal setting = 3.46 AVG CO2 emission (monthly) = 3.39 Social comparison = 3.16 Rewards = 3.03

2. When would you prefer to see information relevant to your driving? 1 Most preferable … 5 Least preferable During driving = 3.97 After driving = 3.70 Aggregated = 3.54 Before driving = 3.23

Appendix 147

3. How would you prefer feedback information to be provided? Visually (93; 52%) Audibly (10; 5%) Both (11; 43%) 4. Many feedback systems display the amount (in grams) of CO2 emissions per

mile or km produced by driving. Are you able to translate the meaning of e.g. 100 g/km to a feasible dimension?

Yes (83; 46%) No (97; 54%) 4a. If Yes: Do you believe that displaying carbon emissions is relevant to

improve your driving efficiency? Yes (54; 67%) No (27; 33%) An average car in the UK emits 167 CO2 g/km driven, whereas for each liter of pet-rol consumed 2.3 kg CO2 are emitted. In order for you to understand, 2.3 kg of CO2 emissions result from: - showering for 35 minutes - cooking on the electric stove for 2.3 hours - operating a lift for 92 minutes - using a computer for 11 working days of 9 hours each - having the fridge on for 14 days - driving an average car for 9.2 km - traveling by bus, train or tube for 37km 4b. If No: Do you believe that displaying carbon emissions is relevant to

improve your driving efficiency once you have a better understanding of the meaning of CO2 emissions?

Yes (69; 71%) No (28; 29%) III. Intro systems description You will now see pictures of feedback systems, which aim at helping you to im-prove your driving behavior. After reading their description carefully, you will have to rate them according to your preference order.

Figure 60: Gauge in cars dashboard

148 Appendix

The leaves are part of the gauge in the car's dashboard 40. They react instantly to your driving behavior by flourishing when you drive efficiently and disappearing when you don't.

Figure 61: Screen on top of the car dashboard

This screen is on top of the car's dashboard.41 It changes colors according to your driving behavior; it turns dark blue when you consume more fuel and bright green when you drive more efficiently.

Figure 62: iPhone Application

This is an iPhone application.42 While you drive efficiently the silver ball remains into the green area. If you start consuming more fuel or accelerate aggressively, the ball moves to the yellow and red areas. In addition to this you get a driving score, marking your driving behavior at each trip. 40 Source: First Drive: 2010 Ford Fusion Hybrid, autoinsane, 2009 41 Source: The Game (Changer), motive, 2009 42Source: GreenGasSaver, The Apple Seed Story, 2009

Appendix 149

Figure 63: PC Application

This is an application for your PC, which collects data of your driving behavior.43 The amount of CO2 emissions that is saved because of efficient driving is translated to the energy consumed for certain everyday activities, such as lighting, heating or cooking.

Figure 64: iPhone Application

This is an iPhone application.44 It shows the amount of CO2 emissions produced for each speed range, and which proportion is produced by acceleration, aerodynamic drag or rolling resistance. With this information, you can adjust your driving style towards producing less emissions.

43 Source: ecoDrive, Fiat 44 Source: greenMeter Application, Hunter Research & Technology, 2009

150 Appendix

Figure 65: Audi navigation system (under development)

This is an on-board navigation systems.45 It provides three alternative routes, opti-mized for time, fuel efficiency or emissions, based or real-time data it receives for traffic conditions and road topology. Plus, you can see the amount of emissions sav-ings, depending on which route you choose. 1. Now, please rank the previous systems according to the order of your

preference 1 Most preferable … 5 Least preferable Color-changing dashboard = 4.40 Leaves shown in dashboard = 3.72 Navigation system = 3.71 Silver Ball iPhone app. = 3.41 CO2 emissions iPhone app. = 3.24 PC application = 2.60

2. What made you choose the #l_3# (your most preferred one)? Answers see Chapter IV. 3. What didn't you like in #l_4# (your least preferred one)? Answers see Chapter IV. 4. Do you have any further ideas to display information about eco-efficient driving? Answers see Chapter IV. IV. Demographic Questions 1. What is your gender? Male (88; 64%) Female (50; 36%)

45 Source: Audi "Clean Air, a Viable Planet" Initiative, Eco-Driving Systems: Now Your Car Can

Gently Nag You Into Being More Fuel-Wise, Green Car Advisor, 2009

Appendix 151

2. What is your age? 18 - 24 (28; 28%) 40 - 44 (2; 1%) 25 - 29 (67; 49%) 45 - 49 (2; 1%) 30 - 34 (23; 17%) 50+ (2; 1%) 35 - 39 (4, 3%) 3. How many years of driving experience do you have? 0 - 4 (39; 28%) 15 - 19 (10; 7%) 5 - 9 (60, 43%) 20+ (7; 5%) 10 - 14 (23; 17%) 4. How often do you drive? Less than once a month (22; 16%) 2 - 3x per week (17; 12%) 1 - 3x per month (11; 8%) Every day (80; 58%) Weekly (9; 6%) 5. An average European driver drives about 40 km (roughly 25 miles) per

day. How much do you drive per day? Less than that (85; 61%) More than that (54; 39%) 6. Which of the following driving styles do you think best describes

yours? Economic (32; 23%) Aggressive (17; 12%) Careful (45; 32%) Risky (4; 3%) Sportive (41; 30%)

152 Appendix

Appendix B: Pre-survey Transportation Habits of Corporate Car Drivers Conducted in August 2011 I. Questions about Transportation 1. How often do you go to work by public transport? 1 - 2x per week (21; 14%) Every day (7; 5%) 3 - 4x per week (14; 10%) Not applicable (105; 71%) 2. How often do you drive to work by your car? 1 - 2x per week (22; 15%) Every day (71; 48%) 3 - 4x per week (43; 30%) Not applicable (11; 7%) 3. Please estimate your average monthly amount of kilometers driven Not known (10; 7%) Between 2011 – 3000 (39; 21%) Between 0 – 1000 (29; 20%) Above 3000 (24; 16%) Between 1001 – 2000 (45; 31%) 4. Please estimate your monthly average fuel emission (in l/ 100km) Less than 4 l (1; 1%) 10 - 11 l (8; 5%) 4 - 5 l (15; 10%) 11 - 12 l (0; 0%) 6 - 7 l (71; 49%) More than 12 l (0; 0%) 8 - 9 l (36; 32%) Not known (5; 3%) 5. Please estimate and indicate your average monthly cost of petrol? (in CHF) 0 - 200 (37; 25%) Above 600 (7; 5%) 201 - 400 (54; 37%) Not known (38; 26%) 401 - 600 (11; 7%) 6. Please estimate and indicate your average monthly CO2 emission (in kg) 0 - 200 (8; 5%) Above 600 (3; 2%) 201 - 400 (10; 7%) Not known (114; 78%) 401 - 600 (11; 8%)

7. Do you understand the meaning of 100gram of CO2 per km?

Yes (54; 37%) Yes, but I could not relate it (54; 37%) No (39; 26%)

If the answer to the previous question is 'No': In your opinion, would it be easi-er to understand the meaning of '100 grams of CO2 per km' if a comparison would be given. For instance: 100g of CO2 per km produced is equal to the us-age of a computer for 4.5 hours? Yes (79; 60%) No (40; 31%) Other comparison (12; 9%)

Appendix 153

8. Do you believe that knowing your personal CO2 emission is relevant to im prove your average fuel efficiency Yes (77; 53%) No (69; 47%)

9. Please evaluate the importance of receiving following information types, which could influence you to improve your fuel efficiency

1 Strongly disagree … 7 Strongly Agree

Personalized eco-driving tips = 3.83 AVG fuel efficiency (monthly) = 3.77 AVG CO2 emission (monthly) = 3.11 AVG km driven (monthly) = 3.06 AVG fuel savings (monthly) = 2.89

