the systematic improvement of advice given by public ... · the systematic improvement of advice...
TRANSCRIPT
The Systematic Improvement of Advice Given By
Public Sector Call Centres
Thesis submitted to:
QUEENSLAND UNIVERSITY OF TECHNOLOGY
Faculty of Information Technology
Research Centre for Information Technology Innovation
Submitted in fulfilment of the requirements for the degree of
MASTER OF INFORMATION TECHNOLOGY
Neville Lindsay Schefe
Dip.T. (Science/Mathematics)
B.App.Sc. (Comp)
Grad.Dip. (Software Quality)
Supervisor: Greg Timbrell
June 2006
i
Statement of Original Authorship
The work contained in this thesis has not been previously submitted for a degree or
diploma at any other higher education institution. To the best of my knowledge and
belief, the thesis contains no material previously published or written by another
person except where due reference is made.
Neville Lindsay Schefe (Student Number: n0135127)
Signed: _______________________
Date: _______________________
ii
Keywords:
Knowledge Management; Call Centre; Information Systems; Problem Types; Quality
of Service
Abstract:
The persistent demand for increased accountability and value for money in the public
sector from both the public and governments raises the issue of quality of service in
advice-giving by governmental agencies. The goal of this study is to develop a model
to validate frameworks able to contribute to improved advice-giving through the
application of knowledge management principles.
Zack’s (2001) Four Knowledge Problem Model, Brogowicz, Delene, and Lyth’s
(1990) Synthesised Service Quality Model, and Markus’s (2001) Theory of
Knowledge Reuse are used to examine knowledge strategies in advice-giving
through the application of a case study methodology. Two Queensland public-sector
call centres are investigated.
This study confirms that although the studied call centres operate under differing
business drivers, agents have developed strategies generally consistent with those
suggested by Zack (2001) to deal with uncertain, complex, and ambiguous problem
types. No equivocal problems were encountered in the study. The solution of the
former problem pair of uncertainty and complexity relies on knowledge that is
codified and stored in databases, while the latter equivocality and ambiguity, seeks
out experts who apply both technical and functional knowledge to the problem
resolution. Roles performed by call-centre agents predominantly align with those
described by Markus (2001), with the opportunity to enhance performance through
contribution by shared-work producers to knowledge repositories. The problem-
solving strategies employed by agents and the technical capabilities of the call
centres combine to deliver a level of service quality which, although meeting client
expectations, has been able to be improved through the application of knowledge
strategies targeting efficient problem resolution through knowledge reuse.
iii
Table of Contents
Statement of Original Authorship............................................................................................... i
Keywords:.................................................................................................................................. ii
Abstract:..................................................................................................................................... ii
Table of Contents...................................................................................................................... iii
List of Figures............................................................................................................................ v
List of Tables ............................................................................................................................ vi
List of Tables ............................................................................................................................ vi
Acknowledgments ................................................................................................................... vii
Introduction.................................................................................................................................. 1
1.1. Context of the Study..................................................................................................... 1
1.2. Background to the Research ......................................................................................... 3
1.3. Research Aims.............................................................................................................. 5
1.4. Limitations of the Study ............................................................................................... 6
1.5. Outline of Thesis Structure........................................................................................... 6
Context of the Study and Review of Literature ...................................................................... 10
2.1. Introduction ................................................................................................................ 10
2.2. Academic Literature ................................................................................................... 10
2.3. Summary..................................................................................................................... 32
Research Method ....................................................................................................................... 34
3.1. Introduction ................................................................................................................ 34
3.2. Objectives of the Research ......................................................................................... 34
3.3. Approach .................................................................................................................... 35
3.4. Customer Expectations............................................................................................... 43
3.5. Problem Resolution Strategies.................................................................................... 44
3.6. The Knowledge Management Strategy ...................................................................... 47
3.7. Quality and the Call-Centre Response........................................................................ 49
3.8. Pilot work ................................................................................................................... 50
3.9. Implementation........................................................................................................... 51
3.10. Summary..................................................................................................................... 58
Call-Centre Work, Organisation, and Processes .................................................................... 59
4.1. Introduction ................................................................................................................ 59
4.2. A Broad View of Call Centres.................................................................................... 59
4.3. Evolution of the Study: CC-A .................................................................................... 60
4.4. Evolution of the Study: CC-B .................................................................................... 63
4.5. Operational aspects of CC-A...................................................................................... 66
4.6. Operational aspects of CC-B ...................................................................................... 70
4.7. Summary..................................................................................................................... 74
A Framework for Analysis ........................................................................................................ 76
5.1. Introduction ................................................................................................................ 76
iv
5.2. Zack’s (2001) Four Problem Model............................................................................ 76
5.3. Summary..................................................................................................................... 84
Patterns from Observations, Interviews, and Responses........................................................ 85
6.1. Executive Management Interviews............................................................................. 86
6.2. Management and Team Leader Interviews................................................................. 86
6.3. Responses to Calls ...................................................................................................... 95
6.4. Service Quality.......................................................................................................... 101
6.5. Summary................................................................................................................... 104
A Pragmatic Approach to Call-Centre Strategies................................................................. 106
7.1. Service Value............................................................................................................ 107
7.2. Summary................................................................................................................... 119
Conclusion................................................................................................................................. 121
8.1. Introduction............................................................................................................... 121
8.2. Research Question .................................................................................................... 122
8.3. Implications for Theory ............................................................................................ 124
8.4. Implications for Practice ........................................................................................... 125
8.5. Further Research Opportunities ................................................................................ 126
8.6. Summary................................................................................................................... 127
References ................................................................................................................................. 128
Appendices ................................................................................................................................ 134
Appendix 1 - General Information Processes in CC-B.......................................................... 134
Appendix 2 - Average Advice Line Wait Time – CC-A........................................................ 135
v
List of Figures
Figure 1 - Organizational Knowledge Evolution Cycle (Rus et al., 2001, p. 17) ........................ 17
Figure 2 - Capability-Creating Activities: Shared Problem Solving (Leonard, 1995, p. 60)....... 20
Figure 3 - Relationship among service-quality concepts (Dedeke, 2003, p. 280) ....................... 27
Figure 4 - Rules and cycles in resolving an equivocal problem (Weick, 1969, p. 77) ................ 31
Figure 5 - Convergence of Multiple Sources of Evidence (Yin, 1994, p. 93) ............................. 37
Figure 6 - The Structured Case Research (Carroll et al., 1998, p. 65) ......................................... 39
Figure 7 -Query-Response Cycle (Timbrell et al., 2005b, p. 546)............................................... 40
Figure 8 - Tiered Query-Response Cycles (Timbrell et al., 2005b, p. 550)................................. 41
Figure 9 - Micro-performance Approach to Trust (Van de Walle & Bouckaert, 2003, p. 894) .. 44
Figure 10 - Generic Taxonomy of IT Helpdesk Calls (Pentland, 1991, p. 102) .......................... 46
Figure 11 - Capability-Creating Activities: Problem Solving (Leonard, 1995, p. 60)................. 50
Figure 12 - Kinds of Responses to Calls to IT Helpdesk (Pentland, 1991, p. 115) ..................... 54
Figure 13 - Age Demographics of staff employed by the agency hosting CC-A ........................ 62
Figure 14 -Query-Response Cycle (Timbrell et al., 2005b, p. 548)............................................. 65
Figure 15 - Operational Context of the Call-Centres Studied...................................................... 66
Figure 16 - Call logging system – CC-B...................................................................................... 71
Figure 17 – Agent Performance Graph (Interview 2b - Full transcript, June 16, 2004) ............. 73
Figure 18 - Research components................................................................................................ 85
Figure 19 - Complex Problem Devolution................................................................................... 98
Figure 20 - Problem Devolution Hierarchy ............................................................................... 101
Figure 21 - Graphical Representation of Frequency of Quality Criteria.................................... 103
Figure 22 - Resolution Rate - Quality Matrix ............................................................................ 106
vi
List of Tables
Table 1 - Knowledge States (Zack, 2001, table 2) ....................................................................... 32
Table 2 - Summary of Zack’s Four Knowledge Problem Model (Zack, 2001, figure 1)............. 45
Table 3 - Three Cycles of the Structured Case Method as Applied to This Study....................... 51
Table 4 -Coding Scheme for Software Support Process (Pentland & Rueter, 1994, p. 494) ....... 53
Table 5 - Proposed Coding Scheme for Problem Type................................................................ 55
Table 6 - Proposed Coding Scheme for Resolution Strategy ....................................................... 55
Table 7 - Proposed Coding Scheme for Quality Characteristics .................................................. 56
Table 8 - Data definition for data capture tool ............................................................................. 57
Table 9 - General information processes in CC-B ....................................................................... 64
Table 10 - Summary of the Strategies Employed by Both Call Centres ...................................... 74
Table 11 - CC-A, CC-B, and the Q-R Cycle................................................................................ 75
Table 12 - Summary of Knowledge Strategies Used in the Call Centres by Problem Type........ 80
Table 13 - Frequency of Repository References in Interviews .................................................... 87
Table 14 - Frequency of Response Strategies .............................................................................. 94
Table 15 - Frequency of Quality Criteria ................................................................................... 102
Table 16 - Analysis of critical incidents (De Ruyter, Wetzels & Van Birgelen, 1999, p. 1139)109
vii
Acknowledgments
The genesis of the problem being addressed in this study lies with frustration at being
asked to “work smarter, not harder”, to “be open and accountable” and to “do more
with less”. Does this mean I am not working smart, that I am unethical and that I can
be more frugal in the application of resources to my work? On reflection, I decided
that to get different outcomes, I needed to do things differently and started on a
journey that has probably only taken the first few small steps to where I hope to end
up.
The inspiration to try different things in the apparent monolith of government came
from my Executive Director, Jan Archer, who had the job of forming a new team in a
previously untried model of corporate service delivery. Her support and mentoring
destroyed the myths I had long held as being absolute truths in operation of
government. The support of my current supervisors and all the staff responsible for
the two call centres studied has been exceptionally tolerant and obliging over the past
three years. I am extremely grateful for the access provided and the honesty with
which issues have been addressed. The interpersonal exchanges in the call centres
provided me with a valuable insight into the informal activities that contribute to
their operations. Also, their interest in the actual work I was doing and the positive
approach they took reinforced my feeling that well-managed staff in an environment
that provided both personal and organisational challenges and responsibilities will
continue to develop to meet those challenges.
The validation of my work is based in the coding and analysis of actual calls made to
the call centres. I wish to acknowledge the support of Kate Bowman who provided
the assistance, as timelines started to shorten, in the moderated coding of the calls.
She did this accurately and completed the work to meet a tight schedule.
The academic staff members at QUT have cajoled, encouraged, critiqued and
contributed in a way that kept me on track from the very first day I commenced this
programme. To combine a research effort with the demands of full time work means
decisions on priorities and milestones need to be carefully attended to. The work put
in by Greg Timbrell to ensure papers prepared for conferences and publication met
both timelines and an academic standard is greatly appreciated. I know I gave him
some late nights in the editing/review phase. The need for sound methodology was
viii
provided by Professor Guy Gable. He assisted tremendously in the formulation of the
original research question while Professor Alan Underwood was instrumental in
continuing this through to maturity.
Finally, as any researcher knows, the support of family is essential for the successful
completion of such a project. To Joan, who has now caught the bug and spends many
weekends and late nights in the study, Ben and Sam, I appreciate the sacrifices the
family has made to accommodate my excesses. I derive inspiration though from the
fact that my whole family is currently studying and achieving tremendously at
university level. Thank you and best of luck with your own endeavours.
Introduction
1
Chapter 1
Introduction
This dissertation investigates the ability to measure and improve the quality of advice
provided by public servants. The simple question of “How do public service
organisations learn from daily routines and translate this knowledge into better
service performance?” is addressed through a case study of two Queensland
Government Departments. Milner (2000) identifies desired global reforms in the
public sector, and the need for the creation of competence through knowledge
management strategies. Such competencies are also discussed by Pentland (1991) as
they apply to a dynamic, knowledge-intensive task environment. The goal of this
thesis is to validate existing theoretical models for customer interaction and problem
resolution within the Queensland Public Service.
1.1. Context of the Study
Faced with a decline of trust in public and private institutions (Gualtieri, 2004;
Mahon, 2003), chief executives are attempting to rebuild faith through quality of
service and response to client and citizen service expectations. Rather than adhere to
technocratic utopian strategies that look to technology as a saviour, contemporary
leading-edge agencies must now align their business requirements and the
organisational relevance of implementing a service-oriented architecture (Davenport,
Eccles, & Prusak, 1992; Kreizman, Kotsikopoulos, & Shah, 2005). Such alignment is
reflected in the trend to identify business problems correctly, and formulate their
related strategic solutions. Hence technology, such as call-centre automation, is now
seen as a driver for performance-based innovation. Clients also continue to demand
an increasingly higher quality of service, with each cycle of improvement feeding the
expectation of further improvement.
Public-sector Chief Executive Officers (CEOs) have recently added a new dimension
to performance reporting, emphasising integrated service delivery. Their strategic
documents reference e-commerce, electronic transactions for government services,
one-stop shops, and 24 x 7 access – all of which is currently referred to as “virtual
government”. Public expectation acknowledges the importance of merit and equity,
yet demands efficiency and effectiveness (Milner, 2000). How then is a CEO to
Introduction
2
deliver a public service, governed by political outcomes rather than economic ones,
that meets the expectations of an increasingly demanding client group?
This research seeks to investigate the application of a model that is capable of
informing public-service knowledge strategy and its associated technology
investment process. So as to limit the number of uncontrolled variables in the study,
and to take advantage of access to data, the thesis will focus on the analysis of
strategies employed in public-sector call centres. An analysis of problem types
encountered, and their associated mapping to an appropriate strategy based in
knowledge management principles, is used to identify areas of strength and
weakness. This method enables outcomes of such strategies to be predicted and
measured, thus engendering confidence that investment in such strategy will provide
commensurate return.
The research problem has its genesis in the increasing demands for efficiency in the
public sector, with public expectation of increased performance levels of public
officials. Information technology has been promoted as the panacea to many ills in
both commercial and public enterprises; however, the sharpened focus on
accountability of IT systems following Year 2000 projects has caused management
boards to take a more critical look at IT and other investments which purport to
deliver efficiencies.
Concurrently, the knowledge management agenda has gained increasing prominence.
