The Internet of Things (IoT): Applications,investments, and challenges for enterprises

In Lee a,*, Kyoochun Lee b

a School of Computer Sciences, Western Illinois University, Stipes Hall 442F, Macomb, IL 61455-1390, U.S.A.bOlin Corporation, Clayton, MO, U.S.A.

Business Horizons (2015) 58, 431—440

Available online at www.sciencedirect.com

ScienceDirectwww.elsevier.com/locate/bushor

KEYWORDSCloud computing;Internet of Things;Radio frequencyidentification;Real options;Supply chainmanagement

Abstract The Internet of Things (IoT), also called the Internet of Everything or theIndustrial Internet, is a new technology paradigm envisioned as a global network ofmachines and devices capable of interacting with each other. The IoT is recognized asone of the most important areas of future technology and is gaining vast attentionfrom a wide range of industries. This article presents five IoT technologies that areessential in the deployment of successful IoT-based products and services anddiscusses three IoT categories for enterprise applications used to enhance customervalue. In addition, it examines the net present value method and the real optionapproach widely used in the justification of technology projects and illustrates howthe real option approach can be applied for IoT investment. Finally, this articlediscusses five technical and managerial challenges.# 2015 Kelley School of Business, Indiana University. Published by Elsevier Inc. Allrights reserved.

1. The Internet of Things (IoT)

The Internet of Things (IoT), also called the Internetof Everything or the Industrial Internet, is a newtechnology paradigm envisioned as a global networkof machines and devices capable of interacting witheach other. The IoT is recognized as one of the mostimportant areas of future technology and is gainingvast attention from a wide range of industries.The true value of the IoT for enterprises can befully realized when connected devices are able tocommunicate with each other and integrate with

* Corresponding authorE-mail addresses: i-lee@wiu.edu (I. Lee),

kyoochun@gmail.com (K. Lee)

0007-6813/$ — see front matter # 2015 Kelley School of Business, Ihttp://dx.doi.org/10.1016/j.bushor.2015.03.008

vendor-managed inventory systems, customer sup-port systems, business intelligence applications, andbusiness analytics.

Gartner (2014) forecasts that the IoT will reach26 billion units by 2020, up from 0.9 billion in 2009,and will impact the information available to supplychain partners and how the supply chain operates.From production line and warehousing to retaildelivery and store shelving, the IoT is transformingbusiness processes by providing more accurate andreal-time visibility into the flow of materials andproducts. Firms will invest in the IoT to redesignfactory workflows, improve tracking of materials,and optimize distribution costs. For example, bothJohn Deere and UPS are already using IoT-enabledfleet tracking technologies to cut costs and improvesupply efficiency.

ndiana University. Published by Elsevier Inc. All rights reserved.

432 I. Lee, K. Lee

In addition to manufacturers’ adoption of the IoT,various service industries are in the process ofadopting the IoT to increase revenue through en-hanced services and become leaders in their mar-kets. Disney’s MagicBand is a new wristband withRFID chips that serves as a ticket and connects toDisney’s data repository regarding park visitors.Kroger’s new IoT-based system, Retail Site Intelli-gence, is one complete retail platform of videoanalytics, wireless devices, POS devices, handheldsensors, IP cameras, and video management soft-ware that was designed to help customers have abetter shopping experience by more easily findingthe products they want and saving time at checkout.

The adoption of this technology is rapidly gainingmomentum as technological, societal, and compet-itive pressures push firms to innovate and transformthemselves. As IoT technology advances and increas-ing numbers of firms adopt the technology, IoTcost-benefit analysis will become a subject of greatinterest. Because of the potential but uncertainbenefits and high investment costs of the IoT, firmsneed to carefully assess every IoT-induced opportu-nity and challenge to ensure that their resources arespent judiciously.

This article begins with a discussion of the fiveessential IoT technologies used for the deploymentof successful IoT-based products and services andidentifies three IoT categories for enterprise appli-cations. Then, it examines a net present value ap-proach and a real option approach widely used in thejustification of technology projects and discusseshow real option valuation can be applied to IoTinvestment. Finally, this article discusses five techni-cal and managerial challenges: data management,data mining, privacy, security, and chaos.