10. If you are an iPhone user, can I contact you in order to provide more in formation about the EcoDriving Yes (45; 35%) No (85; 65%)

154 Appendix

Appendix C: Screenshots of EcoDriving Application (EcoDrive, 2012)

Figure 66: Description in iTunes Store

Figure 67: Main Screen

Figure 68: Setting Screen

Figure 69: Add Vehicle

Appendix 155

Figure 70: Enable Upload of Data

Figure 71: Shift to N Gear

Figure 72: Shift to 1st Gear

Figure 73: Shift to 2nd Gear

156 Appendix

Figure 74: Shift to 3rd Gear

Appendix 157

Appendix D: Eco-Driving Smartphone Application Stata Output and Description of Variables

A. Stata Output

n id

id_trtmt_

grpf_

type km_fil* l_fil* date_fil*km_drvn fe

avg_fe_

rtmt_tfe_

oem

diff_oem_

trtmt_t

diff_oem_

trtmt_t_perc

avg_fe_y

diff_trtmt_

avg_fe_y1 23 0 D 95185 63.11 27.10.11 586 10.77 8.29 7.10 1.19 16.80 8.38 -1.042 24 0 D 67051 54.02 28.10.11 727 7.43 7.44 7.00 0.44 6.34 7.47 -0.393 26 0 D 10300 37.62 01.11.11 500 7.52 7.82 5.20 2.62 50.44 6.82 14.714 27 0 D 21876 46.45 08.11.11 823 5.64 8.08 7.10 0.98 13.81 7.99 1.115 58 0 D 12070 66.9 29.10.11 578 11.57 12.42 10.60 1.82 17.21 9.40 32.166 29 0 D 143895 51.76 26.10.11 1431 3.62 8.20 8.10 0.10 1.21 8.00 2.427 30 0 D 97630 46.28 26.10.11 720 6.43 6.57 6.20 0.37 5.90 6.62 -0.768 31 0 D 45585 57.1 25.10.11 735 7.77 8.34 5.70 2.64 46.36 8.09 3.099 32 0 D 162157 67.4 26.10.11 2077 3.25 6.84 6.20 0.64 10.26 6.67 2.50

10 33 0 U 97371 58 28.10.11 500 11.60 12.49 10.10 2.39 23.68 12.89 -3.1211 36 0 D 27800 63.56 03.11.11 900 7.06 7.67 6.20 1.47 23.72 7.39 3.8412 38 0 D 34865 60.72 31.10.11 788 7.71 7.64 7.00 0.64 9.21 7.69 -0.6513 39 0 U 17181 43.58 01.11.11 489 8.91 8.43 7.20 1.23 17.05 7.77 8.4014 40 0 D 53200 52.62 25.10.11 1100 4.78 8.82 6.70 2.12 31.70 8.50 3.8415 41 0 D 63317 53.09 07.11.11 790 6.72 6.80 6.00 0.80 13.34 6.95 -2.1516 42 0 U 43425 50.35 24.10.11 607 8.29 7.85 7.20 0.65 9.00 7.87 -0.3017 43 0 D 130770 72.38 31.10.11 820 8.83 8.22 8.70 -0.48 -5.50 8.13 1.1118 45 0 D 61882 44.3 26.10.11 527 8.41 8.70 7.50 1.20 15.96 8.88 -2.0419 47 0 D 68010 59.4 04.11.11 690 8.61 8.81 6.94 1.87 27.01 8.75 0.7820 49 0 U 56387 47.7 04.11.11 -392 -12.17 11.33 8.10 3.23 39.92 9.42 20.3621 53 0 D 80747 35.73 24.10.11 623 5.74 6.31 5.70 0.61 10.71 6.64 -4.9422 54 0 D 5034 51.85 12.11.11 872 5.95 5.92 5.90 0.02 0.31 6.05 -2.1323 55 0 U 118777 24.46 24.10.11 6252 0.39 10.48 8.00 2.48 30.98 9.73 7.6724 56 0 D 25380 96.54 12.11.11 880 10.97 11.10 9.30 1.80 19.37 10.92 1.6625 57 0 D 28100 63.28 25.10.11 950 6.66 7.66 5.90 1.76 29.83 7.38 3.8626 1 1 D 53336 43.99 25.10.11 681 6.46 6.41 5.90 0.51 8.68 6.49 -1.2527 2 1 D 30500 50.78 25.10.11 700 7.25 6.76 5.80 0.96 16.58 6.63 1.9228 3 1 D 104306 58.16 24.10.11 1176 4.95 5.13 5.30 -0.17 -3.21 4.64 10.6729 4 1 D 38400 45.3 25.10.11 750 6.04 6.25 4.80 1.45 30.19 6.24 0.0930 5 1 D 67500 59.2 30.10.11 900 6.58 6.25 6.60 -0.35 -5.26 6.84 -8.5631 6 1 U 22054 26.49 25.10.11 306 8.66 8.95 7.60 1.35 17.81 9.09 -1.5032 7 1 D 71712 45.07 24.10.11 212 21.26 7.59 6.70 0.89 13.34 7.87 -3.5433 8 1 D 28794 64.91 24.10.11 729 8.90 8.32 7.00 1.32 18.80 8.31 0.0234 9 1 D 14567 58.87 18.11.11 948 6.21 6.72 6.10 0.62 10.21 7.65 -12.1535 10 1 D 120720 54.3 25.10.11 659 8.24 7.62 6.00 1.62 26.96 7.55 0.9236 11 1 D 724 61.39 04.12.11 0 6.86 5.20 1.66 31.9937 12 1 D 118573 16.02 25.10.11 798 2.01 6.82 6.00 0.82 13.71 6.78 0.6038 13 1 D 66666 64.33 30.10.11 1666 3.86 8.58 7.50 1.08 14.42 9.37 -8.4439 14 1 D 44432 61.5 31.10.11 886 6.94 6.71 6.40 0.31 4.80 6.72 -0.1240 15 1 D 95675 44.33 26.10.11 658 6.74 6.90 6.00 0.90 15.05 7.17 -3.6941 16 1 U 107200 49.53 24.10.11 700 7.08 7.36 6.60 0.76 11.49 7.61 -3.2842 17 1 D 51000 59.4 30.10.11 700 8.49 8.77 6.20 2.57 41.48 8.74 0.4143 18 1 D 15428 57.92 27.10.11 1041 5.56 5.60 5.20 0.40 7.68 5.97 -6.2644 19 1 D 65700 78.29 25.10.11 950 8.24 8.57 8.20 0.37 4.51 8.46 1.2845 20 1 U 56305 37.86 26.10.11 405 9.35 9.35 8.60 0.75 8.70 9.42 -0.7246 21 1 U 43159 45.4 06.11.11 551 8.24 7.63 7.00 0.63 9.07 7.72 -1.1347 22 1 D 165441 64.65 25.10.11 786 8.23 8.58 7.40 1.18 15.96 8.46 1.38

*Data example shows only last filling. All tank filling data were evaluated for the fuel efficiency calculations.