It is widely accepted that the current decade will see growth and maturity in
knowledge industries where customers will continue to expect high quality, value-
for-money services executed in a timely and convenient manner. The public sector
will not be exempt from this trend. Knowledge management strategies and their
promise of capability improvement will therefore be called on to deliver further
public-sector efficiency and effectiveness. Successful knowledge management
programs sustaining themselves through demonstrable business results are
exemplified in the following statements:
• Ford Motor Company's Manufacturing Best Practices Program improves the
efficiency of manufacturing processes and the quality of products.
• Rolls-Royce's Knowledge Acquisition and Modelling Process improves
project management processes.
• BP's knowledge management program started with the exploration and
production process, but now includes project management processes.
Introduction
3
• Shell Oil's Global Learning and Development involves continuous learning
around the exploration and production process.
• CSC's Communities of Practice started with a bid-and-proposal process and
now extends across the company into all lines of business.
• J.D. Edwards' Knowledge Garden started by focusing on the sales support
process and expanded to include resellers and customer service.
• The Knowledge and Learning Practice at the World Bank Institute focused on
improving development projects funded through the World Bank. (Caldwell, 2004, p. 2)
Notwithstanding, Return On Investment guarantees are being sought prior to
approval of business cases, owing to links with IT and the scepticism of Chief
Executive Officers for fad management innovations. Call centres are an option in the
drive for cost-effective delivery of services.
1.2. Background to the Research
This research has been limited to the call centres of each of the organisations being
studied, since they provide an instance of service provision able to be documented
and analysed. Initial literature searches provided little contemporary rigorous
research on the contribution of call centres to business objectives, instead primarily
focussing on quantitative performance measures such as time-in-queue, abandonment
rate, and talk time. This study is unique in that it links business goals, knowledge
strategy, and quality criteria as they apply to call centres in the Queensland public
sector. Direction has been taken from Yin (1994) to ensure the criteria for quality
research design are met in this study. The call centres provide a consistent and high
volume of rich data, with instances able to be tracked from initial contact to
resolution of queries on up to 200 calls per day per operator. This data has been
available for real-time capture through “double jacking”1 conversations, and
historically through calls recorded for other organisation-specific purposes. This data
is supplemented by interviews, observations, and published materials, which have
been used both to scope the research environment and to validate research findings.
Interviews with senior departmental managers have determined the business drivers
for each call centre, and their integral position to the broader objectives of quality
service provision via multi-channelled delivery mechanisms.
1 “Double jacking” is the process whereby the researcher is able to listen in to the conversation between the caller and the agent without influencing the interactions. Supervisors use this facility as a component of individual agent performance analysis.
Introduction
4
One strategy embraced by the public sector to enhance services is e-government: the
application of communication and information technologies to the organisation and
operation of government (Teicher, Hughes, & Dow, 2002). An element of the
e-government strategy is the utilisation of call centres and their associated
technology. Side by side with Internet-based services, the call centre remains a
central plank in the Government’s service delivery platform for both public and
internal inquiries. This “joined-up government” or “citizen-centric government”
concept has been studied by a number of researchers (Ling, 2002; Milner, 2000).
This concept expresses a fundamental change in paradigm from a traditional
benevolent or patriarchal model to one of active engagement by the broader
community. Call centres provide the opportunity for the efficient and equitable
delivery of quality advisory and support functions to government, which is in-line
with the tax-paying public’s expectation that government will deliver efficient and
effective public-sector operations (Davis & Weller, 2001).
Government agencies are now expected to develop performance indicators for the
services they provide, which is a way of measuring the progress toward government-
declared objectives. Managing for Outcomes (MFO) (2003), a Queensland
Government initiative commenced in 1997, addresses integrated planning, budgeting,
and performance management, and is intended to achieve quality, client-responsive
services, value for money, and improved resource allocation. The emphasis on
measures of efficiency and effectiveness is reinforced in the 2002 MFO Performance
Management Framework document:
“Through monitoring output performance, the efficiency and effectiveness of
service delivery, including the extent, quality and benefit of services in
relation to cost, can be determined.” (Performance Management Framework,
2002, p. 1)
Difficulties have been encountered in the measurement of service outputs and
outcomes, since the effectiveness of certain critical Government functions (e.g.,
compliance functions) has little to do with service delivery, or discrete
measurable outputs (Teicher et al., 2002). More meaningful measurement will
be founded on the decomposition, closer examination, and greater
understanding of these services, including those provided by call centres. In the
quest to improve government service quality, efficiency, and effectiveness, an
Introduction
5
analyst must ask: “How do workforce and technical strategies combine to assist
the enterprise to measure and improve on the quality of advice provided to
external customers and internal clients?” Several authors have provided
tantalising components of the puzzle. These include Zack (2001), who offers a
model of problem types and strategies to solve them; Markus (2001), who
investigates repositories and organisational memory systems in order to develop
a theory of knowledge reuse; and Timbrell Nelson, and Jewels (2003), who
devised a methodology to test the knowledge reuse theories as they are applied
in government agencies. In order to develop a defensible response to the
previous question, the principles and concepts underpinning the work of these
authors have been used to investigate two Queensland Government call centres.
1.3. Research Aims
This research project aims to provide evidence that the quality of advice given by
call-centre staff can be improved through the judicious application of a model both
able to predict the outcomes of investment in a knowledge management strategy, and
to measure the impact of such a strategy. The research question is:
“How well does Zack’s (2001) framework represent problem types, and
hence impact on strategies utilised in response to queries encountered in
a public-sector call-centre environment?”
More specifically, the study aims to:
(1) Develop a model that defines the interrelationships between customer queries
and call-centre knowledge management strategies in the provision of
subsequent responses;
(2) Suggest knowledge management strategies to improve appropriate knowledge
re-use within call centres;
(3) Focus discussion and promote constructive interaction for developing a
sophisticated understanding of the call-centre environment, specifically
within the public sector;
(4) Explore the characteristics and implications of systems-based initiatives and
knowledge management interventions that may impact upon a call centre’s
ability to realise service quality benefits;
Introduction
6
(5) Inform academic research directions in call-centre management; and
(6) Provide practical advice to call-centre management.
The examination of two Queensland Government call centres through a knowledge
management lens has the potential to provide guidance in these pursuits. The models
and descriptions in this thesis represent a better understanding of the knowledge-
intensive environment characteristic of a call centre.
1.4. Limitations of the Study
a) The study was limited to two Queensland public-sector call centres to reduce
the variables impacting on the findings. This has, however, reduced the
capacity to confidently extrapolate or generalise to other advice-giving
environments and to service delivery activities outside the Queensland public
sector.
b) Problems of equivocality have not been encountered in the study, so no
claims are able to be made with respect to their place in the framework. This
also limits the ability to claim a highly coherent alignment of Zack’s (2001)
taxonomy with behaviours in the public-sector call centres.
c) The call–encoding methodology has been based on work done by Pentland
(1991) in IT helpdesk environments, where he develops a grammar to analyse
problem-resolution “moves”. In this study, recorded calls have been double-
coded by two different people with a high degree of correlation. Live calls
were not recorded, and hence have been unable to be coded a second time.
This leads to an inherent weakness due to the inability to verify the reliability
of these coded, but not recorded, calls.
1.5. Outline of Thesis Structure
This thesis has seven chapters. The first chapter articulates the processes used to
cover the research question in a scholarly manner. Chapter Two further delves into
the motivation for this study and reviews the relevant literature to provide the
foundation for the further enhancement of aforementioned theories. The chapter
looks at the literature available to assist decision makers in the pre-assessment and
post-implementation evaluation of knowledge management strategies aimed at
improving the capability to provide advice through the burgeoning incidence of call
Introduction
7
centres. The found literature points to a model which incorporates the needs and
quality perceptions of clients (callers), enables analysis of problems raised by callers
via a taxonomy developed by Zack (2001), focuses on the benefits of knowledge
reuse as a strategy influenced by Markus (2001), and then uses determinants of
quality published by Brogowicz et al. (1990) to assess the quality of advice (and
hence improvements based on the selected strategies). Such a model is underpinned
by the assumption that knowledge is the currency of successful organisations, and
that their leaders are able to appreciate the underlying principles and objectives of
knowledge management.
A case study methodology has been chosen in order to analyse the activities of the
call centres. The rationale for this is argued in Chapter Three. Yin’s (1994) case
study method underpins the robustness of the developed models. Although a
substantial amount of quantitative data is available and used, the multi-method
approach suggested by Gable (1994) is not employed. This is due to the fact that the
data is used to develop the model in a qualitative way with greater reliance placed on
patterns, observations, and rich context rather than a detailed statistical examination.
The stepwise process involved in research question definition in the case
methodology, selection of cases and data gathering/analysis techniques, data
collection, analysis, and report preparation has been enhanced by the application of
the iterative methodology suggested by Carroll, Dawson and Swatman (1998). Each
cycle better informs the conceptual framework underpinning the study and provides a
better starting point for a further investigation of the available data.
Chapter Four contains an overview of the organisational structure, tasks performed,
and the related business drivers for the call centres being studied. The comparison
and contrasts able to be derived indicate differing philosophies and strategies used,
with one call centre effectively being divided into teams of teams in order to develop
expertise in specialist areas of operation, while the other develops generalists and
enhances their ability to locate tier-2 expertise, as defined by access to specialist
knowledge primarily of a scientific nature. It also highlights variation in the use of
technology to achieve organisational business goals. Interactive Voice Response
(IVR) systems, a dedicated single source database of business information, and
systems to collect call metadata each exist in one call centre but not the other. This
notwithstanding the performance indicators remain predominantly those reflected in
Introduction
8
the literature to measure generic outputs such as time-in-queue, time taken to
complete, and abandoned calls (Anton & Gustin, 2000). An assumption underpinning
call-centre operation is that referral to tier-2 expertise is a more expensive activity
than reliance on a local call-centre resolution. Self-sufficiency of call-centre
operations (or the degree to which calls are resolved in the call centre) is a goal of
each department. Knowledge strategies used to support this are consistent with the
nature of calls being made. The policies of the agencies do not always reinforce the
goals stated by management during interviews, however, particularly when conflicts
between efficiency and effectiveness arise.
Since multiple sources of data have been considered for this study, specific
frameworks described in Chapter Five are used to gather and analyse data sources
such as interview reports, brochures, templates, recorded calls, and real-time calls for
each of the organisations. The Four Problem Type Model (Zack, 2001) is used to
determine the relevant resolution strategies. Characteristics concerning quality are
extracted from the work of Brogowicz et al. (1990). Response strategies, based on
the work of Markus (2001) and Zack (2001) are used to assess the resolution of
queries through knowledge reuse. The Query-Response Cycle (G Timbrell &
Shepperd, 2002) provides an appropriate framework to link problem, resolution, and
quality in the call-centre environment. Direct observation is able to be employed as
an analytical tool because of the regimented and static nature of these call centres
(with each having single traditional work sites which were easily accessible).
Although the call centres are compared and contrasted, the sum total of the recorded
and live calls is used to produce a more comprehensive coverage of the problem
types described by Zack (2001).
Chapter Six is used to illustrate the patterns and responses obtained from the data
collected in the call-centres. Interviews with management and team leaders, direct
observations of work practices, and coded calls are analysed using NVivo qualitative
research analysis software to distil trends and learnings from a range of perspectives.
A pragmatic approach to call-centre strategies is discussed in Chapter Seven. The
alignment of organisational goals with the role of the call centre and knowledge re-
use strategies employed contributes to the success of the call centre. The dilemma of
who defines the quality of advice (caller or expert) is resolved through the
introduction of the concept of service value. This uses the service quality paradigm
Introduction
9
of Brogowicz et al. (1990) to incorporate the issues of customer satisfaction and
service quality determinants of personal quality, technical competence, and the
functional service offering of the agency.
The final chapter, Chapter Eight, addresses the implications of the thesis’s findings
for government call centres. It outlines the future extension of this research to a
more general organisational environment. Concluding remarks provide advice to
call-centre management which arise from this study.
Context of the Study and Review of Literature
10
Chapter 2
Context of the Study and Review of Literature
2.1. Introduction
In this chapter, a range of academic literature is investigated in a structured attempt
to provide a robust foundation for the qualitative data gathering employed in the case
studies. Alignment of issues with research interests is achieved through an
understanding of the contemporary public-sector context and the predominant
organisational drivers in Section 2.2. Since call centres have been identified as the
suitable response to efficiency drivers (Milner, 2000), Section 2.3.1 focuses on call-
centre performance. The measurement of advice using quality principles is
investigated in Section 2.3.2, and revisited in Sections 2.3.7 and 2.3.8, where
improvement strategies are introduced based on knowledge management principles.
Since the underpinning assumption is that better problem-solving capability will
drive improved performance through knowledge reuse, knowledge management,
generation, capture, and reuse are the foci of Sections 2.3.3 through 2.3.6. The
strategy employed in this analysis is to identify the significant concepts involved in
the improvement of advice provided by public-sector call centres, and to develop a
framework to converge the work of thought leaders.
2.2. Academic Literature
2.2.1 Call Centres and Performance
Call centres have become the major access point for customer interaction with an
organisation. Many organisations are transforming traditional backroom call centres
through rejuvenated technology to become the primary front-line interface of the
enterprise (Anton, 2000). The major business driver for call-centre expansion,
particularly within the public sector, is cost. In 1999, Riggs and Thyfault estimated
the cost of handling a phone call to be $US3 compared to $US25 for responding to a
written letter, the traditional communication channel for government bureaucracies.
Tsoukas and Vladimirou (2001) confirm this in their study of a call centre in Greece
where speed-to-answer contributes to both the lowering of operational cost and the
delivery of high quality service. They identified the importance of accumulated
Context of the Study and Review of Literature
11
experience and knowledge of other agents to the organisation, and highlighted the
social relations at work compared to those absent in technical systems. Compelling
reasons for improved call-centre service cited in literature include loyalty, attracting
new business, and cost cutting, as well as demand from customers (Wilde, 2000).
This accumulated experience and knowledge requirement of call centres is also
predicted by Gartner researchers (Close, 2003). Here, knowledge management
strategy is proposed as a significant determinant of service performance in a call
centre. The predominance of codification strategies (scripting, issues databases, and
document repositories) employed in call centres to improve efficiency satisfies
criteria determined by Hansen, Nohria, & Tierney (1999): that to be competitive,
organisations need to become expert in either codification or personalisation. A
knowledge strategy that includes “The informal memory system (both individual and
collective) which has gradually built over time, consisting of the individual stocks of
experience held by each operator, and by stories shared in their community”
(Tsoukas & Vladimirou, 2001, p. 985) will provide greater capability for agents to
give improved quality advice to customers. The degree to which this heuristic
contributes knowledge to call-centre performance is significant, but is not able to be
quantified by the traditional statistical reporting.