2. Essential IoT technologies

Five IoT technologies are widely used for the deploy-ment of successful IoT-based products and services:

1. radio frequency identification (RFID);

2. wireless sensor networks (WSN);

3. middleware;

4. cloud computing; and

5. IoT application software.

2.1. Radio frequency identification (RFID)

Radio frequency identification (RFID) allows auto-matic identification and data capture using radio

waves, a tag, and a reader. The tag can store moredata than traditional barcodes. The tag containsdata in the form of the Electronic Product Code(EPC), a global RFID-based item identification sys-tem developed by the Auto-ID Center. Three types oftags are used. Passive RFID tags rely on radio fre-quency energy transferred from the reader to thetag to power the tag; they are not battery-powered.Applications of these can be found in supply chains,passports, electronic tolls, and item-level tracking.Active RFID tags have their own battery supply andcan instigate communication with a reader. Activetags can contain external sensors to monitor temper-ature, pressure, chemicals, and other conditions.Active RFID tags are used in manufacturing, hospitallaboratories, and remote-sensing IT asset manage-ment. Semi-passive RFID tags use batteries to powerthe microchip while communicating by drawing pow-er from the reader. Active and semi-passive RFID tagscost more than passive tags.

2.2. Wireless sensor networks (WSN)

Wireless sensor networks (WSN) consist of spatiallydistributed autonomous sensor-equipped devices tomonitor physical or environmental conditions and cancooperate with RFID systems to better track thestatus of things such as their location, temperature,and movements (Atzori, Iera, & Morabito, 2010). WSNallow different network topologies and multihopcommunication. Recent technological advances inlow-power integrated circuits and wireless commu-nications have made available efficient, low-cost,low-power miniature devices for use in WSN applica-tions (Gubbi, Buyya, Marusic, & Palaniswami, 2013).

WSN have primarily been used in cold chainlogistics that employ thermal and refrigerated pack-aging methods to transport temperature-sensitiveproducts (Hsueh & Chang, 2010; White & Cheong,2012). WSN are also used for maintenance andtracking systems. For example, General Electricdeploys sensors in its jet engines, turbines, andwind farms. By analyzing data in real time, GE savestime and money associated with preventive main-tenance. Likewise, American Airlines uses sensorscapable of capturing 30 terabytes of data per flightfor services such as preventive maintenance.

2.3. Middleware

Middleware is a software layer interposed betweensoftware applications to make it easier for softwaredevelopers to perform communication and input/output. Its feature of hiding the details of differenttechnologies is fundamental to free IoT developersfrom software services that are not directly relevant

The Internet of Things (IoT): Applications, investments, and challenges for enterprises 433

to the specific IoT application. Middleware gainedpopularity in the 1980s due to its major role insimplifying the integration of legacy technologiesinto new ones. It also facilitated the development ofnew services in the distributed computing environ-ment. A complex distributed infrastructure of theIoT with numerous heterogeneous devices requiressimplifying the development of new applicationsand services, so the use of middleware is an idealfit with IoT application development. For example,Global Sensor Networks (GSN) is an open sourcesensor middleware platform enabling the develop-ment and deployment of sensor services with almostzero programming effort. Most middleware archi-tectures for the IoT follow a service-oriented ap-proach in order to support an unknown and dynamicnetwork topology.

2.4. Cloud computing

Cloud computing is a model for on-demand access toa shared pool of configurable resources (e.g., com-puters, networks, servers, storage, applications,services, software) that can be provisioned as In-frastructure as a Service (IaaS) or Software as aService (SaaS). One of the most important outcomesof the IoT is an enormous amount of data generatedfrom devices connected to the Internet (Gubbiet al., 2013). Many IoT applications require massivedata storage, huge processing speed to enable real-time decision making, and high-speed broadbandnetworks to stream data, audio, or video. Cloudcomputing provides an ideal back-end solution forhandling huge data streams and processing them forthe unprecedented number of IoT devices and hu-mans in real time.

2.5. IoT applications

The IoT facilitates the development of myriadindustry-oriented and user-specific IoTapplications.Whereas devices and networks provide physicalconnectivity, IoT applications enable device-to-device and human-to-device interactions in a reli-able and robust manner. IoT applications on devicesneed to ensure that data/messages have been re-ceived and acted upon properly in a timely manner.For example, transportation and logistics applica-tions monitor the status of transported goods suchas fruits, fresh-cut produce, meat, and dairy prod-ucts. During transportation, the conservation status(e.g., temperature, humidity, shock) is monitoredconstantly and appropriate actions are taken auto-matically to avoid spoilage when the connection isout of range. For example, FedEx uses SenseAwareto keep tabs on the temperature, location, and

other vital signs of a package, including when it isopened and whether it was tampered with along theway.

While device-to-device applications do not nec-essarily require data visualization, more and morehuman-centered IoT applications provide visualiza-tion to present information to end users in an intui-tive and easy-to-understand way and to allowinteraction with the environment. It is importantfor IoT applications to be built with intelligence sodevices can monitor the environment, identify prob-lems, communicate with each other, and potentiallyresolve problems without the need for humanintervention.