158 Appendix

B. Explanation of Variables Variable Explanation n Number of participants id The id number of the driver

id_trtmt_grp Sorts the drivers into 0 = Control Group (No scale, just eco-driving tips) 1 = Treatment Group 1 (Discrete scale)

f_type The fuel type of the filling km_fil Total amount of km shown on the board computer at the

time of the filling l_fil Amount of liters filled up (in liters) date_fil Date of the filling km_drvn Calculated difference of the value of km_fil of two con-

secutive fillings fe The fuel efficiency for each filling is calculated as

l/100km avg_fe_trtmt_t Average fuel efficiency in the treatment time

(24.10.2011 - 16.12.2011) (in l/100km) per driver fe_oem Average Fuel Efficiency defined by car manufacturer

diff_oem_trtmt_t Difference between avg_fe_trtmt_t and fe_oem

diff_oem_trtmt_t_perc Percentile difference between avg_fe_trtmt_t and fe_oem

avg_fe_y Average Fuel Efficiency for the whole year until 23.10.2011 (in l/100km) per driver

diff_trtmt_avg_fe_y Difference of the average fuel efficiency between the treatment time and the whole year until 23.10.2011 (in l/100km)

Appendix 159

Appendix E: Data set Example from DriveGain Application Data Logger Name of driver: Driver 1

Trip 1 Trip 2 Trip 3 Trip 4

Time 25.10.2011

05:38 26.10.2011

04:47 26.10.2011

15:34 26.10.2011

15:54

From Bahnhofstr.,

Lengnau (BE) Fabrikstr.,

Lengnau (BE) Weiningen

(ZH) Badenerstr.,

Dittwil

To Vauffelin Birrhard Wettingen Kestenholz

Journey Score

65.32 60.24 70.9 80.37

Distance (m) 6680 80443 11506 49272

Duration (s) 459 3249 900 2498

Fuel (l) 0.48 7.44 1.69 4.7

Vehicle Type

Audi A3 Sportback 2.0




Fuel accel 0 0.4 0.54 0.01

Fuel brake 0.04 0.21 0.12 0.16

Fuel speed 0.05 2.34 0.23 1.43

Fuel savings meter (s)

100 0 802 498

Advanced savings (s)

359 3249 98 2000

160 Appendix

Appendix F: Eco-Driving Smartphone Application Post Survey Dear Colleague, The goal of this questionnaire is to receive feedback from you about the usage of the eco-driving DriveGain application, its functionalities and your preferences re-garding different feedback forms (section I and II). In addition, questions relate to your perception on sustainability, your technological affinity and demographic as-pects (section III, IV, V). Duration: Five sections. Around 12 minutes (pre-tests have been conducted). Answer Options: Multiple Choice (1 or more options)

Scale (Select one option: Strongly disagree, Disagree, Partly dis-agree, Neutral, Partly agree, Agree, or Strongly agree)

Evaluation and Confidentiality: No intention to relate any answer to you as an individual. All answers are confidential and only an anonymzed summary of the answers will be presented. I would be grateful though if you would provide your email address at the end of the survey. Many thanks for your support. I. Questions about the DriveGain Application 1. How often did you use the application per week during the field test on

average? Never (0; 0%) 3-4x (6; 25%) 7-8x (9; 38%) more than 10x per week (1; 4%) 1-2x (2; 8%) 5-6x (4; 17%) 9-10x (2; 8%) 2. For which routes and travel purpose did you use the application? (Multiple selections possible) Motorway (23; 96%) Business related travel (23; 96%) Country/ rural road (18; 75%) For long distances, more than 40 km (22; 92%) Inner cities (10; 42%) For short distances, less than 40 km (11; 46%) Private related travel (12; 50%) 3. The position of the application at the window negatively influenced my

attention level to the road 1 Strongly disagree … 7 Strongly Agree = 2.67

4. How often did you check your journey score in the DriveGain online portal

per week during the 8 weeks of the field test? Never (2; 8%) 3-4x (2; 13%) 7-8x (1; 4%) more than 10x per week (0; 0%) 1-2x (17; 71%) 5-6x (1; 4%) 9-10x (0; 0%)

Appendix 161

5. How often did you discuss your driving behavior with your colleagues/ friends/ family during the 8 weeks of the field test?

Never (5; 21%) 3-4x (11; 46%) 7-8x (0; 0%) more than 10x per week (0; 0%) 1-2x (6; 25%) 5-6x (2; 8%) 9-10x (0; 0%)

II. Questions about Feedback Types/ Meters II A: Feedback by DriveGain Application 1. The feedback about my eco-driving behavior was presented clear and

understandable. 1 Strongly disagree … 7 Strongly Agree = 5.29

2. I am more aware of my driving behavior after receiving eco-driving

feedback by the smartphone application. 1 Strongly disagree … 7 Strongly Agree = 4.79 3. The feedback given by the application helped to change my driving

behavior. 1 Strongly disagree … 7 Strongly Agree = 4.63

4. The feedback given by the application influenced me in… 1 Strongly disagree … 7 Strongly Agree … a smoother acceleration. = 5.13 …smoother braking. = 4.83 ...reducing my average speed on motorways. = 2.83 …shifting gears earlier during acceleration. (please do not answer if you have an automatic gear-box.)

= 5.45

…driving in the highest gear as long as possible. (please do not answer if you have an automatic gear-box.)

= 4.36

5. The feedback I received from the application distracted me during driving.

1 Strongly disagree … 7 Strongly Agree = 3.08

6. Below you can see the feedback meters of the application:

162 Appendix

Figure 75: Fuel Savings Meter

Figure 76: Advanced Savings Meter

Have you been aware of the option to change the feedback Yes (16, 67%) No (8, 33%)

Items 1 Strongly disagree … 7 Strongly Agree I mostly chose to receive feedback via the “Fuel Savings” (feedback every three minutes; left figure).

= 2.41

I mostly chose to receive feedback via the “Advanced savings” (real-time feed-back; right figure). = 6.06 I used both feedback meters evenly. = 2.22 Receiving feedback about the last three minutes of my driving behavior was use-ful.

= 2.00

Receiving real-time feedback about my driving behavior was useful. = 5.06

7. I received enough feedback about how I drive from the DriveGain

application. 1 Strongly disagree … 7 Strongly Agree = 4.71

8. I would like to receive more feedback about how I drive.

1 Strongly disagree … 7 Strongly Agree = 4.13

II B: Form and frequency of feedback

Appendix 163

1. I would like to receive feedback from the application about how I drive …

1 Strongly disagree … 7 Strongly Agree …by a computerized voice = 3.33 …by visualization = 6.00 …by a computerized voice and visualization = 3.33 …by a weekly email = 4.88 …by an online portal = 3.21 2. I would prefer to receive feedback from the application when I request it

(=pull) rather than receiving feedback automatically (=push). 1 Strongly disagree … 7 Strongly Agree = 3.21

1. I would like to receive feedback from the application about how I

drive in comparison to… 1 Strongly disagree … 7 Strongly Agree …drivers with the same car model. = 6.00 …drivers with the same driving style. = 3.42 …drivers who drive on average a similar amount of km. = 5.63 …drivers from my organization. = 2.92 ...drivers with the same job role in my organization. = 3.83 4. I would like to receive feedback from the application about how I drive

(multiple selections possible): Real-time during driving (21; 88%) Summarized once a month (11; 46%) Immediately after driving (12; 50%) Summarized once every 3 months (3; 13%) Summarized once a day (2; 8%) Summarized once a year (3; 13%) Summarized once a week (15; 63%) III. Environmental Behavior Scale The validated scale evaluates your environmental behavior. I would appreciate if you answer all questions (including 5 to 9), since it will provide a complete picture of your ecological perception. Items 1 Strongly disagree … 7 Strongly Agree 1. To drive sustainable is an important part of my personality. = 5.00

2. Through sustainable driving, I would like to show to others that I am an ecologically-aware person. = 3.50

3. I like to talk about my fuel efficiency. = 4.60 4. I discuss sustainable driving tips and new technologies (e.g. electric vehicles) with others = 4.30

5. In hotels I let my towels to be changed every day. = 6.20

164 Appendix

6. If I wear a shirt for one day, it will always be washed. = 5.00

7. At home when I leave the room, I switch off the light. = 6.10

8. At home I leave electrical equipment, such as TV, printer, coffee machine, etc., running on standby mode. = 5.15

9. In winter it is warm enough in my flat/ house to wear a T-Shirt. = 5.40

Total Average = 5.0

IV. Technological Affinity Items 1 Strongly disagree … 7 Strongly Agree 1. My friends and colleagues often ask my opinion about new technologies (e.g. smartphones). = 5.00