Humphrey (1989) refers to the need to be able to measure activities in order to
develop and improve them. An awareness by executives of the impact of
performance management systems on multi-functional teams in the area of service
delivery is minimal (Meyer, 2004). Significant investment is being made in
technologies within call centres which lead to issues of Return On Investment, with
such reporting requiring some form of benchmark to justify the benefit of the outlay
through achievement of outcomes. This is particularly the case in the public sector,
which is under continuous pressure to improve performance in a climate of
decreasing funds. While responsiveness is important, it is not a sufficient condition
to achieve customer satisfaction. The measures predominantly published on call-
centre performance relate to time-in-queue (on hold), time in conversation, calls per
hour, and other similar quantitative measures (Anton & Gustin, 2000). Duder and
Rosenwein (2001), for example, suggest a formula to calculate the number of call-
centre agents to achieve zero abandonment rates. Their argument is that by reducing
time-in-queue, abandonment will fall and the quality of service will improve. The
Context of the Study and Review of Literature
12
US Internal Revenue Service employed a similar operations research approach where
the main objective was to allocate resources in such a way as to minimise the number
of customers who receive a busy signal or who must hold for a long time when they
call for information (Harris, Hoffman & Saunders, 1987). In addition, QANTAS
employed a model utilising queuing and integer linear programming to reduce
waiting time and evaluate the relationships of staff size to waiting time and service
time (Gaballa & Pearce, 1979). Yet, in all call-centre interactions, at the point the
caller makes initial contact, the primary purpose of its existence (i.e., taking of
queries and offering of advice in response) has not been fulfilled. The perceptions
being formed by clients in their early interactions with the call centre will impact on
their impression of the organisation. This serves as an indicator that attention needs
to be paid to the responsiveness attribute of service quality.
2.2.2 Measuring the Quality of Advice
Further measures, particularly with respect to customer satisfaction, are obtained
from exit surveys of call-centre clients. Many factors affect the perception a caller
may have of the quality of advice received, and hence impact on how exit
questionnaires are scored. In some instances such advice may not be palatable to the
caller, but the public servant’s creed of “without fear or favour” requires that such
advice be given to as high a standard as possible. Jungermann (1999) suggests that
there is an implied competence in a receiver to take advice. Hence, a caller who
requires a pragmatic response but receives a political justification would at best be
confused by such a reply. Without the benefit of a face-to-face meeting, which
allows a broader range of interpersonal communications cues, call-centre agents have
to rely on much more subtle changes in intonation or use of qualified paraphrasing.
Harvey et al. (2000) suggest people are better at assessing the quality of advice given
than applying that advice. Many a parent laments the waywardness of their children,
who in turn have no doubt as to the quality of advice from the parent; rather the
inclination is to express their individualism in a way which directly challenges this
advice. The role of a public-sector call-centre agent is to provide quality advice.
Whether the caller acts on this advice is another matter.
The conflict arising from existing literature is that the reliability of tools such as exit
surveys needs to be tested against a known benchmark. While these traditional
measures have served the management of efficiency, it has been at the expense of the
Context of the Study and Review of Literature
13
development of effectiveness measures. In the call-centre environment, effectiveness
can be improved by measures relating to quality of advice. The ability to quantify the
quality of advice given to callers is a necessary precursor to the analysis of calls.
Given that a large volume of calls is to be analysed in this study, this process needs
to be efficient and repeatable. The analysis also demands a different perspective
such as employing a knowledge view of call-centre activities that will progress the
goal of applying appropriate effectiveness measures.
Call-centre operational staff have relatively junior positions and provide advice
based on both formal and informal knowledge acquisition processes (Kjellerup,
2004). Most common of the formal processes are training (which is provided to
varying degrees), and use of information databases, which more often than not
contain static information. Zack (2001) indicates this is quite an adequate response
to queries on quantifiable data where there is an unambiguous response easily found
by the advisor (e.g., “By what date do I have to have submitted a tax return?”), but
not so effective in qualitative areas which may be complex, uncertain, ambiguous, or
equivocal (e.g., “Should I sue a trader who has provided me with defective goods?”).
Goh (2002) suggests that a well-developed knowledge management strategy is
critical to business and therefore public-sector success. However, prior to investing
in such a strategy the question, “How do various knowledge management strategies
affect the quality of advice given by public-sector call-centre staff?” needs to be
answered. By refining models proposed by contemporary knowledge management
practitioners and others with expertise in the measurement of quality of advice, a
framework will be developed to provide confidence in realising organisational
benefit from investment in knowledge reuse. Knowledge reuse is a specific area of
interest in knowledge management which Markus (2001) suggests provides
successful outcomes when allied to information needs.
2.2.3 Defining Knowledge Management
In their critique of knowledge management, Baker and Badamshina (2002) suggest
that the domain’s definitions tend to be so broad and vague as to have little meaning,
and in other cases too narrow in focussing on mechanistic and sequential process
directed at knowledge creation, capture, sharing, and (re)use. Since Polanyi (1967)
argued the two aspects of “knowing what” and “knowing how” were present in any
instance of a person’s knowledge, cognitive psychologists have developed a view of
Context of the Study and Review of Literature
14
knowledge creation that involves a continuous interaction between implicit
knowledge (acquired without a conscious attempt to do so) and explicit knowledge
(as occurs when conscious learning strategies are applied) (Nonaka, 1994; von
Krogh, Nonaka & Aben, 2001; Nonaka & Takeuchi, 1995). Knowledge can also be
viewed as a process of simultaneously knowing and acting and focuses on the
application of expertise (Brown & Duguid, 2000). Another view defines knowledge
as an object that can be viewed as capable of being stored and manipulated (Markus,
2001). Finally, knowledge can be viewed as a capability with the potential for
influencing future action (Carlsson, 2001).
Given the variety of interpretations of the concept of knowledge and its application,
it is expected that a similar variety of explanations exist for knowledge management.
Alavi and Leidner (2001) identify differing implications for knowledge management
based on the perspective held on knowledge. Knowledge viewed as an object implies
knowledge management should focus on building and managing knowledge stocks
and using information systems to store, gather and transfer it to users. Knowledge as
a process requires knowledge management to focus on knowledge flow - the
processes of creation, sharing, and distribution of knowledge. Knowledge as a
capability means knowledge management strategies build on core competencies, and
create intellectual capital. The predominant position taken in this study is that
knowledge is a capability since it actually addresses the link between knowledge,
knowledge management and organisational performance.
The Standards Australia AS 5037(Int)-2003: Knowledge Management Standard
(2003) states that :
“[knowledge management] Is a multi-disciplined approach to achieving
organisational objectives by making best use of knowledge. It involves the
design, review and implementation of both social and technological processes
to improve the application of knowledge, in the collective interest of
stakeholders.” (2003, p. 1)
This definition is consistent with the previously discussed material. The extension of
these concepts and the subject of many texts is the development of the tools and
processes that allow an organisation to systematically create, capture, share, and
leverage this knowledge. Each opinion leader has a comment which challenges or
supports the view that knowledge is able to be managed, and puts into perspective
Context of the Study and Review of Literature
15
the role of knowledge in the performance of an organisation. A consistent theme is
the human aspect – that a person needs to take stimuli from a range of sources and
process them in such a way as to improve or add value to what was previously
known, or to translate this into action. The better an organisation exploits this
capability, the better it is able to meet its business objectives (Choo, 1996).
In their seminal work in knowledge management, Davenport and Prusak (1998)
define “knowledge” in terms of the work of people’s minds trying to incorporate
experiences, values, insight and context to a new experience.
“Knowledge is a fluid mix of framed experience, values, contextual
information, and expert insight that provides a framework for
evaluating and incorporating new experiences and information. It
originates and is applied in the minds of “knowers”. In organisations, it
often becomes embedded not only in documents or repositories but
also in organisational routines, processes, practices and norms.”
(Davenport & Prusak, 1998, p. 5)
Davenport and Prusak (1998) then go on to show how knowledge can be used to
improve organisational performance. By annotating features of a project carried out
by British Petroleum, they list the underpinning principles of knowledge
management as including concepts of sharing, trust, technology as an enabler,
rewards, measurement, and creativity. This pragmatic approach avoids the
philosophical debates which arise from taking an abstract view of knowledge (and
subsequent investigation of the meaning of truth) and identifies the role of
management to ensure knowledge assets are used to produce organisational benefits.
Hence, the knowledge that is peculiar to the organisational context is to be exploited
by (knowledge) managers to improve performance, action, and outcomes (Pentland,
1991). Thompson and Walsham (2004) analyse Blacker’s (1995) five images of
knowledge, and conclude that most organisational change programs seldom address
these contextual components simultaneously. The components are defined as
embrained – or latent mental potential; embodied – or historically developed filters
and routines; encultured – or convergent expectations about intentions of others;
embedded – or organisational alignments; and encoded – or explicit symbolic forms.
The degree of simultaneous consideration of all these factors by knowledge
managers provides an organisational knowledge profile that is unique and distinctive
Context of the Study and Review of Literature
16
to a firm and provides its relative competitive advantage (Tsoukas & Vladimirou,
2001). The flow of knowledge within a firm is a function of the encultured
contextual component and is determined by a knowledge market (Davenport &
Prusak, 1998), which, like the well-studied commodity markets in economics,
determines the value of knowledge. The management of this market determines the
behaviours relating to knowledge strategies.
Competitive advantage of organisations in the twenty-first century will be obtained
through creating and sharing knowledge (Senge, 1997). The ability to integrate
theory, core staff capabilities, and institutional learning through best practice will
ensure sustainability of the organisation. Leonard (1995) provides a knowledge-
based model of core capability improvement for an organisation. She identifies
capability gap (1995, p. 138) as that gap between what the company knows and what
it needs to know in order to be productive and competitive. She refers to employee
knowledge and skill, technical systems, managerial systems, and values and norms of
an organisation as dimensions of core technological capabilities that provide strategic
advantage. Deficits may be met by in-house development of knowledge, brought in
via consultants/technology partners or other strategies such as takeover/strategic
partnering with competent firms. The concept of stolen ideas is also considered as a
legitimate process for capability enhancement. Not only do the ideas have to be
replicated, but improved functionality or creative application is required in order to
exploit the innovation that results from enhanced organisational capability. Sony
Corporation provides an example of the way it successfully gained market share
from other games machine manufacturers by adding games controllers that vibrate.2
This is the corollary to obtaining competitive advantage through the inimitability of
products or services. Opportunities for the public sector to exploit intellectual
property of others are restricted, since any behaviour seen as being remotely
dishonest causes considerable ethical and political concerns. Given the nature of
public sector organisations, options for bridging capability gaps are primarily limited
to internal capability improvement. They are focussed on three components of a
knowledge lifecycle identified by Wiig (1999): namely, that of originate/create
2 The Rumble Pack case (http://news.bbc.co.uk/2/hi/business/4387045.stm) refers to a plagiarism row between Sony and Immersion, a small, California-based developer of digital touch technologies, over game controllers that vibrate.
Context of the Study and Review of Literature
17
knowledge or generation; capture/acquire knowledge; and utilisation, which
combines the remaining phases of transform/organise knowledge, deploy/access
knowledge, and apply knowledge. Figure 1 illustrates the relationships and
organisational knowledge transitions among these phases (Rus, Lindvall, & Sinha,
2001).
Figure 1 - Organizational Knowledge Evolution Cycle (Rus et al., 2001, p. 17)
2.2.4 Knowledge Generation
Knowledge generation is taken as the starting point in the improvement of
organisational capability (Davenport & Prusak, 1998). In keeping with the
knowledge market metaphor, the availability of capable staff is the significant
limiting resource in the public sector. Public policy development requires
considerable research and innovative thinking, particularly in response to the
challenge of winning back the trust of citizens (Davis & Weller, 2001). Selecting
existing staff for the purpose of innovation, providing appropriate challenges, and
recognising and rewarding them for their work are widely utilised strategies in
development of individual capabilities (Peters & Austin, 1985). Support for internal
staff selection exists in one of the presuppositions underpinning Neuro-Linguistic
Programming (NLP)-oriented training that suggests people are ultimately able to
draw on their own capabilities to address any situation, provided they are aware of
their capability and know how to use it (Bradbury, 2000). Leonard (1995) also
addresses this knowledge acquisition, and introduces the concept of prototyping and
experimentation. She presents a research and development model consistent with
Context of the Study and Review of Literature
18
earlier themes, indicating the need for knowledge generation in order to maintain a
sustainable organisation. The tying of experimentation to corporate resilience is
tendered by Leonard (1995) as an alternative to long-term strategic planning that
may be incapable of coping with uncertain (future) business environments. Such
experiences produce actionable knowledge (from both success which motivates to
further develop and failure which provides an opportunity to examine alternate
strategies), which in turn contributes to the core capability and hence sustainability of
the organisation.
Two apparent dilemmas are raised by the literature in regard to knowledge
generation. The first is that if there is too much comfort among staff, there is no
incentive to innovate. One strategy able to be employed here is to intentionally
“…try to evoke a sense of crisis among organisational members by proposing
challenging goals” (Nonaka & Takeuchi, 1995, p. 28) The introduction of creative
tension through challenges, either perceived or actual, or by formation of teams
composed of staff with different skills and values can create the “creative chaos”
referred to by Nonaka and Takeuchi (1995). The other apparent quandary is that
success is often the greatest enemy of innovation. This is a subset of the initial
dilemma discussed, with the point being that success often creates comfort or a sense
that past practices will continue to be successful into the future. This is a lesson IBM
has learned, although for some time grave concern was held for the company that had
defined technology in the workplace for much of the century. During the 1980s
IBM's mainframes suffered from competition from more flexible UNIX servers. The
company responded by channelling massive resources into attempting to stabilise its
market position with its traditional products and services. It sustained multi-billion
dollar losses. Under new leadership, rather than promoting any particular product or
content, IBM focused on end-user services such as assisting companies to transform
themselves into secure Internet-based trading communities ("Blue is the colour.,"
1998).