3. IoT applications to enhancecustomer value

Despite growing popularity of the IoT, few studieshave focused on categorization of the IoT for enter-prises (e.g., Chui, Loffler, & Roberts, 2010). Basedon the technology trends and literature review,this article identifies three IoT categories for enter-prise applications: (1) monitoring and control, (2)big data and business analytics, and (3) informationsharing and collaboration. Understanding how thesethree IoT categories can enhance the customervalue of an organization is a prerequisite to success-ful IoT adoption. This article next discusses thethree IoT categories, along with an illustration ofreal-world IoT applications developed to enhancecustomer value.

3.1. Monitoring and control

Monitoring and control systems collect data on equip-ment performance, energy usage, and environmentalconditions, and allow managers and automatedcontrollers to constantly track performance in realtime anywhere, anytime. Advanced monitoring andcontrol technologies such as smart grid and smartmetering reveal operational patterns, spot areas ofpotential improvement, or predict future outcomesand optimize operations, leading to lower costs andhigher productivity.

The smart home is known to be at the forefront ofinnovation regarding IoT monitoring and controlsystems. The primary value propositions are familyand property protection and energy savings. Forexample, the Verizon Home Monitoring and Controlnetwork uses a wireless communications technologydesigned specifically for remote control applica-tions in home automation. IoT-enabled home appli-ances and devices can be monitored and controlledoutside the user’s home through a computer, tablet,or smartphone. The Verizon Home Monitoring and

434 I. Lee, K. Lee

Control network allows users to adjust the lights,control the climate, manage the security system,receive automatic event notifications, and evenlock and unlock doors.

The IoT is also used to monitor and control variouscomponents in cars. The primary customer valuepropositions are drivers’ personalized experienceand satisfaction. Ford and Intel teamed up in 2014to explore new opportunities to personalize the userexperience using facial recognition software and amobile phone app. The joint research project,called Mobile Interior Imaging, incorporates per-ceptual computing technology to offer improvedprivacy controls and to identify different driversand automatically adjust features based on an in-dividual’s preferences. The in-car experience isthen personalized further by displaying informationspecific to the driver, such as his/her calendar, music,and contacts. The customer value propositionsare appropriately integrated into the connectedcar environment to provide another revenue streamfor Ford.

3.2. Big data and business analytics

IoT devices and machines with embedded sensorsand actuators generate enormous amounts of dataand transmit it to business intelligence and analyticstools for humans to make decisions. These data areused to discover and resolve business issues–—suchas changes in customer behaviors and marketconditions–—to increase customer satisfaction, andto provide value-added services to customers. Busi-ness analytics tools may be embedded into IoTdevices, such as wearable health monitoring sen-sors, so that real-time decision making can takeplace at the source of data.

The IoT and advances in business analytics nowmake it possible to capture vast amounts of individualhealth data. The IoT enables healthcare serviceproviders to personalize patient care. New IoT tech-nologies provide data about a patient’s everydaybehaviors and health, creating opportunities for careproviders to influence patients far more frequentlyand effectively. For example, Humana’s HealthsenseeNeighbor1 remote monitoring system reportschanges in the member’s normal patterns of move-ment and activity to Humana care managers–—via in-home sensors that measure routine daily activitieswith data analytics–—to help trigger interventions andhelp prevent adverse events from escalating to emer-gency room visits or hospital stays.

IoT-based big data are also transforming thehealthcare product industry. For example, Proctor& Gamble developed the Oral-B Pro 5000 interactiveelectric toothbrush to provide users with a smarter,

more personalized oral care routine. The interactiveelectric toothbrush records brushing habits with mo-bile technology while giving mouth-care tips along-side news headlines. This innovation provides userswith unprecedented control over their oral care.Tests of the interactive electric toothbrush haveshown that when connected, brushing time increasesfrom less than 60 seconds with a manual toothbrushto 2 minutes and 16 seconds with an electric tooth-brush, surpassing the 2-minute session recommendedby dental professionals.

3.3. Information sharing and collaboration

Information sharing and collaboration in the IoT canoccur between people, between people and things,and between things. Sensing a predefined eventis usually the first step for information sharingand collaboration. In the supply chain area, infor-mation sharing and collaboration enhance situa-tional awareness and avoid information delay anddistortion. For example, if sensors are placedthroughout a retail store where refrigeration is nec-essary, alerts can be sent to the store manager’smobile device whenever the refrigerators malfunc-tion. The manager can then check the employeestatus report to see who is available and sendtask assignments to that employee via his or herIoT-enabled mobile device.