2. My friends and colleagues usually know more about new technologies than I do. = 4.60

3. I am always up to date with new technological improvements which interest me. = 5.00

4. I have fun to test new technologies. = 5.80 5. For me technology is only a means to an end (= dient nur als Mittel zum Zweck). = 5.10

Total Average = 5.10 V. Demographic Questions 1. What is your gender? Male (18; 75%) Female (6; 25%) 2. What is your age? Below 30 (1; 4%) 40 - 44 (6; 25%) 50 - 54 (3; 13%) 30 - 34 (4; 18%) 45 - 49 (7; 28%) 55 - 60 (3; 13%) 3. How many years of driving experience do you have? 0 - 9 (0; 0%) 15 - 19 (1; 4%) 25 - 30 (8; 33%) 10 - 14 (5; 21%) 20 - 24 (6; 25%) More than 30 years (4; 17%) 4. How often do you drive? Less than once a month (0; 0%) 1 - 2x per week (0; 0%) 1 - 2x per month (0; 0%) 3 - 4x per week (1; 4%) Every week (0; 0%) Every day (18; 75%) 5 - 6x per week (5; 21%)

Appendix 165

5. How many km on average do you drive per year? Less than 6000 km (0; 0%) Between 24.000 - 30.000 km (4; 17%) Between 6000 - 12.000 km (0; 0%) Between 30.000 - 36.000 km (10; 42%) Between 12.000 - 18.000 km (0; 0%) More than 36.000 km per year (8; 33%) Between 18.000 - 24.000 km (2; 8%) 6. Which of the following driving styles do you think best describes

yours? Ecological (14; 58%) Aggressive (1; 4%) Conservative (6; 25%) Risky (1; 4%) Sportive (14; 58%) Depends on the situation (14; 58%) 7. Are you available for a short follow up interview (max. 20 Min) to gain

more ‘qualitative’ understanding? Yes (17; 71%) No (7; 29%) 8. What is your E-Mail address?

All data will be anonymzed, but for a better evaluation it would be beneficial to relate the data collected from the DriveGain application to the answers from this survey.

Thank you for your time and support!

166 Appendix

Appendix G: Eco-Driving Smartphone Application Post Interviews Interview Dates

Corporate Car Driver Date 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

March 22nd, 2012 March 26th, 2012 March 28th, 2012 April 03rd, 2012 April 12th, 2012 April 13th, 2012 April 16th, 2012 April 18th, 2012 April 20th, 2012 April 23rd, 2012 April 25th, 2012 April 28th, 2012 April 29th, 2012 April 30th, 2012 April 30th, 2012

Interview Questions A. DriveGain Application 1. What do you think was the greatest benefit of the application?

2. Besides the feedback you have received, which further feedback criteria are rel-

evant for you?

3. Do you think a monthly e-mail with a comparison of your FE to another driver would be useful?

i. If yes, what kind of comparison would you prefer?

4. Which other aspects would be relevant for you to use the application regularly in your company car?

B. Conceptual Aspects 5. How should the management roll out a fuel improvement/ CO2 reduction initia-

tive within the organization? i. What steps do you feel are crucial for implementation?

ii. Where do you see problems?

6. Any further remarks/ questions?

Appendix 167

All categorized answers from interviews:

Questions 1. Driver Level 2. Management Level 1. What do you think was the greatest benefit of the application?

Awareness: - Less about the feedback as such, more about using the application (15) - Raise awareness (13) - Sensitize (11) - Learn/ familiarize eco- driving practices (5) - Testing a new mobile technology (4)

2. Besides the feedback you have received, which further feedback criteria are relevant for you?

Feedback: - Real time feedback (15) - Should not distract from driving (12) - Transparent feedback (12) - If per voice, it should provide various driving tips (9) - Comparison with other colleagues (9) - More feedback related to eco-driving techniques (8) - Weekly feedback per email with summarized fuel consumption information (8) - Feedback has to show progress (8) - Individual feedback according to the driving type (7)

3. Do you think a monthly e-mail with a compari-son of your FE to another driver would be useful? If yes, what kind of comparison would you prefer?

- Yes (15) - Comparison with the same car model type (15) - Same amount of total km per year (14) - Transparent comparison (10) - Similar driver type (8) - The same job profile is not so important (3) - Same route (2)

168 Appendix

4. Which other aspects would be relevant for you to use the applica-tion regularly in your corporate car?

Time importance: - Everyday on the road (15) - Important to be home fast (15) - Many customer visits per day (10) - Drive faster if no speed regulations (8) Incentives: - High influence financial incentives (12) - Incentives through deciding to get a car with lower CO2 emission (8) - Becoming the eco-driver of the month (4)

Provide Information: - Provide personalized information (14) - Expose fuel costs (13) - Expose fuel efficiency (9) - CO2 Emission (6) Rewards: - Realistic/ fair rewards (12) - Bonus point program (9) - Bonus Points difficult (4)

5. How should the management roll out a fuel im-provement/ CO2 reduction initia-tive within the organization? i) What steps do you feel are cru-cial for imple-mentation? ii) Where do you see problems?

Punishment/ Control vs. Awareness: - No punishment (14) - Unfair to punish drivers who need their car for daily business (13) - No control from management (12) - Sensitize corporate car drivers, since no understanding of any fuel consumption (comparison with mobile phone) (9) Goal setting - Goal setting necessary (12) - Realistic goals related to corporate car drivers (11) - Goals enhance motivation (7) - Goals not important (3) Driving/ Routes - Consider different routes (9) - Depends on drivers driving behavior, sometimes no further reductions are possible (7)

6. Any further remarks/ ques-tions?

- What are the next steps (12) - Application updated to our needs (8)

Appendix 169

Appendix H: Monthly Personal Car Consumption Email with discrete Scale

FOR INTERNAL USE ONLY

Your personal Car Consumption and EcoDriving Tips

Dear Name of Driver, Please review your personal fuel emissions. Your average fuel consumption per month (liters per 100 km) until end of October 2011, compared to the average consumption rec-ommended by your car manufacturer, is in the middle area:

- Average fuel consumption: 8.5 Liters per 100 km (difference of -0.2 Liters compared to the previous month) - Difference of 19% between recommended and actual fuel consumption - Average kilometers driven: 3025 km per month - Average CO2 emissions: 168 gram CO2 per km

Recommendations, how your fuel efficiency can be improved through eco-driving: 1. Accelerate quickly and upshift to next gear by latest 2500 rpm (petrol) or 2000 rpm (diesel) (applicable for manual gear only); 2. Downshift as late as possible (applicable for manual gear only); 3. Drive with the highest gear as possible (applicable for manual gear only); 4. Drive anticipatory and smooth: do not accelerate aggressively and maintain constant speed (applicable for manual and automatic gears); 5. Use cruise mode when possible (applicable for automatic gear only); 6. Check your tire pressure monthly (applicable for manual and automatic gears). Best regards, Johannes SAP Research Note: Due to the low average fuel consumption figures stated by car manufacturers and in order to improve the comparison between the drivers, the categories in the graphic above are based on average fuel consumption values from eligible SAP (Switzerland) AG corpo-rate car and car allowance drivers. These drivers were randomly selected for the field test. The consumption rates are taken from the data received from the petrol cards, are treated highly confidential and are calculated individually for each driver. Please contact Johan-nes Tulusan for any further questions or recommendations.