2.2.5 Knowledge Capture
Knowledge capture is the second key systematic process identified as a critical
organisational capability. Differing from the generation of knowledge, technology is
seen to provide a tool to codify knowledge in such a way that it becomes useful in
other circumstances (Davenport & Prusak, 1998). Dixon (2000) introduces the
Context of the Study and Review of Literature
19
concept of translation or modification where the same issue is likely to be repeated,
but in different localities and under differing circumstances. Hansen et al. (1999)
assert that a firm’s knowledge strategy should reflect its competitive strategy; it
should have an 80-20 mix of either codification or personalisation methods,
specialising in only one of the areas, but without ignoring the other. Codification
strategies, which rely heavily on investment in IT and reward for people contributing
to the document databases, are more aligned with maintenance of high revenues
through reuse of the knowledge asset. High profit goals driven by a strategy of
personalisation requires the development of person-to-person networks and reward
for direct interaction with others. Simply capturing the information does not,
however, make it useful. Thompson and Walsham (2004) support an integrated
strategy, but extend the model to include embrained, embodied, encultured,
embedded and encoded contextual components. Bryar (2001) raises issues of
information overload, but suggests that the combination of a well-defined taxonomy
and associated technology at least ensure information can be relocated if correctly
classified and tagged for metadata searching. Attempting to excel in both
personalisation and codification strategies increases risk of failure, while ignoring
either strategy has a similar impact (Hansen et al., 1999).
2.2.6 Knowledge Reuse
The final aspect identified in the literature refers to the sharing of knowledge and its
socialisation: the process whereby tacit knowledge is able to be transferred between
individuals and organisational groups through shared experience, space, and time.
This important value-creating process allows the organisation to obtain competitive
advantage through better decisions or quicker time to market (Handzic &
Chaimungkalanont, 2004). Knowledge sharing is referred to by various terms
including “knowledge transfer” (Davenport & Prusak, 1998; Dixon, 2000; Goh,
2002; G. Timbrell, Andrews, & Gable, 2001), and “knowledge reuse” (Kucza, 2001;
Markus, 2001). Davenport and Prusak (1998) suggest organisational value is created
by this knowledge transfer when the two actions of transmission (sending or
presenting knowledge to a potential recipient) and absorption (by that person or
group) improves its ability to perform new actions, innovate, or change behaviours.
Value is obtained through the efficient reuse of knowledge in the execution of
repetitive tasks or solving similar problems (Majchrzak, Neece, & Cooper, 2001).
Context of the Study and Review of Literature
20
Markus (2001) identifies four knowledge reuse situations based on the purpose of the
knowledge and the role of the seeker and refers to them as shared-work producers,
shared-work practitioners, enterprise-seeking novices, and secondary data miners.
This is validated by Timbrell et al. (2003) with the identification of a further group
identified as “primary data miners”. Given that call-centre agents predominantly act
as individuals, Markus’s (2001) model provides a framework for focussing further
study on knowledge sharing behaviours by government call-centre operators, in an
attempt to improve contextual problem solving both at the individual and team
levels. Figure 2 illustrates the role of shared problem solving as a core capability,
necessary to address the increasing complexity from globalisation and the
proliferation of formal education (Leonard, 1995). Provided an appropriate
repository exists, benefits accruing from knowledge reuse in a call centre are
reflected in:
• Reduction in search time through the application of informed search
strategies;
• Consistency of responses through alignment of past scenarios with current
queries; and
• Continuous improvement in quality of response through opportunity to reflect
on knowledge sources, resolution strategies and robustness of advice
(Leonard, 1995, p. 60).
Figure 2 - Capability-Creating Activities: Shared Problem Solving (Leonard, 1995, p. 60)
PRESENT
INTERNAL
FUTURE
Shared Problem Solving
Improving Knowledge
Experimenting
Implementing and Integrating
EXTERNAL
Core
Capabilities
Context of the Study and Review of Literature
21
Shared problem solving is not the recognised method of operation of Queensland
Government call centres where a traditional management paradigm requires
individual agents to take responsibility for their client callers. However, in order to
improve performance through knowledge reuse, the contribution by individual agents
impacts on the overall problem solving performance of the call centre. All of the
reuse situations defined by Markus (2001) and Timbrell, Nelson, and Jewels (2003)
are applicable to this study of knowledge reuse in call centres. The following
analysis suggests that shared-work producers are able to improve efficiency through
codification; shared-work practitioners are significant players in current knowledge
reuse; expert-seeking novices (both callers and less experienced agents) need to
develop a range of knowledge sources and search strategies; while primary and
secondary data miners are more likely to perform data mining activities in an attempt
to identify strategic activities to improve efficiency and effectiveness of call-centre
operations. These differing knowledge transfer strategies are addressed by role which
for some individuals may change depending on the type of task allocated at the time.
Shared-work producers are categorised as a homogeneous team producing
knowledge for their own later use. They also have a need to know the “what,”
“how,” and “why” of previous decisions. Although primarily gathered for personal
use, benefits would accrue to all agents if the embedded knowledge were accessible
to all call-centre staff, since they frequently keep good (personal) records as a by-
product of their work, but often omit the rich context surrounding the incident, which
is a necessary component for extracting learnings for future projects. Anton (2000)
refers to the proportion of customer intelligence available to call-centre staff as
strongly favouring static information.3 Information is collected by call-centre agents
in their informal (personal) working notes, by filling in forms (both paper- and
electronic-based), and in formal reports resulting from research required to answer
the caller’s query. Utilisation of this range of data collection methods is not
universally adhered to. In terms of efficiency, performance measures reduce the
motivation to either codify or share such information in a formal way, thus reducing
3 Modern technology is more able to deliver identification details of a caller to the desktop, but is expensive to implement and maintain. At the lowest level, intelligence provides the caller’s phone number. If linked to customer management systems (CMS), further information such as name, address, preferred products, previous call history, etc. are able to be accessed.
Context of the Study and Review of Literature
22
the benefit from knowledge reuse and the consequential increased service quality
opportunity (Anton & Gustin, 2000).
Shared-work practitioners do similar work, but may be employed in different
environments. They acquire new knowledge from external sources and have a need
to know how and why a particular procedure works. Several roles within the call
centre reflect this reuse situation, since the establishment of call centres addresses the
modern organisation’s need to balance flexibility and stability with increased
openness to customers’ needs and the protection of core practices and routines
(Arzbacher, Holtgrewe & Kerst, 2002). The static information databases, pamphlets,
and other cues used by agents are developed by experts from outside the call centre,
and may not have the call-centre agents as their target audience.
Part of the expertise of the shared-work practitioner is the ability to locate material
relevant to the query in a timely way. For this reason, initial training for call-centre
staff focuses on the ability to locate specialist knowledge, either through codification
processes (use of Yellow Pages-type directories) or personalisation processes
(attendance at functional work group meetings, for example) (Hansen et al., 1999). A
number of call centres have a tiered structure where reference to tier-2 staff will
facilitate the application of their specialist knowledge to the query. These shared-
work practitioners are able to improve the efficiency of the call centre by sharing
their learnings (in the categories of close knowledge distance4 and most useful) with
other specialists, as well as other agents who may be able to exploit it at a later time.
Alternatively, functional specialists within the call centre may have calls directed to
them by integrated voice response (IVR) technology. In all cases, the shared-work
practitioners take calls referred by call-centre staff and carry out more detailed
activities, which often relates to problem resolution. It is likely that as performance
pressures impose on call-centre staff, more referrals are made to these practitioners.
Business strategies for the use of call centres vary from firm to firm, and will range
from clearing house for routine transactional processes (e.g., renewal of licenses and
4 Knowledge distance is a term used by Markus (2001) and Timbrell et al. (2003) when referring to the gap between what a person knows and what they need to know in order to perform some task or solve some problem. In this instance, if a person makes notes about a particular event, they know the context in which the information was collected and hence will know how to apply it in other similar situations.
Context of the Study and Review of Literature
23
permits), to complex problem-solving roles such as the use of an IT helpdesk. An
appropriately selected knowledge management initiative tied to the firm’s business
strategy will improve the ability of front-line staff to manage a wider variety calls,
thus leading to a more efficient call-centre operation (Zack, 1999).
Expert-seeking novices seek answers to questions not previously encountered by
them, and acquire knowledge through advice from other experts. Often operating
without a capability to identify the relevance of contextual information, expert-
seeking novices may have difficulty defining the problem at hand, and may rely on
knowledge designed for novice audiences, or on access to specialists. This role
category equally applies to callers and novices within call centres. In technical areas,
manuals are seldom written for the novice, which leads users with a problem (callers)
to seek assistance from a service provider. The primary reason for the existence of
many “helpdesk”-based call centres is to provide advice and assistance. In his study
of IT helpdesk software support actions, Pentland (1991) distinguishes between a
problem and a question. He suggests questions may well be better answered by
reference to available information sources such as program documentation. This is a
problem defined by Zack (2001) as one of uncertainty. In many cases, the caller
probably has easy access to such information, but will call either due to lack of
confidence in locating the required information, or inability to interpret it in a
meaningful way. Problems relating to software support, which Zack (1999) defines
as complex, ambiguous, or equivocal, are the domain of the helpdesk, where experts
can apply diagnostic processes in order to resolve the issue causing concern. The
knowledge distance in these instances is much larger than when discussions relate to
operating instructions (since the call-centre agents are technical experts and the
callers primarily applications users), and is complicated by the reliability of the
information supplied by the caller, and the understanding of the resolution
instructions. In their study of a Greek call centre, Tsoukas and Vladimirou (2001)
also discovered the difficulty of communication between customers and operators
due to the lack of sophistication of the caller.
Consideration must also be made for the expert-seeking novices who exist within the
staff of the call centre. Due to the turnover in staff in call centres, there is an
abundance of these novices who have a knowledge of the local context, but are
unlikely to know which aspects of this context are important (Tuten & Neidermeyer,
Context of the Study and Review of Literature
24
2004). Knowledge of the operations is gained on the job through degrees of training,
but more likely through informal networks of perceived experts, with that decision
being based as much on personality as actual expertise. Strategies based on advanced
technology are often cited as an effective option for call centres, and are reflected by
comments such as:
“One major component of the Government Call-Centre project is to put in
place the necessary hardware, software, and networking infrastructure to
support the various call centres…The customer service officer is able to
access back-office host computer systems and databases.” (Milner, 2000, p.
109)
Milner (2000) goes on to predict that call centres and the supporting infrastructure
will be at the forefront of public sector strategies within the next ten years, so as to
deliver client-focussed modes of operation. The benefits of this investment are not
well identified in research literature on call-centre performance, however, although
Anton (2000) predicts that the call centre of the future will integrate the Internet with
calls to allow the caller to see the agent servicing them. This technology will lead to
reduced talk times and time to answer, as well as service quality improvements in
accuracy of information (both from caller and agent).
Secondary data miners attempt to develop new knowledge though analysis of
records. This study, which accesses, analyses, and reflects on documents created in
the call centres among other things, aims to develop new knowledge able to be
applied in more general advice-giving situations. It thus satisfies the secondary data
miner typology. In traditional call-centre environments, analysis has generally
referred to the efficiency measures of average speed of answer, time-in-queue,
resolution averages, total calls, and a range of other quantitative outputs which are
generally easily extracted from call-centre management software. Timbrell et al.
(2003) refer to call-centre managers as primary data miners because of the close
“proximity to knowledge context” and that they reuse this knowledge to discover
processes to improve performance or remove activities that may be in conflict with
business goals. Feinberg, Kim, Hokama, de Ruyter & Keen (2000) suggest that
interest in these measures may simply be an artefact of the technology since they
become important because they can be measured. Their research shows that fast
access to the agent and successful resolution of the caller’s problem are the essential
Context of the Study and Review of Literature
25
elements of an effective call-centre operation. Milner (2000) extends this concept by
suggesting that analysis of call-centre operations needs to provide senior
management with awareness and knowledge of client needs and the capability of the
organisation to deliver on these. This suggests a degree of sophistication above that
of mere qualitative analysis of traditional sets of output metrics, and may well
provide the competitive advantage achieved through call-centre agents becoming
“truly knowledge workers” (Anton & Gustin, 2000). The ability to reuse knowledge
in a call centre lies in its capacity to connect “want-to-knowers” efficiently with
“knowers” (or at least access to documents and a context to make sense of the
information contained within them). Contemporary call-centre managers are
responding to the changing business environment5 through their desire to develop a
capability in their agents to recall reasons for decisions (or the “when, what, how,
and why”) – a classification referred to as rationale knowledge by Markus (2001, p.
62). It is reasonable to assume significant benefits will accrue to call centres able to
put such knowledge management practices in place.
2.2.7 Defining Quality of Advice
The nature of a call centre is that it provides relative anonymity to the caller.6 This
may influence the quality of advice given since environmental factors affect the
quality of advice given (Rogers, Hassell, Noyce & Harris, 1998). Rogers et al. (1998)
found that lower socio-economic groups received a lower standard of advice in
community pharmacies than those of higher status groups. In a call-centre
environment, unless a caller provides details to identify characteristics which may
introduce bias, a caller’s identity should not influence the advisor. However,
telecommunications innovations are able to deliver significant profile information to
a call-centre desktop which may include caller identification (from the
telecommunications carrier) and other profile information from internal systems
(e.g., if the caller telephone number is linked to company transaction databases)
(Anton, 2000). The agent’s advice can then be tailored to fit the caller’s profile,
5 Shorter product life cycle, global competition, and superior service quality are some of the challenges faced by organisations in their quest to sustain a competitive position in the marketplace (Zairi, 2002). 6 A caller may withhold identifying information and block technical triggers (such as call identification), but may be required to provide unique qualifiers that allow identification (such as Medicare number), if material benefits are being negotiated.
Context of the Study and Review of Literature
26
which is advantageous, but may also lead to degradation in quality if socio-economic
bias is introduced.
A determination needs to be made as to who defines the quality of the advice given.
Callers who receive advice contrary to what they see as acceptable are likely to seek
a second opinion elsewhere. This is not at issue in this study, since the public sector
organisations studied are effectively monopolistic. Its significance is rather that a
reliable measure of quality is available and that as strategies are put in place to
improve the advice, impacts are able to be reported. A simple resolution to this
situation is that an expert is used to rate the advice being given, which is a trivial task
in terms of monitoring, but a complex one in terms of resource allocation and impact
on employees being monitored. Jungermann (1999) provides a model for advice-
giving. He asserts that advisors attempt to match options to the clients, rather than
provide a detailed analysis of issues being considered. This may be in order to satisfy
the person who is requesting the information, or it may result from the advisor
prejudging a decision about the competence of the caller to understand the range and
complexity of issues being considered. Jungermann (1999) also analyses the take-up
of advice based on the opinion-related and personality-related attributes of both the
giver and receiver, and how they relate to degrees of expertise, credibility, and
confidence. Harvey et al. (2000) provide some hope for using the receiver of advice
as a reliable source of information able to contribute to quality measures, in that they
conclude that a receiver of information is better at assessing the quality of advice
than using it. If it were possible to obtain post-advice information independent of the
need of the receiver to use that advice, such data may provide a guide to its quality.