To enhance information sharing and collaborationwith shoppers, Macy’s is deploying shopkick’s shop-Beacon technology, an enhanced mobile location-based technology that uses ultrasound BluetoothLow Energy (BLE). ShopBeacon provides shopkickapp users with personalized department-leveldeals, discounts, recommendations, and rewards.As shoppers enter Macy’s, shopBeacon remindsthose shopkick app users who have opted in. Thisenhancement in Macy’s information sharing withshoppers allows for increased consumer engage-ment and promotional and marketing relevancy thatlead to higher customer satisfaction and increaserevenues. In September 2014, following a pilot testof the application, Macy’s decided to roll out shop-Beacon in all of its 4,000 U.S. locations. Other majorretailers such as Target, American Eagle Outfitters,and JCPenney also partnered with shopkick andlaunched shopBeacon in 2014. Due to competitivepressure, there is expected to be a rapid adoption ofshopBeacon at other national retailers, too.

4. Evolution of the foundational IoTtechnologies

Various types of IoT applications have emerged, andthe willingness of enterprises to utilize them is

The Internet of Things (IoT): Applications, investments, and challenges for enterprises 435

growing rapidly. According to Bradley, Barbier, andHandler (2013), the IoTwill generate $14.4 trillion invalue; the combination of increased revenues andlower costs will migrate among companies and in-dustries from 2013 to 2022. From an industry per-spective, four industries make up more than half ofthe $14.4 trillion in value. These leading four indus-tries in terms of value at stake include manufactur-ing at 27%; retail trade at 11%; information servicesat 9%; and finance and insurance, also at 9%. Otherindustries such as wholesale, healthcare, and edu-cation lag behind in terms of value generation, witha range between 1% and 7%. Much of the value formanufacturers comes from greater agility and flex-ibility in factories, and from the ability to make themost of workers’ skills. Additionally, a large amountof the value for retailers comes from connectedmarketing and advertising. Geographic distributionsof the value are heavily driven by each region’srelative economic growth rate and by the relative

Table 1. Evolution of key IoT technologies

Before 2010 2010—201

Network � Sensor networks � Self-aware anorganizing ne� Sensor netwolocationtransparency� Delay-tolerannetworks� Storage netwopower networ� Hybrid netwotechnologies

Software andAlgorithms

� Relational databaseintegration� IoT-oriented RDBMS� Event-based platforms� Sensor middleware� Sensor networksmiddleware� Proximity/Localizationalgorithms

� Large-scale,

semantic softmodules� Composablealgorithms� Next generatbased social s� Next generatbased enterpapplications

Hardware � RFID tags and somesensors� Sensors built intomobile devices� NFC in mobile phones� Smaller and cheaperMEMs technology

� Multiprotocolmultistandardreaders� More sensors

actuators� Secure, low-c(e.g., Silent T

DataProcessing

� Serial data processing� Parallel dataprocessing� Quality of services

� Energy, frequspectrum-awprocessing� Data processicontext adap

Source: Adapted from Sundmaeker, Guillemin, Friess, and Woelffle

size of industry sector in each region. In the UnitedStates, $4.6 trillion of value is most prevalent in theservices area. However, in China, $1.8 trillion ofvalue is derived from rapid economic growth, mainlyin the manufacturing sector.

Table 1 shows projected evolution in the area offoundational IoT technologies: network, softwareand algorithms, hardware, and data processing.The network is the backbone of the IoT. It refersto uniquely identifiable objects (things) and theirvirtual representations in an Internet-like structure.Network technology is moving to unobtrusive wire-free communication technology that allows device-to-device applications to be deployed more flexibly.Network technology is evolving toward a context-aware autonomous network.

Objects rely on software to communicate effec-tively with each other and to deliver enhancedfunctionality and connectivity. Software shouldbe developed with the IoT’s interoperability,

5 2015—2020 Beyond 2020

d self-tworksrk

t

rks andksrking

� Network contextawareness

� Network cognition� Self-learning, self-repairing networks

openware

ion IoT-oftwareion IoT-rise

� Goal-orientedsoftware� Distributedintelligence,problem solving� Things-to-Thingscollaborationenvironments

� User-orientedsoftware� The invisible IoT� Easy-to-deploy IoTsoftware� Things-to-Humanscollaboration� IoT 4 All

,s

and

ost tagsags)

� Smart sensors(biochemical)� More sensorsand actuators(tiny sensors)

� Nanotechnology andnewmaterials

encyare data

ngtable

� Context-awaredata processingand dataresponses

� Cognitive processingandoptimization

(2010, p. 74)

436 I. Lee, K. Lee

connectivity, privacy, and security requirementsin mind. The focus of software development isshifting to user-oriented, distributed intelligenceand machine-to-machine and machine-to-humancollaboration.