Unsubscribe | Subscribe | Copyright/Trademark | Privacy | Impressum

Appendix I: Environmental Behavior Survey

https://www.sap.com/profile/unsubscribe.epx

http://www.sap.com/mk/get/g_basic_un_subscribe

http://www.sap.com/company/legal/copyright/index.epx

http://www.sap.com/company/legal/privacy.epx

http://www.sap.com/company/legal/impressum.epx

170 Appendix

I. Environmental Behavior / Umweltbewusstsein

Items 1 Strongly disagree … 7 Strongly Agree 1. To drive sustainable is an important part of my personality. = 5.50

2. Through sustainable driving, I would like to show to others that I am an ecologically-aware person. = 4.66

3. I like to talk about my fuel efficiency. = 3.65 4. I discuss sustainable driving tips and new technologies (e.g. electric vehicles) with others = 4.48

5. In hotels I let my towels to be changed every day. = 5.99 6. If I wear a shirt for one day, it will always be washed. = 4.95 7. At home when I leave the room, I switch off the light. = 6.14 8. At home I leave electrical equipment, such as TV, printer, coffee machine, etc., running on standby mode. = 5.26

9. In winter it is warm enough in my flat / house to wear a T-Shirt. = 5.33

Total Average = 5.11 II. Technological Affinity (= Technologieaffinität) Please evaluate your perception of statements 1 to 5 by choosing one of the answer options: Items 1 Strongly disagree … 7 Strongly Agree 1. My friends and colleagues often ask my opinion about new technologies (e.g. smartphones). = 4.62

2. My friends and colleagues usually know more about new technologies than I do. = 3.42

3. I am always up to date with new technological improvements which interest me. = 4.92

4. I have fun to test new technologies. = 5.44 5. For me technology is only a means to an end (= dient nur als Mittel zum Zweck). = 4.19

Total Average = 4.52

References 171

References 1. Agostinelli, G., Brown, J.M. and Miller, W.R. (1995), “Effects of normative

feedback on consumption among heavy drinking college students,” Journal of drug education, Vol. 25 No. 1, pp. 31–40.

2. Aitken, R.C. (1969), “Measurement of feelings using visual analogue scales.,” Proceedings of the royal society of medicine, Vol. 62 No. 10, p. 989.

3. Ajzen, I. (1991), “The theory of planned behavior,” Organizational behavior and human decision processes, Vol. 50 No. 2, pp. 179–211.

4. Ajzen, I. and Fishbein, M. (1977), “Attitude-behavior relations: A theoretical analysis and review of empirical research.,” Psychological bulletin, Vol. 84 No. 5, p. 888.

5. Alder, G.S. (2007), “Examining the relationship between feedback and per-formance in a monitored environment: A clarification and extension of feed-back intervention theory,” The Journal of High Technology Management Re-search, Vol. 17 No. 2, pp. 157–174.

6. Alder, G.S. and Tompkins, P.K. (1997), “Electronic performance monitoring: An organizational justice and concertive control perspective.,” Management Communication Quarterly, Vol. 10, pp. 259–288.

7. Amabile, T.M. (1993), “Motivational synergy: Toward new conceptualiza-tions of intrinsic and extrinsic motivation in the workplace,” Human Resource Management Review, Vol. 3 No. 3, pp. 185–201.

8. Angel, N.F. (1998), “Evaluating employees by computer,” Personnel Adminis-trator, Vol. 34 No. 11, pp. 67–72.

9. Arroyo, E., Sullivan, S. and Selker, T. (2006), “CarCoach: a polite and effec-tive driving coach.,” Proceedings of the CHI, Presented at the Conference on Human Factors in Computing Systems, pp. 357–362.

10. Baeriswyl, M., Przepiorka, W. and Staake, Thorsten. (2011), “Identifying in-dividuals’ preferences using games: A field experiment in promoting sustaina-ble energy consumption,” Proceedings of International Conference on Infor-mation Systems. from https://edit.ethz.ch/im/people/bmichi/Baeriswyl-Identifying_Preferences_using_Games_2011.pdf

11. Bahnmüller, K. (2012), “Ein Jahr der Freude,” Schweizer Fahrzeugmarkt, Vol. 12 No. 1, p. 40.

12. Balcazar, F., Hopkins, B.L. and Suarez, Y. (1986), “A critical, objective re-view of performance feedback,” Applied Psychology, Vol. 66, pp. 314–323.

13. Bamberg, S., Ajzen, I. and Schmidt, P. (2003), “Choice of travel mode in the theory of planned behavior: The roles of past behavior, habit, and reasoned ac-tion,” Basic and applied social psychology, Vol. 25 No. 3, pp. 175–187.

14. Bamberg, S. and Schmidt, P. (2003), “Incentives, morality, or habit? Predict-ing students’ car use for university routes with the models of Ajzen, Schwartz, and Triandis,” Environment and behavior, Vol. 35 No. 2, pp. 264–285.

172 References

15. Baron, J. (2008), “Normative Models of Judgment and Decision Making,” pp. 19–36.

16. Belz, A. and Kow, E. (2011), “Discrete vs. continuous rating scales for lan-guage evaluation in nlp,” Proceedings of the 49th Annual Meeting of the Asso-ciation for Computational Linguistics (ACL-HLT’11). Retrievedfrom http://www.newdesign.aclweb.org/anthology-new/P/P11/P11-2040.pdf

17. Bernstein, I.H. and Nunnally, J.C. (1994), Psychometric theory, New York: McGraw-Hill.

18. Beswick, D. (2007), “Intrinsic Motivation and Extrinsic Rewards.” Retrieved September 18, 2012, from http://www.beswick.info/psychres/management.htm

19. Black, T.R. (1999), Doing Quantitative Research in the Social Sciences: An Integrated Approach to Research Design, Measurement and Statistics, Sage Publications Ltd, 1sted.

20. Boriboonsomsin, K., Vu, A. and Barth, M. (2010), “Eco-Driving: Pilot Evalu-ation of Driving Behavior Changes among US Drivers,” Faculty Research, University of California Transportation Center, UC Berkeley, pp. 17–26.

21. Bostrom, R. and Heinen, J.S. (1977), “Problems and Failures: A Socio-Technical Perspective,” MIS quarterly, Vol. 1 No. 3.

22. Burton-Jones and Gallivan. (2008), “Toward a Deeper Understanding of Sys-tem Usage in Organizations: A Multilevel Perspective,” MIS quarterly, Vol. 31 No. 4, pp. 657–679.

23. Cengel, Y.A. and Boles, M.A. (2002), Thermodynamics: An Engineering Ap-proach, McGraw-Hill, New York.

24. Chapnick, N. (2007), “Hypermiling: Quest for Ultimate Fuel Economy.” Re-trieved January 26, 2012, from http://www.edmunds.com/cars/fueleconomy/articles/120880/article.html

25. Cialdini, R.B. (2003), “Crafting normative messages to protect the environ-ment,” Current directions in psychological science, Vol. 12 No. 4, pp. 105–109.

26. Cialdini, R.B., Kallgren, C.A. and Reno, R.R. (1991), “A focus theory of nor-mative conduct: A theoretical refinement and reevaluation of the role of norms in human behavior,” Advances in experimental social psychology, Vol. 24 No. 20, pp. 1–243.

27. Climate Company. (2011), “Environmental zones in Germany.” Retrieved January 26, 2012, from http://www.umwelt-plakette.de/int_england.php?SID=js9eb7lu4kg3kcssqi8vq1ijq4

28. Corbett, A.T. and Anderson, J.R. (2001), “Locus of feedback control in com-puter-based tutoring: Impact on learning rate, achievement and attitudes,” Proceedings of the SIGCHI conference on Human factors in computing sys-tems, pp. 245–252.

References 173

29. Cornford, T. and Smithson, S. (2006), Project Research in Information Sys-tems: A Student’s Guide, Palgrave Macmillan.