Brogowicz et al. (1990) provide a service quality model synthesised from two
sources referred to as “The Nordic School” and ”The North American School”. The
model is multi-dimensional and separates technical quality (staff, systems, etc.) from
function quality (policies and procedural issues) to provide a service gap instance.
The smaller this gap, the better the perceived quality of service. The determinants,
able to be empirically defined, include “tangibles, reliability, responsiveness,
assurance and empathy” (Brogowicz et al., 1990, p. 33). Courtesy is another term
used to cover these interpersonal interactions (Tsoukas & Vladimirou, 2001).
Pragmatic thinking towards exceeding customer expectations in marketing generally
supports the strong positive relationship between customer and service provider
Context of the Study and Review of Literature
27
(Estelami & Maeyer, 2002). Brogowicz et al. (1990) suggest competitive advantage
can be obtained through a strategy aimed at decreasing customer expectation, rather
than increasing quality of service. This is an interesting approach and may have been
used for a long time in the public sector, given the degree of esteem with which it is
held. The aim of this study is, however, to identify a range of strategies able to be
applied in an advice-giving environment with a high degree of confidence that they
will deliver performance improvement. Unfortunately, the actual levels of quality as
determined by accuracy, timeliness, and appropriateness tend to become enmeshed in
the perceptions of the callers – whether owing to issues of courtesy, empathy, or
expectation of a beneficial result.7 The concept of service value is investigated by
Dedeke (2003). Figure 3 shows the relationship between service quality, service
value, and client satisfaction. Dedeke (2003) establishes that customers may tolerate
poor service in certain circumstances, and be dissatisfied with excellent quality in
others. In areas of complaints management, these apparently irrational outcomes are
highly probable.
Figure 3 - Relationship among service-quality concepts (Dedeke, 2003, p. 280)
7 An aggrieved caller may call looking for an outcome which may not be legally or morally appropriate. In this case, no matter what the advice, the caller will not assess it as being of high quality.
Service Value Behaviours
Service
Quality Satisfaction
Perceived
Costs
Attitudes
Attachments
− Performance quality of outcomes
− Performance quality of processes
− Performance quality of process system
− Sacrifices
− Direct costs
− Inconveniences
Context of the Study and Review of Literature
28
2.2.8 Knowledge Strategies to Improve Advice-giving
The drive for efficiency in call centres, while delivering better service quality, has
concurrently caused the knowledge required to do work such as claims processing to
be embedded in systems. This requires prescribed patterns of interaction, but
minimises relationships both within and beyond the organisation (Cross, Liedtka &
Weiss, 2005). The focus of such systems changes from distinct roles (as
characterised by staff external to the call centre) to one of process flows (role
anonymity). Routine response networks, which rely on scripted and easily located
knowledge sources, are structured to reduce effort in resolution of repeat service
calls. They are also structured to locate experts for second-tier referrals quickly to
address well-structured problem types having minor degrees of abstraction. Zack
(2001) refers to these problem types as uncertain (i.e., “having insufficient factual
information about the goal, situation or task, and some lack of confidence in the
consequent inferences, estimates, or predictions required” (Zack, 2001, table 1)).
However, since they are unable to instigate required cognitive and affective
processes, such routine response networks become ineffective in problems which are
complex (i.e., with “too many situational elements and relationships to coordinate or
consider simultaneously” (Zack, 2001, table 1)); ambiguous (i.e., showing
“inadequate knowledge about, no explanation for, or understanding of a goal,
situation, or task” (Zack, 2001, table 1)); or equivocal (i.e., providing “multiple
interpretations of a goal, situation, or task” (Zack, 2001, table 1)), since by definition
it is not possible to define a single process flow to address these problem types
(Jonassen, 2000). Cross et al. (2005) conclude that management needs to develop
strategic views on social networking methodologies for the purposes of knowledge
reuse/referral services, which are best suited to their business goals. A modular
response network, consisting of experienced practitioners, provides the capacity to
solve complex problems through re-sequencing and reuse of knowledge acquired in
previous (similar) situations. The defining characteristic of modular responses is their
ability to disaggregate complex problems into smaller ones that are both familiar and
solvable by the experienced network members. Resolution of unstructured, ill-
defined problems (ambiguous and equivocal) is enhanced through the development
of customised response networks that link experts capable of developing innovative
solutions.
Context of the Study and Review of Literature
29
Core capability to enable innovation is also discussed by Leonard (1995). She
extends her concept of capability (or capability gaps) to the problem-solving arena
and concludes that shared problem solving is a critical competency of a successful
organisation. She warns that core rigidities established in prior successful problem-
solving situations can cause dysfunctional behaviours that hinder innovative
thinking. Leonard (1995) thus recommends the introduction of creative abrasion,
controlled introduction of tension into the workplace, to expand the available
knowledge resource. The management problem that arises from her work is to
balance the benefits accrued from the application of solutions from highly skilled
staff with blockages that come from emotional attachment to their mindsets and
problem-solving biases.
Leonard (1995) addresses the primary purpose of problem solving for innovation; a
broader analysis of problem is required to address all the activities of organisations,
since a knowledge-based view of a firm recognises the multi-dimensional nature of
knowledge. Zack (2001) contends that “Different literatures have addressed only
portions of this knowledge-problem space” (Zack, 2001, p. 17). He suggests
knowledge is an organisation’s most enduring and strategic resource, and presents a
taxonomy of problem types (complexity, uncertainty, equivocality, and ambiguity) to
integrate work done by other researchers. These problem types are explored further
below.
Complexity
An organisation’s response to complex problems, as those with “many interrelated
variables, solutions and methods” (Zack, 2001, p. 18), is determined by its capability,
and includes actions to:
• Develop richer knowledge in order to allow issues to be dealt with as a
familiar single problem; and
• Reduce complexity through decomposition to simplified (solvable/familiar)
components.
Uncertainty
Problems of uncertainty, characterised by the “lack of information or factual
knowledge about current and future states” (Zack, 2001, table 1), are addressed
through:
Context of the Study and Review of Literature
30
• Acquisition of additional knowledge; or
• Improvement in knowledge and ability to predict, infer, or estimate.
Ambiguity
Ambiguous problems represent “an inability to interpret or make sense of” (Zack,
2001, table 1) a particular situation. Tsoukas and Vladimirou (2001) suggest
ambiguous problems occur when callers are unclear of what they want and therefore
often provide redundant information. Skilled operators guide callers to articulate
their problems clearly, which enables an appropriate response by the agent. Zack
(2001) provides the following strategies to solve ambiguous problems:
• Reframing the problem to a more meaningful situation;
• Iterative cycles of incremental points of clarification; and
• Interactive face-to-face conversations.
Equivocality
Problems which appear to have “multiple meanings for or interpretations” (Zack,
2001, table 1) of the particular situation are referred to by Zack (2001) as problems
of equivocality. Coordinated action is required to resolve the range of interpretations
in a problem of equivocality by:
• Cycles of interpretation;
• Interactive discussion; and
• Negotiation to converge on one meaning.
Weick (1969) states that there is a requirement for an organisation to disrupt order to
make it possible to create order: to take information in equivocal form and transform
it into one that is unequivocal. He suggests an inverse relationship exists between the
degree of equivocality in a situation and the number of “rules” available to resolve
the problem. Since this process is cyclical, Weick (1969) finds an inverse
relationship exists between the number of rules available and the number of cycles
required to simplify unequivocal problems, as is represented in Figure 4. This
implies that a highly equivocal problem will have few rules (either formal or
heuristic) available, and hence would require many incremental cycles to define an
agreed meaning of the situation accurately.
Context of the Study and Review of Literature
31
Figure 4 - Rules and cycles in resolving an equivocal problem (Weick, 1969, p. 77)
Zack (2001) further develops a matrix, displayed in Table 1, which relates the
knowledge types (declarative, procedural, causal, and relational) to a range of
knowledge states determined by problem type. This presents both a taxonomy for
codification of requests to the call centre, and guidance in the strategic management
of knowledge needed to satisfy such requests. The matrix of 16 cells implies a
targeted knowledge management strategy for each cell, with the impact of such a
strategy able to be assessed in terms of its effectiveness in improving capability to
resolve such problem types.
Amount of equivocality in input
Number of rules used to assemble the process
Number of interlocked cycles selected for application to input
Amount of uncertainty removed from input
Inverse relationship Inverse relationship
Inverse relationship Direct relationship
Context of the Study and Review of Literature
32
Knowledge State
Knowledge Type
Ambiguous Equivocal Uncertain Complex
Declarative (know about)
No schema for classifying or describing something
Multiple possible classifications; or vague boundaries between classes
Classified or described within some level of confidence
Classified with confidence, but within a schema having many intricately related attributes
Procedural
(know how)
No script for describing a procedure or event sequence (e.g. tacit skill)
Multiple possible procedures or sequences
One procedure defined within some level of confidence of appropriateness or predictability of sequence
One procedure or sequence defined with confidence but having many intricately related steps
Causal (know why)
No explanation for why something occurs or occurred
Multiple possible explanations
Once explanation proposed within some level of confidence of appropriateness or predictability of cause/effect
One explanation defined with confidence, but having may intricately related causes and/or effects
Relational (know with)
No understanding of relationships among elements
Multiple possible relationships among elements
One set of relationships defined within some level of confidence or predictability
One set of relationships defined with confidence, but having many intricately related elements.
Table 1 - Knowledge States (Zack, 2001, table 2)
2.3. Summary
Initial research indicates a gap in the literature for the application and evaluation of
knowledge-based principles to call-centre situations. This research effort to address
the question “How well does Zack’s (2001) framework represent problem types, and
hence impact on strategies utilised in response to queries encountered in a public-
sector call-centre environment?” will make a contribution to the literature. Markus
(2001) and Timbrell et al. (2003) describe the characteristics of actors and their roles
in knowledge generation, capture, and reuse; Brogowicz (1990) provides a
framework for investigating service quality; and Zack (2001) presents a taxonomy of
problem types likely to be encountered. The existence of general principles and
practices for client interaction, problem resolution, and knowledge management
implies the existence of a conceptual matrix which informs response strategies to
customer queries that address these generic categories to produce optimal outcomes
for both caller and the service provider. Hence, the desired outcome from this
Context of the Study and Review of Literature
33
research project to improve quality of call-centre advice resulting from the
application of targeted knowledge strategies will include a range of strategies
targeting problem-types specific to the organisation.
Research Method
34
Chapter 3
Research Method
3.1. Introduction
Chapter Two explored a range of contemporary theoretical models that provide
guidance in the application of principles of knowledge management in the
improvement of advice-giving in a public-sector environment. This chapter focuses
on the selection of a research methodology and an associated protocol to ensure that
a logical and defensible approach is applied. The aims of the research and adopted
approach are initially articulated. The characteristics of customer needs and
classification of their problems is then investigated in order to inform the data-
gathering component of the protocol along with activities to identify responses of
call-centre agents. Finally, the practicalities of encoded data gathering are discussed.
3.2. Objectives of the Research
This research project aims to provide evidence that quality of advice given by call-
centre staff can be improved through judicious application of a model both able to
predict the outcomes of investment in a knowledge management strategy, and to
measure the impact of such a strategy. The research question is:
“How well does Zack’s (2001) framework represent problem types, and
hence impact on strategies utilised in response to queries encountered in
a public-sector call-centre environment?”
More specifically, the study aims to:
(1) Develop a model that defines the interrelationships between customer queries
and call-centre knowledge management strategies to provide subsequent
responses;
(2) Suggest knowledge management strategies to improve appropriate knowledge
reuse within call centres;
(3) Focus discussion and promote constructive interaction for developing a
sophisticated understanding of the call-centre environment, and within the
public sector in particular;
Research Method
35
(4) Explore characteristics of, and implications for, systems-based initiatives and
knowledge management interventions that would impact upon a call centre’s
ability to realise service quality benefits;
(5) Inform academic research directions in call-centre management; and
(6) Provide practical advice to call-centre management.
3.3. Approach
The study has primarily relied on qualitative analysis. By focussing on two
organisations, participant observation and interviews have provided the opportunity
to capture current organisational behaviour. To develop the richness of the study
further, archival data have been randomly selected to support the real-time
observations. The criteria for selecting a suitable research technique for this study
have included:
• An exploration of what is actually happening compared to what executives,
management, and agents suggest is happening; and
• Aligning these events with the theoretical frameworks identified as probable
determinants of improved advice-giving in public-sector call centres.
Access to many years of archival data from one of the organisations has provided a
temptation to develop a strong analytical tool based on statistical methods. The
combination of the complete data set (qualitative observations plus the statistical
analysis of archived data) would deliver benefits beyond that of a pure case
methodology (Gable, 1994). However, the study is attempting to identify causal
relationships (such as: better problem-solving capability leads to better advice),
which are not well served through descriptive statistics (Kaplan & Duchon, 1988).
Hence, the archival data are used to validate activities identified through observation
and interview by following a pure case study method. Yin (1994) recommends the
use of a case study methodology in a situation that allows a more complete
examination of factors which may produce a specific phenomenon. It gives support
to questions of why events happened, and how they have impacted on a
contemporary social organisation. Yin (1994) suggests this approach is particularly
applicable if boundaries between context and phenomenon are not clearly evident.
The major mode of analysis recommended by Yin is pattern matching. The analysis
Research Method
36
requires the investigation of patterns as they occur in the case study, and comparison
of findings with the theoretical model being investigated. A significant correlation of
predicted patterns with empirical examples implies a high level of internal validity.8
This test for quality of research design is satisfied in the study through the
investigation and verification of Zack’s (2001) problem types, Timbrell, Koller et
al.’s (2005) Query-Response (QR) Cycle, and Brogowicz et al.’s (1990) quality
criteria.