The news that Google is paying $3.2 billion in cashto buy Nest, a smart thermostat business, demon-strates the value of hardware in the IoT. Hardware isinnovatively designed and robustly produced, drivenby the consumerized IoT devices which have myriadfeatures, functionalities, and operating environ-ments. While RFID tags and sensors have been thefocus of hardware innovation, miniaturizationof hardware and nanotechnology is leading theenergy-efficient, low-power hardware evolution.

IoT devices generate enormous quantities of datathat need to be aggregated and analyzed in real timeto provide information regarding status, location,functionality, and environment of the devices. Thetraditional data processing method does not workwell in the real-time streaming data process of theIoT environment. Since processing large quantitiesof IoT data in real time will increase workloads ofdata centers at an exponential rate, data processingwill become more context-aware, optimized, andcognitive.

In the IoT environment, a large number ofdevices are connected with each other, and it is notfeasible to process all the streaming data availableto those devices. Context-aware data processingenables sensors and devices to use context-specificinformation such as location, temperature, and theavailability of a certain device to decide what datato collect and interpret to provide relevant infor-mation to other devices or users. For example,context-aware data processing can deliver rele-vant information to a user by knowing the user’scurrent location (e.g., within a department store,a park, or a museum). Cognitive data processingintegrates the human cognition process into IoTapplications. Rather than being programmed todeal with every possible data-processing need, acognitive data-processing application is trainedusing artificial intelligence algorithms to sense, pre-dict, infer, and learn tasks and environments. Forexample, cognitive data processing uses image rec-ognition techniques to understand the surroundingenvironment, processes data for a user, and utilizesfeedback from the user to learn further. The opti-mization of data processing is critical to timelyprocessing of the continuous stream of massiveamounts of data. Technological advances in opti-mized data processing help make timely decisionsin time-critical big data applications such assmart grids, environmental monitoring, and smartmanufacturing.

5. IoT investment opportunities andevaluation (net present value vs. realoption approach)

Our survey shows the IoT is penetrating a widerange of industries including retailing, manufactur-ing, healthcare, insurance, home appliances, heavyequipment, airlines, and logistics. The benefits ofIoT technologies such as RFID-based merchandisetracking and home networking are concrete andimmediately measurable. Other IoT technologiessuch as intelligent automobiles and intelligent hos-pital robot systems are in the experimental stageand their benefits may be realized in the long term.While the IoT is relatively new, investment oppor-tunities abound, along with the development ofvarious foundational technologies summarized inTable 1. Companies are expected to take advantageof the wave of IoT innovations in the coming years.

In general, companies are going to take an im-mediate investment or a wait-and-see approach toinvestment based on the maturity level of the spe-cific IoT technologies. This section discusses twoinvestment evaluation methods widely used in thejustification of technology projects.

5.1. Net present value and real optionapproach

With so much potential value in the investment ofIoT technology, firms need an appropriate measureby which to properly assess its risks and rewards.The standard measure firms typically use to valueprojects, net present value (NPV), is inappropriateto use for several reasons. Chief among these in thiscircumstance is that it ignores flexibility in invest-ment such as reversibility and scalability in theevaluation horizon. No other technology investmenthas the flexibility that information technology in-vestments in general have (Fichman, Keil, & Tiwana,2005). All of the aforementioned IoT technologiesmay have had value arising from flexibility ininvestment. Thus, NPV tends to undervalue aproject’s worth and is not suitable for high-riskprojects. In order to value the IoT more appropri-ately, real option valuation may be an appropriateevaluation method. The following section discusseshow the real option valuation can be applied for IoTinvestment.

5.2. Real options

As implied by their namesake, real options are theright–—but not the obligation–—to take an actionduring a period of time. These include the optionsto expand, contract, and wait. Real options can

Table 2. Types of real option approaches

Type of Option Description

Option to Abandon/Switch This option gives management the option to abandon a project that is operatingat a loss and sell or redeploy the assets.

Option to Contract Similar to the option to abandon, this gives management the option to scale backa project that is operating at a loss.

Option to Defer/Postpone This gives management the option to wait/learn more to see if a project will beprofitable.

Option to Expand This gives management the option to expand/scale up the project based on itssuccess.