30. Darby, S. (2006), The effectiveness of feedback on energy consumption: a re-view for DEFRA of the literature on metering, billing and direct displays., pp. 1–21. Retrievedfrom http://www.eci.ox.ac.uk/research/energy/electric-metering.php

31. Dataforce. (2011), “Fuhrparkmanagement in deutschen Unternehmen.” Re-trieved January 26, 2012, from http://www.dataforce.de/web/guest/infocenter

32. Deci, E.L. and Ryan, R.M. (1985), Intrinsic Motivation and Self-Determination in Human Behavior, Springer, 1sted.

33. Department for Transport, G.M.H. (2011), “Transport Statistics Great Britain 2011,” Web page, . Retrieved September 19, 2012, from http://www.dft.gov.uk/statistics/releases/transport-statistics-great-britain-2011/

34. Deslauriers, B.C. and Everett, P.B. (1977), “Effects of intermittent and contin-uous token reinforcement on bus ridership.,” Journal of Applied Psychology, Vol. 62 No. 4, p. 369.

35. Donmez, B. (2007), “Safety implications of providing real-time feedback to distracted drivers,” Accident analysis and prevention, Vol. 39 No. 3, p. 581.

36. Doppelt, B., Markowitz, E.M. and Initiative, C.L. (2009), “Reducing Green-house Gas Emissions Through Behavioral Change: An Assessment of Past Re-search On Energy Use, Transportation and Water Consumption.”

37. DriveGain Ltd. (2012), “DriveGain.” Retrieved September 19, 2012, from http://drivegain.com/about

38. DTLR. (2001), “Department for Transport: Transport statistics 2000.” Re-trieved January 26, 2012, from http://dftsearch.dft.gov.uk/client=default_frontend&proxyreload=1&proxystylesheet=default_frontend&output=xml_no_dtd&q=transport+statistics+2000&pg=1&Search.x=0&Search.y=0

39. Dula, C.S. and Geller, E.S. (2003), “Risky, aggressive, or emotional driving: Addressing the need for consistent communication in research,” Journal of Safety Research, Vol. 34 No. 5, pp. 559–566.

40. EcoDrive. (2012), “The Golden Rules of EcoDriving.” Retrieved January 26, 2012, from http://www.ecodrive.org/en/what_is_ecodriving-/benefits_of_ecodriving/

41. Eisenhardt, K.M. (1989), “Building theories from case study research,” Acad-emy of management review, Vol. 14 No. 4, pp. 532–550.

42. Enerdata. (2011), “World Energy Statistics | World Energy Consumption & Stats| Enerdata.” Retrieved January 26, 2012, from http://yearbook.enerdata.net/

43. European Commission. (2010), “Energy: Energy Efficiency Directive - Euro-pean commission.” Retrieved January 26, 2012, from http://ec.europa.eu/energy/efficiency/eed/eed_en.htm

174 References

44. Flick, U. (2006), An Introduction to Qualitative Research, Sage Publications Ltd, Third Edition.

45. Fogg, B.J. (2002), “Persuasive technology: using computers to change what we think and do,” Ubiquity, Vol. December, p. 5.

46. Formosa, D. (2009), “Ford SmartGauge: Designing an Extra 9 MPG?,” Smart Design.,from www.smartdesignworldwide.com/pdf/SmartDesign_SmartGauge.pdf

47. Forward, S. (2004), “The prediction of travel behaviour using the Theory of Planned Behaviour,” International Conference of Traffic and Transport Psy-chology.

48. Foxx, R. and Hake, D. (1977), “Gasoline conservation: A procedure for meas-uring and reducing the driving of college students,” Journal of Applied Behav-ior Analysis, Vol. 10 No. 1, p. 61.

49. Foxx, R. and Schaeffer, M.H. (1981), “A company-based lottery to reduce the personal driving of employees,” Journal of Applied Behavior Analysis, Vol. 14 No. 3, p. 273.

50. Freyd, M. (1923), “A graphic rating scale for teachers,” The Journal of Educa-tional Research, Vol. 8 No. 5, pp. 433–439.

51. Froberg, D.G. and Kane, R.L. (1989), “Methodology for measuring health-state preferences—II: Scaling methods,” Journal of clinical epidemiology, Vol. 42 No. 5, pp. 459–471.

52. Froehlich, J., Dillahunt, T., Klasnja, P., Mankoff, J., Consolvo, S., Harrison, B. and Landay, J.A. (2009), “UbiGreen: investigating a mobile tool for track-ing and supporting green transportation habits,” Proceedings of the 27th inter-national conference on Human factors in computing systems, pp. 1043–1052.

53. Froehlich, J., Findlater, L. and Landay, J. (2010), “The design of eco-feedback technology,” Proceedings of the 28th international conference on Human fac-tors in computing systems, pp. 1999–2008.

54. Frost & Sullivan. (2009), 2010 Outlook & Forecast: Mobile & Wireless Communications.

55. Fujii, S. and Gärling, T. (2003), “Application of attitude theory for improved predictive accuracy of stated preference methods in travel demand analysis,” Transportation Research Part A: Policy and Practice, Vol. 37 No. 4, pp. 389–402.

56. Gardner, G.T. and Stern, P.C. (2008), “The short list: The most effective ac-tions US households can take to curb climate change,” Environment: Science and Policy for Sustainable Development, Vol. 50 No. 5, pp. 12–25.

57. Gartner. (2010), Forecast: Mobile Devices, Worldwide 2003-2014, 2Q10 Up-date.

58. Gartner. (2011), “Smartphone Sales Increased 42 Percent.” Retrieved January 26, 2012, from http://www.gartner.com/it/page.jsp?id=1848514

References 175

59. Geller, E.S., Winett, R.A., Everett, P.B. and Winkler, R.C. (1982), Preserving the environment: New strategies for behavior change, Pergamon Press Elms-ford, New York.

60. German Federal Environmental Agency. (2009), Ratgeber: Sprit sparen und mobil sein ( No. 2), p. 19.

61. Goldenbeld, C., Twisk, D. and Houwing, S. (2008), “Effects of persuasive communication and group discussions on acceptability of anti-speeding poli-cies for male and female drivers,” Transportation Research Part F: Traffic Psychology and Behaviour, Vol. 11 No. 3, pp. 207–220.

62. Goldstein, N.J., Cialdini, R.B. and Griskevicius, V. (2008), “A room with a viewpoint: Using social norms to motivate environmental conservation in ho-tels,” Journal of Consumer Research, Vol. 35 No. 3, pp. 472–482.

63. Graham, J., Koo, M. and Wilson, T.D. (2011), “Conserving energy by induc-ing people to drive less,” Journal of Applied Social Psychology, Vol. 41 No. 1, pp. 106–118.

64. Graml, T., Loock, C., Baeriswyl, M. and Staake, T. (2011), “Improving Resi-dential Energy Consumption at Large Using Persuasive Systems.,” Presented at the . 19th European Conference on Information Systems (ECIS), Helsinki, Finland.

65. Graml, T., Loock, C., Staake, T. and Fleisch, E. (2010), “Online Energy Be-havior Change Systems: What makes people stick? pdf free download,” Pre-sented at the Persuasive Conference, Copenhagen.

66. Grant, R. and Higgins, C. (1989), “Monitoring service workers via computer: The effect on employees, productivity, and service,” National Productivity Review, Vol. 8 No. 2, pp. 101–112.