Another test which establishes the quality of empirical research is construct validity.9
Yin (1994) recommends the use of multiple sources of evidence and identification of
a chain of evidence to achieve this goal. He identifies the strengths and weaknesses
of each of the following six sources of evidence, and represents the convergence of
these in Figure 5. These strengths and weaknesses are summarised as:
1) Documentation as a source of evidence is stable, unobtrusive, and exact, and
has broad coverage, but suffers from irretrievability, bias from selectivity and
author access may be a problem.
2) Archival records are similar to documentation, and although they tend to be
more precise and quantitative, they are also less accessible due to privacy
issues.
3) Interviews can be targeted and insightful, but may reflect bias and
inaccuracies due to recall and reflexivity.
4) Direct observations have strengths in their reality and ability to capture
context, but are apt to be expensive, time consuming and reflexive.
5) Participant observation is similar to direct observation and can provide
insight into behaviours and motives but again may suffer from bias.
6) Artefacts provide insight into cultural features and technical operations but
suffer from selectivity and availability (Yin, 1994).
8 Internal validity establishes a causal relationship whereby certain conditions are shown to lead to other conditions (as opposed to spurious relationships) (Yin, 1994, p. 33).
9 Construct validity establishes correct operational measures for concepts being studied (Yin, 1994, p. 33).
Research Method
37
FACT
Archival records
Documents
Observations (direct &
participant) Structured interviews
and surveys
Open-ended interviews
Focused interviews
Figure 5 - Convergence of Multiple Sources of Evidence (Yin, 1994, p. 93)
Open-ended interviews with senior management of the two call centres studied have
been used to obtain information on the strategic roles of the call centre, business
drivers, and performance expectations. Interviews structured on the knowledge roles
of participants, as defined by Markus (2001), investigate the operational aspects of
the call centres from the perspective of team leaders. Observations (including direct
coding of calls) have been carried out over an extended period, with supplemented
documents provided to support claims in relation to operational procedures.
Use of replication logic in multiple case studies provides external validity.10 This
methodology has been applied to two public-sector call centres. This provides
limited confirmation of the replication logic, and also provides the opportunity to
expand the scope of future studies to commercial and non-call-centre environments.
No generalisations will be claimed as a result of this study, apart from the
verification of the models under investigation in public-sector call centres.
Reliability,11 the final quality test identified by Yin (1994), is obtained through the
definition of a protocol and the utilisation of a database to collect data in a pre-
10 External validity establishes the domain to which a study’s findings can be generalised (Yin, 1994, p 33).
11 Reliability demonstrates that the operations of a study – such as the data collection procedures can be repeated, with the same results (Yin, 1994, p 33).
Research Method
38
determined format. The protocol is able to be followed in multiple case studies
because the application of theoretical frameworks defined in academic literature
supports the protocols used in this study. A forms-driven database has been used to
codify the calls. A pilot set of calls has been used to train both the researcher and the
research assistant in the codification of calls. The recorded data used in the research
has been coded twice, once by the researcher and again by the assistant, and later
reconciled to ensure consistency. Variations have been discussed and resolved
through revisiting the underpinning framework developed by Zack (2001).
High quality analysis in empirical research is accomplished through addressing the
following four principles that should attract the researcher’s attention:
• Show that the analysis relied on all the relevant evidence;
• Include all major rival interpretations in the analysis;
• Address the most significant aspect of the case study;
• Use the researcher's prior, expert knowledge to further the analysis. (Yin,
1994, pp. 123-124).
These principles are used in Chapter Six to converge the sources of evidence to give
a clear and lucid account of the knowledge strategies and their effectiveness in
public-sector call centres.
Carroll et al. (1998) describe a process to undertake case study research. This
methodology is iterative in that the research cycle (as shown in Figure 6) is able to be
repeated as many times as is necessary to validate the issues under investigation. The
cycle commences with the development of a conceptual framework, and is followed
by development of a research design (modified as needed in successive cycles), data
collection, analysis, and reflection. Outcomes from each cycle inform the conceptual
framework and may lead to the development of theory relating to the research topic.
The study activities have traversed the loop three times. The first cycle was used to
establish that the overall objectives of the research were likely to be achieved within
the initial scope of the project. The meetings with senior management in both
organisations indicated a deficiency in both academic literature and pragmatic
models to inform the performance measurement and capability improvement of
public-sector service provision. Access to work environments was approved subject
Research Method
39
to caveats which protect the confidentiality of government business and the privacy
of individuals concerned.12 This cycle was completed with a need to identify a solid
protocol to form the basis of future data collection. This protocol was developed and
tested in the second cycle. This process involved structured interviews (based on the
work of Markus (2001)), data collection,13 categorisation (based on the work of Zack
(2001), and Brogowicz et al. (1990)), and codification (based on the work of
Pentland (1991)).
Conceptual Framework
Theory & knowledge Research themes
Literature
Insights
Theoretical foundations
Analyse Collect
Data
Plan Reflect
Figure 6 - The Structured Case Research (Carroll et al., 1998, p. 65)
The final cycle required the researcher to revisit the objectives of the research
project. It further entailed a final investigation of the contemporary literature to
determine if any recent work had been developed. The conceptual framework,
referred to as the Query-Response Cycle (Figure 7), had developed further, and was
utilised to refine the focus of the literature searches and reviews. It provided a
knowledge-based view of call-centre activities, where callers consign their query to
the call centre in order to have their problem resolved (Timbrell et al., 2005b). The
12 Ethical issues were also covered by the ethical research approval process within the Queensland
University of Technology (QUT).
13 A forms-driven Microsoft Access database was developed to assist in the collection, categorisation, and codification of data.
Research Method
40
cyclical nature of this model implies a continuous pay-off assessment is performed
by the caller, who may choose to be content with the response received (resolved),
re-query when the knowledge distance has been measurably reduced by a previous
cycle and there appears to be a benefit from continuing (re-query), or drop out of the
cycle without resolution if the response is not converging to a solution
(abandonment).
Figure 7 -Query-Response Cycle (Timbrell et al., 2005b, p. 546)
Within the call centre, a strategy of reference to second and third tiers simply
commences a new Query-Response (Q-R) Cycle with the directing of the caller’s
problem to more specialised (expert) agents. As reported by Timbrell et al. (2005b),
this escalation represents additional (opportunity) cost to the organisation through
employment of labour of specialists, and the diversion of these specialists from other
activities. Queries may be consigned several times. Each consignment starts another
Query-Response Cycle (Figure 8) until the resolution is passed back, either directly
or via the tiers, to the original querist. The elements on the left-hand side of the
Query-Response Cycle model, i.e., query reformulation, search strategies, use of
knowledge sources, creation of the response set, and the final offering of one or more
Research Method
41
responses, will vary according to the maturity of the knowledge infrastructure. The
actions of the respondent will vary according to the perception of service quality of
the querist (Brogowicz et al., 1990). The type of knowledge problem entering the
cycle has an effect on the knowledge processing and subsequent consignments to
more expensive knowledge resources such as expert-group tiers. Variations to this
model include referral of an action back to the caller (such as returning to the
problem source to gain more information), or with the agent taking the problem off-
line to carry out research free of the restriction of the telephone-bound
communication. These variations are specific to the search strategies and knowledge
resources represented in the left section of the model.
Figure 8 - Tiered Query-Response Cycles (Timbrell et al., 2005b, p. 550)
A simple model of the service process was also devised by Pentland (1991).
Although not cyclical, he cautions that, “While the steps in the process can be
thought of as though they were linear, this is not always the case” (Pentland, 1991, p.
92). The process for software support is defined as
“Opening ⇒ Classifying ⇒ Assigning ⇒ Diagnosing ⇒ Responding ⇒ Closing.”
Research Method
42
Opening (which equates to the query in Timbrell et al.’s (2005) model) is the process
of getting basic information about the caller, identifying the problem, and arriving at
the mutual agreement that the call centre will be able to assist with the problem.
Classification (Timbrell et al.’s (2005) query reformulation) is the next step towards
assembling a response. Pentland (1991) notes that:
“Even when an entitled caller states explicitly what he or she wants,
further discussion may reveal that something else is actually the problem,
or that there are multiple problems, or that there is no problem at all. The
classification of a call remains problematic until the call is actually
resolved, because one can never be sure that a problem has been correctly
identified until it is resolved.” (Pentland, 1991, p. 95)
This aligns with the problem types of Uncertainty, Ambiguity and Complexity in
Zack’s (2001) framework.
Assigning the call is the process of redirecting the problem to a specialist – a process
which management is attempting to minimise in the call centres being analysed in
this study. Diagnosis creates a more certain and agreed interpretation of the problem.
This may not be easily established, particularly if the knowledge distance between
the caller and agent is significant. Responding should involve an acceptable outcome
for the caller, but may cause the caller to consider that no solution will be found. In
this instance, the agent needs to act to preserve the relationship with the caller. A call
is closed only if the caller agrees that it is.
The objective of this research is to identify strategies and practices embedded in the
left sector of the Q-R Cycle (Timbrell et al., 2005b) which are able to improve the
capability of public-sector agencies. This enabling capability14 for the agency may
then provide competitive advantage if it is inimitable (i.e., a core capability), as
identified by Leonard (1995). Requests for information from a call centre in this
environment are varied and generally not predictable, even though they often relate
to political activity at that time. Factors to be addressed through application of the Q-
R Cycle include the nature of the problem (through a taxonomy provided by Zack
14 Public-sector agencies focus on effectiveness and efficiency principles in the delivery of government services. Hence, the preferred practice would be to assist other public-sector agencies to gain the same capability (hence it should be imitable) in order to enhance the service quality of the whole government further – not just of the agency.
Research Method
43
(2001)), determinants of quality of advice (Brogowicz et al., 1990), and the ability to
match and measure the impact of a particular knowledge management strategy,
particularly knowledge reuse (Markus, 2001; Timbrell et al., 2003).
3.4. Customer Expectations
There is a continuos and growing expectation that decisions and advice from public
servants is informed and rational. The Queensland legislation Freedom of
Information Act 1992 and Judicial Review Act 1991 embeds these principles. The
concept of informed citizens is significant in the current political climate: people
have an expectation that the systems of government are transparent, efficient, and
deliver a public benefit. The consequence of the information society is the increasing
options for choice. An expectation has been created that the public sector will
provide accurate and complete responses to requests for information, advice on
actions relating to government regulation, and simple personal benefit from services
provided by the public purse.
The problem for the public sector is the translation of such expectation into
successful strategy. How is a call-centre staffer, quite junior in status, able to respond
confidently to the range of requests directed from the public at large, which often has
an over-inflated expectation that this individual can provide the answer? A range of
knowledge strategies identified in the literature are able to be applied in a public-
sector call centre to solve this problem. Recruitment of intelligent staff, adoption of
structured approaches to training, use of both formal and informal networks of
“experts,” targeted systems development, involvement in discussions around the
coffee machine, and formal after-action reviews are several recommended
approaches (Choo, 1996; Cross et al., 2005; Davenport & Prusak, 1998; Leonard,
1995; Wiig, 1999). These strategies will be successful provided organisational
creativity, operational effectiveness, and quality of services are all improved through
competent use of knowledge (Todd & Southon, 2001).
Figure 9 illustrates that effective and efficient performance leads to satisfied citizens,
in turn inducing them to trust government (Van de Walle & Bouckaert, 2003). A
subsequent benefit is the establishment of trust between caller and agent, which
Jungermann (1999) sees as crucial in a service environment. This case study
approach will investigate the interrelationships between effectiveness and service
Research Method
44
quality on the basis that client trust is an outcome determined by these former two
criteria.
Figure 9 - Micro-performance Approach to Trust (Van de Walle & Bouckaert, 2003, p. 894)
3.5. Problem Resolution Strategies
In order to improve operational effectiveness, Zack’s (2001) taxonomy simplifies
what appears to be an impossible array of problem types into a manageable,
unambiguous set of knowledge states. The problem types of complexity, uncertainty,
equivocality, and ambiguity matched with know about, know how, know why, and
know with, allow classification of problem types. These can then be analysed
statistically to identify priorities for knowledge management strategy development.
According to Zack (2001), different types of knowledge problems are best processed
by differing knowledge and information systems strategies. Zack’s (2001)
knowledge problems are summarised in Table 2. As is illustrated, for each of his
knowledge problems, Zack (1999) suggests a number of information systems
strategies.
Research Method
45
Information Knowledge
Lack of…
Uncertainty
Insufficient factual information about the goal, situation, or task, and some lack of confidence in the consequent inferences, estimates or predictions required.
Ambiguity
Inadequate knowledge (patterns/concepts) about, no explanation for, or understanding of a goal, situation, or task
Variety /
Diversity
of…
Complexity
Too many situational elements and relationships to coordinate or consider simultaneously
Equivocality
Multiple interpretations of a goal, situation, or task
Table 2 - Summary of Zack’s Four Knowledge Problem Model (Zack 2001, figure 1)
For problems of uncertainty, Zack (2001) suggests:
1) Providing central repositories to enhance the ability to locate codified and
documented information;
2) Providing automated capabilities to analyse large amounts of information;
3) Configuring communication networks in highly flexible ways to respond to
unpredictable information processing needs;
4) Enabling communication regardless of geography or time; and
5) Enabling broadcast at-large requests for information and knowledge, thus
eliminating the need to know precisely where it is located.
Becker’s (2001) strategy to reduce uncertainty involves acquiring increasing
amounts of information until the problem is diminished in terms of its uncertainty.
Becker also suggests that a problem defined as ambiguous will not be resolved by
this strategy; rather, it may make it worse. The information already at hand needs to
be processed in order to reduce an ambiguous problem to a complex one.
For problems of complexity, Zack (2001) suggests:
1) Auxiliary high-capacity memory for managing and analysing complex sets of
information rapidly, i.e. computer-based decision support systems, database
systems, and expert systems;
2) Develop searchable online repositories of explicit knowledge to leverage the
organisation’s experts;
3) Develop the ability to locate experts spontaneously and quickly; and
Research Method
46
4) Facilitate decentralised decision-making by making local information
available globally and global information available locally.
Utilisation of “embrained knowledge,” the cognitive skills and conceptual abilities
used by people to understand complex causations, is a strategy widely relied on that
releases the latent mental potential in problem solvers (Thompson & Walsham,
2004).
For issues of ambiguity and equivocality, Zack (2001) suggests the provision of
communication technologies to best support dialogue between a flexible and
responsive network of experts and associates. The purpose of the dialogue is to apply
the available knowledge resources to transform problems of ambiguity and
equivocality into problems of complexity and uncertainty. In some ways Zack’s
(2001) approach is similar in nature to Hansen et al. (1999), who describe their two
knowledge strategies of codification and personalisation, promoting the use of
information systems in their codification strategy, and communication technologies
in their personalisation strategy. The nature of the problem types (queries) directed at
the call centre affects the processes employed to respond to that problem.