Table 3. Real options variables

Real Options Variable FinancialOptions

Present Value ofProject

S CurrentStock Price

Investment Cost ofa Project

X OptionExercise Price

Riskiness/Uncertaintyof the Project

s2 Stock PriceUncertainty

Time Window ofthe Project

T Time toExpiration

Time Value of Money r Risk-free Rate

Source: Adapted from Li & Johnson (2002)

The Internet of Things (IoT): Applications, investments, and challenges for enterprises 437

prove particularly valuable in fields of high uncer-tainty and risk, such as information technology. Asreviewed by Li and Johnson (2002), two main char-acteristics make real options an appropriate appli-cation for IT investments. First, IT projects typicallyrequire high initial investments and are often irre-versible. Second, IT investments can have very highuncertainty and risk. IT projects such as IoT projectsinherently contain technical uncertainty as well asmarket uncertainty (Fichman et al., 2005). Also, ITcan advance at a rapid pace and change directionquickly. These characteristics make real option val-uation ideal in valuing IoT investment projects, asit can capture value that otherwise would beoverlooked.

Good managers intuitively understand real op-tions. They understand that simple cash profits arenot the only value a project can add; other oppor-tunities may arise from engaging in projects. Realoption valuation allows management to quantifythese options to more accurately reflect the valueof a project and to have a real strategic impact onthe value of a project. Table 2 lists four generaltypes of real options.

There are several examples of how real optionsare used in valuation today. Any firm that operatesin a field of high uncertainty (e.g., pharmaceuti-cals) is likely to employ real options. Pharmaceuti-cal companies face uncertainty not only in drugdevelopment (akin to IT technical uncertainty) butalso in other external factors such as regulationsand patents (akin to IT market uncertainty). Realoption valuation is also used with movie deals. Moviestudios often purchase rights–—that is, real options–—to produce films. Film rights give the purchasingstudio the right but not the obligation to produce amovie. [Note that this is not the cost of producing themovie, simply the value of the right to produce it.]After rights are purchased, studios can then employsurveys and analysis to determine whether or not aproduction will be profitable. If the timing is notright, these options allow the studio to wait andperhaps produce the film at a later time.

5.3. Valuation

Like financial options, real options can be calculatedusing the Black-Scholes model or decision trees. Forreal options, using decision trees may be moreappropriate, as that will allow setting up possibili-ties of the project according to what managementbelieves them to be. When valuing real options, it isespecially important to stage the problem correctlyand to understand how real options are analogous tofinancial options. Table 3 provides a guide on howthey are related.

S represents the present value of cash flows fromthe project; X represents the cost to invest in theproject; s2 represents the riskiness of the project; Trepresents the period of time in which managementcan take an action; and r represents the risk-freerate the investment capital would earn.

Using decision trees to calculate the real optionvalue, one can stage the possible values a project cantake, exercise the option at the optimal time/valueof the project, and discount backward in order to findthe value of the option. For example, with a one-period decision tree, we begin with the startingvalue today, S0, and move forward one period. Thevalue can either increase to Su or decrease to Sd. Fromhere we can use the risk-free rate, r, and determine

438 I. Lee, K. Lee

the probability of success (p) and failure (1-p),determine the expected payoffs using the projectvalue less the investment amount, and essentiallywork backward to determine the value of the option.

5.4. Real options example

Here we offer an example. A company is looking toinvest in new smart vending machines that willreduce costs and increase profits. Management be-lieves there is a probability of 0.55 that there will bea high demand in this technology with a marketvalue of $140m, and a probability of 0.45 that therewill be a low demand in this technology with amarket value of $40m. Figure 1 shows a decisiontree without real options.

If the cost of investing in this technology at timezero is $100m and the discount rate is 8%, then froma simple NPV calculation the value of the project is—$12.04m, which the standard NPV rejects.

Present Value of Cash In flow

¼ ð140mÞð0:55Þ þ ð40mÞð0:45Þð1:08Þ1

¼ $87:96m

NPV ¼ $87:96m � $100m ¼ �$12:04m

However, management can use a real option ap-proach to evaluate this investment as a phased fi-nancing and scaling option. The company could startwith a pilot project and better learn the market overtime. In the following year, management could avoidfull investment of $100m into this smart vendingmachine technology if the market turns out to be$40m, and only invest in this technology if themarket turns out to be $140m. Therefore, manage-ment can value the option using a decision treethat takes the higher value as the exercisedoption. The option value of this project from thereal option perspective is ($140m — $100m)*0.55 +($0m)*0.45 = $22m/(1.08)1 = $20.37m. As long as thepilot project costs less than $20.37m, this pilot proj-ect with a following investment in the smart vendingmachine technology is worth doing. Figure 2 shows

Figure 1. A decision tree without real options

the option value calculation, discounted back oneperiod.