67. Greengard, S. (1996), “Privacy: Entitlement or illusion?,” Personnel Journal, Vol. 75 No. 5, pp. 74–76.

68. GreenRoad. (2008), “Is Safe Driving More Economical? Driver Safety and Fuel Consumption.” Retrieved May 29, 2011, from http://www.greenroadtech.com/documents/driver_safety_and_fuel_consumption_us.pdf

69. Henriques, V.E. (1996), “In defence of computer monitoring,” Training, Vol. 75 No. 5, pp. 74–76.

70. Holmes, T.G. (2007), “Eco-visualization: combining art and technology to reduce energy consumption,” Proceedings of the 6th ACM SIGCHI conference on Creativity & cognition, C&;C ’07, New York, NY, USA, ACM, pp. 153–162. doi:10.1145/1254960.1254982

71. Van Houwelingen, J.H. and Van Raaij, W.F. (1989), “The effect of goal-setting and daily electronic feedback on in-home energy use,” Journal of con-sumer research, pp. 98–105.

72. Hussey, J. and Hussey, R. (1997), Business Research: A Practical Guide for Undergraduate and Postgraduate Students, Macmillan.

176 References

73. Hutton, K.A., Sibley, C.G., Harper, D.N. and Hunt, M. (2001), “Modifying driver behaviour with passenger feedback,” Transportation Research Part F: Traffic Psychology and Behaviour, Vol. 4 No. 4, pp. 257–269.

74. Johansson, H., Gustafsson, P., Henke, M. and Rosengren, M. (2003), “Impact of EcoDriving on emissions,” Transport and Air Pollution. Proceedings from the 12th Symposium, pp. 97–105.

75. Keen, P.G.W. (1981), “Information systems and organizational change,” Commun. ACM, Vol. 24 No. 1, pp. 24–33.

76. King, G., Keohane, R.O. and Verba, S. (1994), Designing social inquiry: sci-entific inference in qualitative research, Princeton: Princeton University Press.

77. Kluger, A.N. and DeNisi, A. (1996a), “The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feed-back intervention theory.,” Psychological bulletin, Vol. 119 No. 2, p. 254.

78. Kluger, A.N. and DeNisi, A. (1996b), “The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feed-back intervention theory.,” Psychological bulletin, Vol. 119 No. 2, p. 254.

79. Lawson, T. (2008), Introduction to Critical Realism and Social Science: Con-text and Contribution., Presented at the MP0-2 Lecture, Judge Business School, Cambridge University.

80. Lee, H., Lee, W. and Lim, Y.K. (2010), “The effect of eco-driving system to-wards sustainable driving behavior,” Proceedings of the 28th of the interna-tional conference extended abstracts on Human factors in computing systems, pp. 4255–4260.

81. Loock, C., Graml, T., Baeriswyl, M. and Staake, T. (2011), “How to Motivate Energy Efficiency Online,” Presented at the Proceedings of 20 th International Conference on Management of Technology, Florida, USA.

82. Loock, C., Staake, T and Landwehr, J. (2011), “Green IS Design and Energy Conservation: An Empirical Investigation of Social Normative Feedback,” Presented at the International Conference on Information Systems (ICIS), Shanghai, China.

83. Mankoff, Jennifer, Matthews, D., Fussell, S. and Johnson, M. (2007), “Lever-aging Social Networks To Motivate Individuals to Reduce their Ecological Footprints,” IEEE, pp. 87–87. doi:10.1109/HICSS.2007.325

84. Marshall, C. and Rossman, G.B. (1989), Designing qualitative research, Sage Publications.

85. Meschtscherjakov, A., Wilfinger, D., Scherndl, T. and Tscheligi, M. (2009), “Acceptance of future persuasive in-car interfaces towards a more economic driving behaviour,” Proceedings of the 1st International Conference on Auto-motive User Interfaces and Interactive Vehicular Applications, pp. 81–88.

86. Van Mierlo, J., Maggetto, G., Van de Burgwal, E. and Gense, R. (2004), “Driving style and traffic measures-influence on vehicle emissions and fuel

References 177

consumption,” Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, Vol. 218 No. 1, pp. 43–50.

87. Mingers, J. (2001), “Combining IS research methods: towards a pluralist methodology,” Information systems research, Vol. 12 No. 3, pp. 240–259.

88. Morgan Stanley. (2010), “Internet Trends,” Presented at the CM Summit, New York City.

89. Mountain, D. (2008), Real-Time Feedback and Residential Electricity Con-sumption: British Columbia and Newfoundland and Labrador Pilots, Ontario, Canada: Mountain Economic Consulting and Associates Inc.

90. National Research Council. (2010), Assessment of Technologies for Improving Light Duty Vehicle Fuel Economy ( No. Traffic Safety).

91. Newnam, S., Watson, R.T. and Murray, W. (2004), “Factors predicting inten-tions to speed in a work and personal vehicle,” Transportation Research Part F: Traffic Psychology and Behaviour, Vol. 7 No. 4-5, pp. 287–300.

92. Nussbaum, K. and duRivage, V. (1986), “Computer monitoring: Mismanage-ment by remote control,” Business and Society Review, Vol. 56, pp. 16–20.

93. Onoda, T. (2009), “IEA policies–G8 recommendations and an afterwards,” Energy Policy, Vol. 37 No. 10, pp. 3823–3831.

94. Orlikowski, W.J. and Baroudi, J.J. (1991), “Studying information technology in organizations: Research approaches and assumptions,” Information systems research, Vol. 2 No. 1, pp. 1–28.

95. Pace, T., Ramalingam, S. and Roedl, D. (2007), “Celerometer and idling re-minder: persuasive technology for school bus eco-driving,” CHI’07 extended abstracts on Human factors in computing systems, pp. 2085–2090.

96. Polk, M. (2003), “Are women potentially more accommodating than men to a sustainable transportation system in Sweden?,” Transportation Research Part D: Transport and Environment, Vol. 8 No. 2, pp. 75–95.

97. Reno, R.R., Cialdini, R.B. and Kallgren, C.A. (1993), “The transsituational influence of social norms.,” Journal of personality and social psychology, Vol. 64 No. 1, p. 104.

98. Romm, J.J. and Frank, A.A. (2006), “Hybrid vehicles gain traction,” Scientific American, Vol. 294 No. 4, pp. 72–79.

99. Rotter, J.B. (1966), “Generalized expectancies for internal versus external control of reinforcement,” Psychological monographs, Vol. 80 No. 1, pp. 1–28.

100. Schultz, P., Nolan, J.M., Cialdini, R.B., Goldstein, N.J. and Griskevicius, V. (2007), “The constructive, destructive, and reconstructive power of social norms,” Psychological Science, Vol. 18 No. 5, p. 429.

101. Schwartz, B., Ward, A., Monterosso, J., Lyubomirsky, S., White, K. and Leh-man, D.R. (2002), “Maximizing versus satisficing: Happiness is a matter of

178 References

choice.,” Journal of Personality and Social Psychology, Vol. 83 No. 5, p. 1178.

102. Scott, W.R. and Huskisson, E.C. (1977), “Measurement of functional capacity with visual analogue scales,” Rheumatology, Vol. 16 No. 4, pp. 257–259.

103. Shipworth, M. (2000), Motivating Home Energy Action: A handbook of what works, Canberra, Australia: Australian Greenhouse Office.

104. Siero, S., Boon, M., Kok, G. and Siero, F. (1989), “Modification of driving behavior in a large transport organization: A field experiment.,” Journal of Applied Psychology, Vol. 74 No. 3, pp. 417–423.

105. Simon, H.A. (1955), “A behavioral model of rational choice,” The quarterly journal of economics, Vol. 69 No. 1, p. 99.

106. Smiley, A. and Michon, J.A. (1989), “Conceptual Framework for generic in-telligent driver support.,” Traffic Research Centre, Vol. University of Gro-ningen Haren, Netherlands, p. 37.

107. Stead, M., Tagg, S., MacKintosh, A.M. and Eadie, D. (2005), “Development and evaluation of a mass media Theory of Planned Behaviour intervention to reduce speeding,” Health Education Research, Vol. 20 No. 1, p. 36.