Figure 10 - Generic Taxonomy of IT Helpdesk Calls (Pentland, 1991, p. 102)
Research Method
47
A generic call classification system (Figure 10) was developed by Pentland (1991) in
his study of IT helpdesks. This hierarchical model initially divides between
“technical” and “non-technical” calls.15 Pentland (1991) in turn breaks technical calls
into “problem” and “question.” “Question” is then broken into “general,” “how to,”
and “compatibility.” These question calls mirror the problems of uncertainty
identified by Zack (2001). The “problem” calls are further subdivided into “ordinary
problem,” “bug,” “user error,” and “other vendor’s problems.” The difference
between a problem and a question is defined in terms of the callers’ expectations.
Pentland (1991) suggests that a caller does not have a problem until they have
attempted a solution and it did not work. A framed question (such as “How do I…?”)
makes the problem easy to identify, but this may become an issue if the helpdesk
response is a suggested solution already tried without success.
The problems will then rely on the abilities of the helpdesk staff to identify the
causalities, whether that of software or the user expectations of that software.
Complex problems, as defined by Zack (2001), would cover a range of calls
identified by Pentland (1991), but when callers have an expectation that the software
will perform certain functions (when in fact it will not or needs intervention from
another expert), it becomes a problem of ambiguity. Equivocality arises when the
system designers misinterpret functional requirements.
3.6. The Knowledge Management Strategy
The case study method also captures the strategies used by agents to resolve queries
to the call centres, allowing identification of relationships between roles, problem
types, and resolution methods. The works of Markus (2001) and Timbrell et al.
(2003) provide a guide to development of an appropriate knowledge management
strategy. The quality of advice given relies substantially on the relationship
established between the caller and receiver. For call-centre staff to be confident, they
need to be knowledgeable and responsible for their decisions. The inability to answer
a query immediately and accurately is not regarded as poor advice; rather, poor
advice comes with an attempt to cover for such a knowledge gap. Dixon (2000)
15 Pentland (1991) does not expand on the non-technical calls, since only technical calls were the focus of his study, which referred to a strategy of immediate referral back to a switchboard. His research indicates a fairly clear distinction between technical and non-technical calls, but admits some calls were of a hybrid nature.
Research Method
48
provides advice that a knowledge reuse strategy encompasses a “complete solution”.
Simply relying on technologies such as repositories and fast information technology
will not deliver improvement in performance. Also, design of incentives to increase
contribution to, and use of, information repositories is to be considered, since this
will provide support for the codification and storage of reusable information
identified as being essential to a successful knowledge reuse strategy.
The reuse situations identified by Markus (2001) and Timbrell et al. (2003) include
activities by data miners who do not interact directly with callers in that role, and so
will not be considered further in this study. Of the other roles identified by these
authors, shared-work producers are to be encouraged to make information available
to the whole team, rather than storing it for their own reuse (which is far simpler to
do). Shared-work practitioners also need to become producers, as they add value
though their work. Expert-seeking novices, as they extract relevant knowledge from
experts, are to make that knowledge available through community of practice
discussion, contribution to a repository, or by some other means. The goal is to
manage effectively both the codified knowledge of the organisation, and the expert
knowledge that resides within call-centre agents and every other employee of the
organisation. In each reuse situation, the information system that supports the
knowledge management initiative must be capable of accessing expertise as well as
the experts.
The role of the knowledge manager is of critical importance for the development of
organisational strategies. It is not sufficient to rely on call-centre staff to develop,
contribute, and disseminate relevant knowledge. Three key functions of such a role
include:
a) Ensuring valuable information is recorded in a knowledge repository;
b) Reviewing such information to ensure its currency and finally providing
assistance in locating; and
c) Linking historical information relating to areas of interest (Markus, 2001).
This case study investigates the existence of knowledge strategies in two Queensland
Government call centres and the roles individuals play in implementing them.
Research Method
49
3.7. Quality and the Call-Centre Response
Quality criteria, as identified by Brogowicz et al. (1990), are coded in the data
collection phase of the project. The three areas of focus are:
1) Training of staff, which is critical to ensure the determinants of quality
identified by Brogowicz et al. (1990) are addressed. These determinants of
tangibles (those visible signs of quality such as work space and contemporary
technology), reliability, responsiveness, assurance, and empathy are personal
qualities that require support from both technical and functional expectations.
It is anticipated that interviews with mangers and staff will provide data that
identifies training strategies.
2) The capability to incorporate knowledge captured in information systems into
general problem-solving and learning heuristics is referred to as the technical
determinant of quality. The technical determinants within a call centre are
substantially impacted by the individual staffer’s knowledge of the context of
the problem presented by the querist. Both individual and organisational
capabilities are able to be improved through experience, training, and access
to codified information. Physical resources identified as technical
determinants of quality include the information technology infrastructure,
communications infrastructure, and published materials.
3) Functional service quality, the expressive performance of a service, reflects
the way solutions derived as a result of technical capability are transferred
through personal attributes of attitude to service, accessibility, and public
relations (Gronroos, 1984, p. 39). Selective recruitment and training, and an
organisational culture overtly attuned to client service are precursors to
meeting the functional service quality gap.
Quality of advice is a core capability of successful service organisations. Leonard
(1995) proposes a model for the creation of capability (as shown in Figure 11)
through shared problem solving, internal implementing and integrating, future
innovation, and importing knowledge from external sources. Improvement in the
quality of advice provided by call-centre agents may be continuous and incremental
if many of the fundamentals of the model are in place. The requirement is for agents
to transform from the individualistic shared-work producers (who create knowledge
Research Method
50
primarily for their own later reuse) to team-based shared-work practitioners (who
acquire new knowledge from external sources in context). This suggests that
judicious prioritisation of knowledge reuse strategies would lead to improved
capability, and hence performance improvement in problem-solving exercises. As
knowledge management strategies are implemented to address each problem type,
successful outcomes are able to be embedded in practice, and unsuccessful ones
discarded or re-evaluated. The case study methodology provides an opportunity to
document the degree to which problem solving in the call centres is a shared or
personal activity.
Figure 11 - Capability-Creating Activities: Problem Solving (Leonard, 1995, p. 60)
3.8. Pilot work
Initial rounds of consultations were held with senior staff of the host agencies to
provide input into the further refinement of the framework. One call centre was
identified as having a complete recording of all customer conversations over the past
eight years. An initial sample of these calls was used to establish a reliable and valid
coding protocol (based on Zack’s (2001) problem types, Markus’s (2001) reuse
strategies and Brogowicz et al.’s (1990) quality assessments). Team leaders from the
call centres took part in workshops to review the protocols and processes, and to
develop an instrument to assist in the documentation of the information gathered.
PRESENT
INTERNAL
FUTURE
Shared Problem Solving
Improving Knowledge
Experimenting
Implementing and Integrating
EXTERNAL
Core
Capabilities
Research Method
51
3.9. Implementation
3.9.1. Participants
Three cycles of the Structured Case Methodology (Carroll et al., 1998) make up the
high-level processes in the study and focus on different staff groupings, as shown in
Table 3.
Stage 1
Establishing
the
Context
Stage 2
Embedding
the
Process
Stage 3
Performing
the
Analysis
Table 3 - Three Cycles of the Structured Case Method as Applied to This Study
Senior executives responsible for the management of the call centres have been
consulted in Stage 1 of the research project and gave approval to proceed. Through
informal and unstructured interviews, they provided insights into the business drivers
Strategic Requirements
Business Drivers Quality Criteria Problem Categorisation
Identify expected individual and organisational
behaviours
Unstructured interviews
Why does the call-centre
exist?
Link knowledge strategies to
expected outcomes
Operational Protocols
Operational Constraints Personal Impacts Problem Resolution
Develop data gathering tool for
call analysis
Structured interviews and
meetings
Gain support of call-centre
staff
Test and agree on coding protocol &
Pilot
Implementation
Literature (Zack, Brogowicz, Markus etc.) Insights Heuristics
Secondary code in NVivo to identify relationships and chain of causality
Collect and code the data
Identify relevant sources of
research data & information
Reflect on the research findings
Research Question
Resolution
Research Method
52
behind the call centres and the knowledge strategies they employed in relation to
knowledge reuse and its contribution to agency outcomes (being improved advice).
The executives contributed further by providing feedback on the initial conceptual
framework proposed to analyse call-centre performance.
Call-centre supervisors were interviewed to develop the data-capture process, and
also to assist in the validation of coding. Particularly in the pilot phase, these team
leaders were used extensively to ensure the coding rules reflected the nature of the
work being performed by the agents. They also assisted in the identification of key
phrases which provide “hints” for that coding process. Another role for the team
leaders has been to ensure that their staff members were comfortable with the
research project. To this end, several staff briefings were organised for each call
centre where the research project was presented, and queries relating to both
concerns and benefits were addressed. As well as providing call-centre operators
with general background of the project (a development exercise in its own right), this
ensured researcher’s access to staff during the data collection phases in a relaxed and
non-threatening environment.
The majority of the data has been taken from archival recordings taken by one of the
call centres. Live calls by the centre staff have also been monitored to assist in the
protocol development on both problem types and quality of advice. This is consistent
with a current practice where callers are advised that monitoring may take place,
primarily for the purpose of training. Callers have the option to veto such a practice.
3.9.2. Data Gathering
A rich source of data has been made available through allowing interviews with
managers, access to procedural and factual documents, recorded phone calls, and
archived materials covering the lives of the call centres. Over this period, several
strategies have been applied in order to improve call-centre performance. Anecdotal
evidence exists to indicate the success (or otherwise) of these strategies. Advice-and–
complaint-type calls have been sampled around these events and coded to capture:
1) Problem type (Zack, 2001);
2) Quality of advice (Brogowicz et al., 1990) and
3) Identification (if possible) of the resolution strategy used.
Research Method
53
In particular, the use of codified information (e.g., database, intuitive, training,
referred to tier 2) was recorded as it is a strong example of knowledge reuse. By
sampling around known interventions, significant changes in agent behaviours would
indicate a consequence of that event.
Pentland and Rueter (1994) have performed a detailed analysis of calls and responses
to helpdesks for software support of complex computerised systems in order to
develop a grammar to describe organisational routines. Their analysis defines the
processes specialists used to resolve problems, as shown in Figure 12, such as
transferring the problem to other areas, escalating it to engineering staff, making
reference to vendors, or by transferring back to the customer for further investigation
and information collection. Pentland and Rueter (1994) have devised and validated a
coding scheme for this process, as per Table 4. Their analysis focuses on non-routine
problem-solving work with a high variation in actual problems, but with a similarity
in call types (e.g., a familiar problem on a different platform). The problems
encountered have been classified as having low analysability owing to the user
uncertainty as to what caused the problem, whether a feasible solution exists,
inability to construct an explanation for the problem, and not always having
sufficient information to isolate the symptoms of the problem.
Code Explanation
O Open the call.
W Work on the call.
C Close the call.
I Declare problem inactive, usually because it can't be reproduced, so no further work can be done.
D Defer the problem, can't solve now.
T Tentative fix completed.
R Resolve problem (but leave call open for confirmation by customer).
F Fix given to customer.
E Explain reason for closing.
US Transfer responsibility to user.
PS Transfer responsibility to product support.
DV Transfer responsibility to development.
TR Transfer call to another functional area.
Table 4 -Coding Scheme for Software Support Process (Pentland & Rueter, 1994, p. 494)
Research Method
54
Figure 12 - Kinds of Responses to Calls to IT Helpdesk (Pentland, 1991, p. 115)
Using the Pentland and Rueter (1994) analysis as a model, the codes in Tables 5, 6,
and 7 have been proposed to incorporate the activities being investigated in this
study. Sufficient records are available to obtain statistical significance, but the coding
protocol will limit the amount of detail that will be recovered about any particular
interaction. This will not impact on the fit of the research data with the proposed
model being investigated, but will afford the necessary degree of confidentiality to
callers and call-centre agents.
Research Method
55
Zack’s (2001) Problem Types
Code Type Description
A Ambiguity “Inability to interpret or make sense” (Zack, 2001, p. 21). This,“cannot be coded precisely into mutually exhaustive and exclusive categories” (Weick, 1995, p. 92)
C Complexity “A large number of parts that act in a non-simple way” (Zack, 2001, p. 18)
E Equivocality “Multiple meanings” (Zack, 2001, p. 21)
U Uncertainty “Lack of enough information to chose an exhaustive and well defined set of possible states, preferred outcomes and actions” (Zack, 2001, p. 19)
X Not Coded Not coded – does not fit the Zack (2001) model
Table 5 - Proposed Coding Scheme for Problem Type
Resolution Strategies
Code Type Description
PD Took Personal Details
Personal details taken, such as name, phone, address.
CR Cyclical reframing
Cyclical reframing of the problem through testing partial solutions (Weick, 1995).
DP Decompose Decompose into smaller manageable problems and resolve (Zack, 2001).
MD Meaning through discussion
Meaning through discussion (trial and error, sounding out) (Weick, 1995).
RI Requires research
Requires time to research further information – will return call.
SC Social construction
Social construction and intervention through scanning and discovery. The environment is not just given, but rather can be “enacted” through the actions of agents, and produced by knowledge rather than the opposite (Weick, 1995).
TC Terminate Call Terminate Call due to unforseen factors such as caller hang-up, etc.
TU Transfer to user Transfer responsibility to user (more information or some action by caller needed).
TR Transfer to Functional Area
Transfer call to another functional area.
TS Transfer to supervisor
Transfer responsibility to call centre supervisor.
CI Use codified information
Access codified information (such as databases, fact sheets etc.).
Table 6 - Proposed Coding Scheme for Resolution Strategy
Research Method
56
Quality Characteristics
Code Type Description
EA Ease of access Ease of access to the call-centre service (Brogowicz at al., 1990).
EC Empathy with customer Empathy with customer (Brogowicz et al., 1990).
CS Satisfactory closure Satisfactory closure of the call.
CT Technically competent Technically competent and knowledgeable (Brogowicz et al., 1990).
TQ Time waiting in queue Time waiting in queue < 30 seconds (Brogowicz et al., 1990).
RC Technical resource capability
Technical resource capability (eg access to Databases, IVR. etc.) (Brogowicz et al., 1990).