While this example was simplified for readers fromnon-finance backgrounds, more complicated scenar-ios can be analyzed using the same principles. ManyIoT projects have unclear project scopes and goalsand are using breakthrough technologies; in suchscenarios, there is a higher risk of project failureand greater irreversibility of investments than withtraditional technology projects. Our example high-lights the value of real option approaches to IoTprojects.

6. Challenges in IoT development

Based on the survey of IoT practices, this sectiondiscusses challenges in IoT development by enter-prises. As with any disruptive innovation, the IoTwill present multiple challenges to adopting enter-prises. For example, due to the explosion of datagenerated by IoT machines, Gartner (2014) suggestedthat data centers will face challenges in security, theenterprise, consumer privacy, data itself, storagemanagement, server technologies, and data centernetworking. This section discusses five technical andmanagerial challenges: data management, data min-ing, privacy, security, and chaos.

6.1. Data management challenge

IoT sensors and devices are generating massiveamounts of data that need to be processed andstored. The current architecture of the data centeris not prepared to deal with the heterogeneousnature and sheer volume of personal and enterprisedata (Gartner, 2014). Few enterprises would beable to invest in data storage sufficient to house allthe IoT data collected from their networks. Con-sequently, they will prioritize data for operationsor backup based on needs and value. Data centerswill become more distributed to improve process-ing efficiency and response time as IoT devicesbecome more widely used and consume morebandwidth.

Figure 2. A decision tree with real options

The Internet of Things (IoT): Applications, investments, and challenges for enterprises 439

6.2. Data mining challenge

As more data are available for processing and analy-sis, the use of data mining tools becomes a necessity.Data consist not only of traditional discrete data, butalso of streaming data generated from digital sensorsin industrial equipment, automobiles, electrical me-ters, and shipping crates. These streaming data areabout location, movement, vibration, temperature,humidity, and even chemical changes in the air. Datamining tools can invoke corrective processes to ad-dress immediate operational issues or inform man-agers of discoveries regarding competitors’ strategicmoves and customers’ preference changes thatwill impact their short-term and long-term businessactivities.

Data need to be tamed and understood usingcomputer and mathematical models. Traditionaldata mining techniques are not directly applicableto unstructured images and video data. Coupledwith the need for the advanced data mining toolsto mine streaming data from sensor networks andimage and video data, there is a shortage ofcompetent data analysts. McKinsey Global Insti-tute estimated that the United States needs140,000 to 190,000 more workers with analyticalskills and 1.5 million managers and analysts withanalytical skills to make business decisions basedon the analysis of big data (Manyika et al., 2011).

6.3. Privacy challenge

As is the case with smart health equipment andsmart car emergency services, IoT devices can pro-vide a vast amount of data on IoT users’ locationand movements, health conditions, and purchasingpreferences–—all of which can spark significant pri-vacy concerns. Protecting privacy is often counter-productive to service providers in this scenario, asdata generated by the IoT is key to improvingthe quality of people’s lives and decreasing serviceproviders’ costs by streamlining operations. The IoTislikely to improve the quality of people’s lives.According to the 2014 TRUSTe Internet of ThingsPrivacy Index, only 22% of Internet users agreed thatthe benefits of smart devices outweighed any privacyconcerns (TRUSTe, 2014). While the IoT continuesto gain momentum through smart home systemsand wearable devices, confidence in and acceptanceof the IoT will depend on the protection of users’privacy.

6.4. Security challenge

As a growing number and variety of connecteddevices are introduced into IoT networks, the

potential security threat escalates. Although theIoT improves the productivity of companies andenhances the quality of people’s lives, the IoT willalso increase the potential attack surfaces for hack-ers and other cyber criminals. A recent study byHewlett Packard (2014) revealed that 70% of themost commonly used IoT devices contain seriousvulnerabilities. IoT devices have vulnerabilitiesdue to lack of transport encryption, insecure Webinterfaces, inadequate software protection, andinsufficient authorization. On average, each devicecontained 25 holes, or risks of compromising thehome network. Devices on the IoT typically do notuse data encryption techniques.

Some IoT applications support sensitive infra-structures and strategic services such as the smartgrid and facility protection. Other IoT applicationswill increasingly generate enormous amounts ofpersonal data about household, health, and finan-cial status that enterprises will be able to leveragefor their businesses. Lack of security and privacy willcreate resistance to adoption of the IoT by firms andindividuals. Security challenges may be resolved bytraining developers to incorporate security solutions(e.g., intrusion prevention systems, firewalls) intoproducts and encouraging users to utilize IoT secu-rity features that are built into their devices.