108. Steg, L. and Sievers, I. (2000), “Cultural theory and individual perceptions of environmental risks,” Environment and Behavior, Vol. 32 No. 2, p. 250.

109. Stenner, F. (2009), Handbuch Automobilbanken: Finanzdienstleistungen für Mobilität, Gabler Wissenschaftsverlage.

110. Stoppani, S. (2009), “Bliss TrekTM,” Overview of the BlissTrek Application. Retrieved August 5, 2012, from http://blisstrek.com/

111. Strauss, A.C. and Corbin, J. (1998), Basics of Qualitative Research: Second Edition: Techniques and Procedures for Developing Grounded Theory, Sage Publications, Inc, 2nded.

112. Svensson, E. (2000), “Comparison of the quality of assessments using contin-uous and discrete ordinal rating scales,” Biometrical Journal, Vol. 42 No. 4, pp. 417–434.

113. Tasca, L. (2000), “A review of the literature on aggressive driving research,” Presented at the Aggressive driving issues conference, p. 25.

114. Tester, J., Fogg, B.J. and Maile, M. (2000), “CommuterNews: a prototype of persuasive in-car entertainment,” CHI’00 extended abstracts on Human fac-tors in computing systems, pp. 24–25.

115. Thambusamy, R. and Salam, A.F. (2010), “Corporate Ecological Responsive-ness, environmental ambidexterity and IT-enabled environmental sustainabil-ity strategy,” Presented at the International Conference on Information Sys-tems, Lima.

116. Thogersen, J. and Moller, B. (2008), “Breaking car use habits: The effective-ness of a free one-month travelcard,” Transportation, Vol. 35 No. 3, pp. 329–345.

References 179

117. Toulmin, S.E. (1960), The Philosophy of Science: An Introduction, Harper & Row, Harper torchbooks.

118. Tversky, A. and Kahneman, D. (1986), “Rational choice and the framing of decisions,” Journal of business, pp. 251–278.

119. Ullmenstein, T. (2012,September13), “Ökologie wie Ökonmie ins Spiel brin-gen,” Neu Zürcher Zeitung, p. 11.

120. Van Velsor, E., Leslie, J.B. and Fleenor, J.W. (1997), Choosing 360. A Guide to Evaluating Multi-Rater Feedback Instruments for Management Develop-ment., Center for Creative Leadership, Greensboro,.

121. Van der Voort, M., Dougherty, M.S. and Van Maarseveen, M. (2001), “A pro-totype fuel-efficiency support tool,” Transportation Research Part C: Emerg-ing Technologies, Vol. 9 No. 4, pp. 279–296.

122. Wahlberg, A. (2007), “Aggregation of driver celeration behavior data: Effects on stability and accident prediction,” Safety science, Vol. 45 No. 4, pp. 487–500.

123. Watson, R.T., Boudreau, M.C. and Chen, A.J. (2010), “Information systems and environmentally sustainable development: energy informatics and new di-rections for the IS community,” MIS quarterly, Vol. 34 No. 1, p. 23.

124. Williams, M. (2000), “Interpretivism and Generalisation,” Sociology, Vol. 34 No. 2, pp. 209–224.

125. Winkler, R.C. and Winett, R.A. (1982), “Behavioral interventions in resource conservation: A systems approach based on behavioral economics.,” American Psychologist, Vol. 37 No. 4, p. 421.

126. Worsnop, R.L. (1993), Privacy in the workplace, CQ Press. 127. Yin, R.K. (2008), Case Study Research: Design and Methods, SAGE Publica-

tions. 128. De Young, R. (1985), “Encouraging environmentally appropriate behavior:

The role of intrinsic motivation,” Journal of Environmental Systems, Vol. 15 No. 4, pp. 281–292.

129. Zahorsky, D. (2010), “Ignite the Two Fires of Employee Motivation.” Re-trieved September 18, 2012, from http://sbinformation.about.com/cs/benefits/a/033003.htm

180 References

Publications

Tulusan, J., Steggers, H., Staake, T., Fleisch, E., Supporting eco-driving with eco-feedback technologies: Recommendations targeted at improving corporate car driv-ers’ intrinsic motivation to drive more sustainable, Energy Informatics 2012 (EI 2012), Atlanta, Georgia, United States, October 2012.

Tulusan, J., Staake, T., Fleisch, E., Direct or indirect sensor enabled eco-driving feedback: Which preference do corporate car drivers have?, Internet of Things 2012 – Third International Conference on the Internet of Things (IoT 2012), Wuxi, P.R. China, October 2012.

Tulusan, J., Staake, T., Fleisch, E., Providing eco-driving feedback to corporate car drivers: what impact does a smartphone application have on their fuel efficiency, 14th ACM International Conference on Ubiquitous Computing (UbiComp), Pitts-burgh, Pennsylvania, United States, September 2012.

Tulusan, J., Soi, L., Paefgen, J., Staake, T., Eco-efficient feedback technologies: Which eco-feedback types prefer drivers most?. World of Wireless, Mobile and Multimedia Networks IEEE Conference, Lucca, Italy, June 2011

Tulusan, J., Brogle, M., Staake, T., Fleisch, E., Becoming a Sustainable Driver: The Impact of Mobile Feedback Devices. 4th ERCIM eMobility WWIC Conference, Lulea, Sweden, June 2010

Curriculum Vitae 181

Curriculum Vitae Personal Data Johannes Tulusan Born 29th of October 1979 in Erlangen (Germany)

Education 01/2010 – 12/2012 University of St. Gallen Institute of Technology Management, St.Gallen (CH) PhD Student in Management

10/2008 – 12/2009 Cambridge University Judge Business School, Cambridge (UK) PhD Candidate in Management (First year)

09/2006 – 09/2007 London School of Economics & Political Science IS and Innovation Department, London (UK) MSc ADMIS (Analysis, Design and Management of Information Systems)

09/2000 – 06/2003 Brunel University Management Department, London (UK)

BSc (Hons) Business and Computing

09/1994 – 06/1999 Richard-Wagner-Secondary School Bayreuth (Germany)

Graduated with German Abitur Work Experience 01/2013 – now SAP AG, Munich / Walldorf (Germany) Business Development Manager (full time) Strategic Business Development Department

01/2010 – 12/2012 SAP AG, Zurich (Switzerland) Research Associate (full time) SAP Research Switzerland

182 Curriculum Vitae

10/2007 – 09/2008 Viewmy.TV Ltd., London (UK) Head of Business Development (full time) Headquarter London

03/2004 – 03/2006 SAP AG, Walldorf (Germany) Business Development Specialist (full time) Advisory Office EMEA

01/2004 – 02/2004 Siemens Medical Services, Philadelphia (USA) Healthcare IT Consulting Internship ‘Sorian’ Consulting Department

08/2003 – 12/2003 SAP AG, Walldorf (Germany) Technology Consulting Internship Technology Consulting II Department

06/2002 – 09/2002 Siemens Business Services, Singapore Business and IT Consulting Internship Application Hosting Department

07/2001 – 09/2001 Suse Linux Ltd., Borehamwood (United Kingdom) Sales Internship – Sales and Marketing Office

07/1998 – 10/2000 Courissima Ltd., Bayreuth (Germany) Co-founder and Business Development Manager (full

time) Awards and Scholarships 01/2010 – 12/2012 PhD funded from the cooperation bursary of SAP

and canton St. Gallen. 10/2008 – 12/2009 Economic and Social Research Council Full Scholar-

ship.

10/2008 – 12/2009 Cambridge European Trust Scholarship.

10/2007 MSc dissertation result award.

Author: Mr. Johannes Tulusan Title: Combining ICT with Eco-driving Concepts to Improve Corporate Car Drivers’ Fuel Efficiency

combining ict with eco-driving concepts to …...combining ict with eco-driving concepts to improve...

Documents