6.5. Chaos challenge

The evolution of IoT technologies (e.g., chips,sensors, wireless technologies) is in a hyper-accelerated innovation cycle that is much fasterthan the typical consumer product innovation cycle.There are still competing standards, insufficientsecurity, privacy issues, complex communications,and proliferating numbers of poorly tested devices.If not designed carefully, multi-purpose devices andcollaborative applications can turn our lives intochaos. In an unconnected world, a small error ormistake does not bring down a system; however, in ahyper-connected world, an error in one part of asystem can cause disorder throughout. Smart homeapplications and medical monitoring and controlsystems consist of interconnected sensors and com-munication devices and controllers. If a sensor of amedical monitoring and control system malfunc-tions, the controller may receive an incorrect sig-nal, which may prove fatal to the patient. It is notdifficult to imagine smart home kits such as thermo-stats and residential power meters breaking down orbeing attacked by hackers, creating unexpectedsafety problems. The Internet bandwidth can getsaturated with data traffic of proliferating devices,creating system-wide performance problems. A sin-gle device may have an insignificant problem, but

440 I. Lee, K. Lee

for the system as a whole, the chain reactions ofother connected devices can become disastrous. Toprevent chaos in the hyper-connected IoT world,businesses need to make every effort to reducethe complexity of connected systems, enhancethe security and standardization of applications,and guarantee the safety and privacy of users any-time, anywhere, on any device.

7. Conclusion

Because the IoT is such a recent development, thereis still a paucity of studies on the social, behavioral,economic, and managerial aspects of the IoT. Thismakes it very challenging for companies to makeinformed decisions as regards IoT adoption/implementation. Our article is one of the first studieson a conceptual model of IoT applications for enter-prises. In this article we identified three categories ofIoT applications: monitoring and control, big dataand business analytics, and information sharing andcollaboration. We also presented investment oppor-tunities and investment evaluation with NPV andreal options. Finally, we discussed five challengesin implementing IoT applications for enterprises.

Acknowledgment

We offer special thanks to Editor Dr. Marc J.Dollinger for his valuable comments and sugges-tions.

References

Atzori, L., Iera, A., & Morabito, G. (2010). The Internet of Things:A survey. Computer Networks, 54(1), 52787—52805.

Bradley, J., Barbier, J., & Handler, D. (2013). Embracing theInternet of Everything to capture your share of $14.4 trillion.Cisco White Paper. Retrieved from http://www.cisco.com/web/about/ac79/docs/innov/IoE_Economy.pdf

Chui, M., Loffler, M., & Roberts, R. (2010). The Internet of Things.McKinsey & Company. Retrieved from http://www.mckinsey.com/insights/high_tech_telecoms_internet/the_internet_of_things

Fichman, R., Keil, M., & Tiwana, A. (2005). Beyond valuation:‘‘Options Thinking’’ in IT project management. CaliforniaManagement Review, 47(2), 74—96.

Gartner. (2014, March 19). Gartner says the Internet of Thingswill transform the data center. Retrieved from http://www.gartner.com/newsroom/id/2684616

Gubbi, J., Buyya, R., Marusic, S., & Palaniswami, M. (2013).Internet of Things (IoT): A vision, architectural elements,and future directions. Future Generation Computer Systems,29(7), 1645—1660.

Hewlett Packard. (2014, July 29). HP study reveals 70 percent ofInternet of Things devices vulnerable to attack. Retrievedfrom http://www8.hp.com/us/en/hp-news/press-release.html?id=1744676#.VOTykPnF-ok

Hsueh, C.-F., & Chang, M.-S. (2010). A model for intelligenttransportation of perishable products. International Journalof Intelligent Transportation Systems Research, 8(1), 36—41.

Li, X., & Johnson, J. (2002). Evaluate IT investment opportunitiesusing real options theory. Information Resources ManagementJournal, 15(3), 32—47.

Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R.,Roxburgh, C., et al. (2011, May). Big data: The next frontierfor innovation, competition, and productivity. McKinsey &Company. Retrieved from http://www.mckinsey.com/insights/business_technology/big_data_the_next_frontier_for_innovation

Sundmaeker, H., Guillemin, P., Friess, P., & Woelffle, S. (2010).Vision and challenges for realising the Internet of Things.Accessible at http://www.researchgate.net/publication/228664767_Vision_and_challenges_for_realising_the_Internet_of_Things

TRUSTe. (2014). TRUSTe Internet of Things privacy index–—USedition. Retrieved from http://www.truste.com/resources/privacy-research/us-internet-of-things-index-2014/

White, C. C., III, & Cheong, T. (2012). In-transit perishableproduct inspection. Transportation Research Part E: Logisticsand Transportation Review, 48(1), 310—330.