CB Not available - Call Back
Attempt to transfer call to a 2nd tier - not available so gave call back details.
NEA Difficult to access Negative - Access to the call centre service is difficult (Brogowicz et al., 1990).
NEC No Empathy with customer
Negative – Displays no empathy with customer (Brogowicz et al., 1990).
NQC No quality criteria used No quality criteria used.
NSC Unsatisfactory closure Negative - Call closed without an agreed resolution.
NTC Not Technically competent
Negative - Unable to address problem resolution in a logical manner (Brogowicz et al., 1990).
NTQ Excess Time waiting in queue
Negative - Time waiting in queue > 30 seconds (Brogowicz et al., 1990).
NTR No Technical resource capability
Negative - Lack of technical resource capability (eg no access to Databases, IVR etc.) (Brogowicz et al., 1990).
Table 7 - Proposed Coding Scheme for Quality Characteristics
3.9.3. Validity of Results and Data Analysis
This research has been limited to three cycles. The first cycle has been used to
validate the coding protocols and to establish a mechanism to ensure such a process
is learnable and repeatable. Thirty recorded calls were initially hand-coded by the
researcher. The results were then presented to the call-centre team leaders and
reviewed item-by-item. This process has also been used to identify “triggers” to be
used in the real-time coding of calls. Once an agreement was reached with team
leaders, a research assistant again coded the test set using the queues and other
information relating to the research topic. This exercise has been performed to ensure
that the researcher and assistant were able to code consistently. It has also established
that no more than five strategies have been applied for both resolution and quality
actions. The capture tool to be used in the major cycles has thus required only five
fields for storage of these codes. An ancillary benefit to this process has been the
Research Method
57
demonstration of the removal of identifying data necessary to meet the ethical and
confidentiality requirements of the agencies and the University. The test for construct
validity has been satisfied by the application of the restricted coding set (Yin, 1994).
Data collected in the second cycle is from recorded calls in one of the call centres.
Each call has been double coded for consistency. The coding process has been
assisted through the use of a database which provided timing, sequencing, and
coding automation (see the data definition in Table 8). Drop-down menus and
enforced metadata capture have meant that the coded data has been able to be later
scrutinised to confirm consistency and repetition of the process. Double-coding has
revealed few anomalies, which have been reviewed and changes agreed. These
results have again been discussed with call-centre team leaders, who have again
confirmed the coding to be consistent.
Field Name Data Type Description
time_Encoded Date/Time Default value = system time
call_Centre Number Either CC-A or CC-B (toggle)
call_Source Number To record whether a recorded or real time - recorded calls capture the media details of the recording (eg CD#99)
call_ID Text Unique sequential system identifier
call_Date Date/Time Default value = system time; able to be overwritten to reflect recorded call time
call_Topic Text General description - allows possible analysis of motivation for making the call
zack_Code Text Any of the 4 codes and option not to satisfy the Zack (2001) criteria (drop-down)
qualCode1 Text One of the 14 codes of the Quality criteria (drop-down)
qualCode2 Text Another of the 14 codes of the Quality criteria (drop-down) - code(s) eliminated if used in any previous quality field
qualCode3 Text As for qualCode2
qualCode4 Text As for qualCode2
qualCode5 Text As for qualCode2
resolveCode1 Text One of the 11 codes of the resolution criteria (drop-down)
resolveCode2 Text Another of the 11 codes of the resolution criteria (drop-down) - code(s) eliminated if used in previous quality field
resolveCode3 Text As for resolveCode2
resolveCode4 Text As for resolveCode2
resolveCode5 Text As for resolveCode2
call_Comment Memo Available to enter comments which may be pertinent to the call
Table 8 - Data definition for data capture tool
A final cycle has been used to capture real-time data from the second call centre.
This extra data has been gathered to compare and contrast data from different call
centres. The information collected has been analysed to establish its consistency
within the framework. The purpose of this action has been to establish the external
validity and reliability through replication (Yin, 1994).
Research Method
58
3.10. Summary
Chapter Three has been used to detail both the reasons for selecting a case study
methodology and the protocols used to enact the methodology. Detractors of case
study methodology point to a weakness associated with repeatability (Gable, 1994).
Since the framework proposed is a performance improvement model, it is not able to
provide an absolute measure for outcomes. It does, however, provide a model which
informs practice, and offers indicators for changes to outcomes. This aspect of the
research is repeatable in any related advice-giving environment. Yin (1994)
recommends a range of tactics to address validity and reliability criteria for quality
research design. The protocols developed for this study follow his recommendations.
The convergence of multiple sources of data, adoption of a structured approach, and
the incorporation of two organisations in the case study have ensured Yin’s (1994)
quality criteria have been addressed.
Work, Organisation, and Processes
59
Chapter 4
Call-Centre Work, Organisation, and Processes
4.1. Introduction
This chapter depicts life in two Queensland public-sector call centres. The systematic
observation of activity in two call centres identifies a variety of behaviours which are
characterised by different abilities to utilise knowledge strategies contributing to the
establishment advice-giving practices in the public sector. In effect, the approach
emphasises the commonalities and differences in systems and the behaviour of
agents. At the core is the focus on service quality by individual agents, and the fact
that knowledge is abundant, but that performance is limited by the ability to use it.
Management is, however, concerned with the allocation of scarce resources and their
alignment with localised business drivers.
4.2. A Broad View of Call Centres
Milner (2000) analyses the growth of public-sector call centres, and concludes that
the public’s familiarity with and access to telephones has allowed these centres to
flourish. She draws on the example of the Brisbane City Council, which has
articulated two key criteria for change needed to implement a call centre. These
criteria are “advantage to customer” and “demonstrate the achievement of value for
money”. Technology has provided the enabling role in the provision of information
access via the Internet and corporate Intranets, mobile and remote service delivery,
and in creating a sense of empowerment for users. The leveraged service quality
through consistency and ease of access to services supports the criteria for success
and provides a blueprint for the operation of public-sector call centres. The value-for-
money criteria is reflected in the knowledge self-sufficiency of call centres, i.e., the
ratio of queries resolved to the total queries within a call centre ( Timbrell et al.,
2003). The lower cost of problem resolution within the call centre, relative to tier-2
or specialist referral areas, means this measurable and meaningful ratio is indicative
of call-centre performance. It also supports the customer-advantage principle, since a
highly self-sufficient call centre has the capability to handle the majority of calls
without the need to redirect them to a third party.
Work, Organisation, and Processes
60
Customer advantage also exists for callers in their ability to connect with the call
centre conveniently. Toll-free numbers (where the call centre accepts the charge) are
widely used, but a normal call (where the caller bears the cost of the call themselves)
still exists for some centres. Having connected with the call centre, the caller is
welcomed by either an agent or an interactive voice response (IVR) system. An IVR
system may take the caller through a layered menu of options, resulting in a
connection with a specialist agent, a general-enquiries agent, or an automated system
that requires no human intervention. Contemporary examples of this practice include
flight arrival information, recorded weather reports, and stock market prices. In
some cases, the caller can exit from the IVR system and connect directly to, or queue
for, an agent. The following provides an insight into the two call centres, designated
as CC-A and CC-B.16
4.3. Evolution of the Study: CC-A
The first call centre, CC-A, has been the initial focus of the study. This study
commenced in early 2002. CC-A was (at the time) the focus of efficiency drives, as
well as responding to changing legislation that contributed to significant growth in
the number of calls to the agency. Management was primarily concerned with
throughput, and allowed team leaders considerable freedom in the day to day
operations of the call-centre. CC-A’s operation is characterised by technology-driven
improvement, with performance reports reflecting the volume and time spent
responding to calls. However, a structural change in the organisation caused the
appointment of a new Executive Manager whose responsibility covered certain
aspects of the call centre.
External performance measures were instigated in an attempt to gauge the service
quality of a major component of the agency, and in particular the services delivered
via CC-A.
16 The identities of the organisations, staff, and callers in this study are disguised through coding. Although people familiar with the operations of the departments may identify the particular call centres being studied, no individual will be exposed. There is no intent to make value judgements on performance; rather, observation of activities and responses is documented in order to develop the frameworks necessary to provide the repeatability of findings. Nothing is lost in this anonymity since the assurance of confidentiality has allowed participants to behave in “normal” ways, and to be honest in face-to-face interviews.
Work, Organisation, and Processes
61
Another organisational restructure transferred responsibility for the call centre to a
senior officer who initiated a number of changes to improve its performance. From a
knowledge perspective, each initiative has had a positive impact:
a) The introduction of “teams of teams” encouraged the sharing of knowledge
across the agency (including regional offices) via constant communication,
thereby ensuring all agents were kept current. This is a strategy referred to as
“knowledge streaming” (Timbrell et al., 2005b). These teams were initially
facilitated externally to share their vision and develop action plans in a two-
day workshop. Six-monthly face-to-face meetings were also planned, with a
commitment to improving intra-office communications.
b) A mentoring (“buddy”) system was also introduced where junior officers
were linked with senior members. Each mentor had only one buddy. These
buddies were nominated by management to ensure appropriate match of
experience, skills, and personalities. The knowledge issues addressed by
these strategies is both the sharing of knowledge, in much the same way as
apprentices learn their trades through socialisation and internalisation
(Nonaka & Takeuchi, 1995), and also to address the potential loss of
knowledge to the organisation through retirement of experienced officers.
This is of concern as the age demographic shows 24 percent of the agency is
50 years or older (i.e., potential retirees), while 40 percent are over 45 years
of age. See Figure 13 for a graphic representation of the age demographic
existent in the agencies.
c) The final initiative was the employment of a Training Officer, to maintain the
static database of information used, design, and deliver induction and
training, and to document processes and standards of operation formally
within the call centre. This strategy targeted the management of explicit
knowledge as identified by Nonaka and Takeuchi (1995) in the
externalisation and combination modes of knowledge conversion. In an
interview in December 2002 (Interview 1b - Annotated notes, December 23,
2002), the team leader suggested this regime was aimed at multi-skilling the
call-centre workforce, and that it has had a positive impact on both the
motivation of individual agents and the performance of the call-centre
operations.
Work, Organisation, and Processes
62
Agency Age Distribution cf State Employed Population
0%
5%
10%
15%
20%
15-
19
20-
24
25-
29
30-
34
35-
39
40-
44
45-
49
50-
54
55-
59
60-
64
65-
74
75 +
Age Range
%
State %
Agency %
Figure 13 - Age Demographics of staff employed by the agency hosting CC-A
In an unstructured interview with the Executive Manager of CC-A in early 2003
(Interview 1a - Annotated notes, January 29, 2003), plans to obtain hard data were
discussed. Issues of concern were identified by the manager as a result of unsolicited
feedback from clients. These concerns primarily related to the consistency and
accuracy of advice provided by the call-centre and other front-line service staff as
they interpreted and applied the range of regulations being administered. The
absence of reliable data to support these concerns led to the instigation of a project to
obtain hard data capable of guiding decision making for improvement strategies.
Employing a technique developed by a sister agency in another Australian State, an
external analyst was engaged to develop, deliver, and analyse a “mystery shopper”
exercise. This had the intention of testing call-centre staff on matters of accuracy
and consistency of advice being offered. Approximately 300 telephone calls and 50
email queries were performed. Call-centre staff were assessed on greeting, call
times, helpfulness, and accuracy, as compared to model answers. Client exit
interviews and mystery shopper activities produced sets of data which were analysed
and reported on by the external analyst. In order to ensure the cooperation of the call-
centre staff, however, an embargo was applied on the findings by the agency. The
scenarios presented were made available but due to the focus on restricted function-
Work, Organisation, and Processes
63
specific issues sets17 (primarily recall with little problem solving), they failed to
provide the breadth required for a valid18 and repeatable research exercise. The focus
of the research then changed to CC-B due to its reputation for being a best-practice
model within the Queensland public sector.
4.4. Evolution of the Study: CC-B
In the initial meeting with the manager and the team leader of the second call centre,
CC-B, the strategic role for the call-centre was discussed. The management of
people, processes, and performance was focussed on delivery of a service based on
quality principles, rather than efficiency. The team leader made the point that in this
organisation “quality is more important than time.” The call centre has fewer than
ten permanent staff members who have been recruited on the basis of attitude rather
than experience. This recruitment is followed by a training period of up to six
months, which allows the agent to develop a deep knowledge of the organisation.
CC-B has a well-documented set of processes covering both inbound and outbound
call functions. The provision of “tele-research” is the major outbound service offered
by CC-B on a fee-for-service basis. Call-centre agents work closely with clients to
develop survey questionnaires to ensure that the business group’s goals in
undertaking the survey are met. Consultancy regarding the timing, accuracy, and
type of questions is provided before data collection begins. Management has
emphasised the importance of the business group liaison with the interviewers, in
order to provide an in-depth understanding of the questions and intent of the survey
and the profile of the client base. A database has been created for the clients’
responses. If a hard copy is required by the business group, records can be updated
from paper-based questionnaires. Before the survey instrument is universally applied,
a pilot is performed on a sample of five to ten clients, and a post-pilot report is
provided. This report includes client response, length of time to complete the survey,
clients’ understanding of questions, and accuracy of information obtained. On
17 The scenarios developed were mainly problems relating to uncertainty in Zack’s (2001) framework and tested either accurate agent recall or the use of the dedicated database for scripted answers only. 18 An interesting phenomenon reported was that some of the experienced call-centre agents were able to identify the mystery shopper calls through the inconsistent responses to in-depth questioning by the agent. The knowledge distance was apparently greater than normally expected of callers to the call centre.
Work, Organisation, and Processes
64
completion, a final report and synopsis is provided to the business group in electronic
format. This report contains the final figures, percentages, graphs, and any other
information requested by the business group. The synopsis contains interviewer
observations and comments, and adds anecdotal information regarding the conduct of
the survey for the information of the business groups. Although market research has
not been a conventional undertaking for public sector agencies, it is clear from the
activities of CC-B that an agency has the capability to deliver services both to
external and internal clients. Such a role will always be supplementary to the general
information/advice functions of government.
Functions performed
Preparation Negotiation & Performance
Acceptance
Client
Call Centre
Resp
on
sib
le
Are
a
Business Area
Table 9 - General information processes in CC-B
General information processes in CC-B are documented in process charts having a
similar format to Appendix 1, and reflect a matrix structure assigning function
preparation, negotiation, performance, and acceptance to client, call centre, or
business area. Calls to the call centre are managed in a way that is consistent with the
Query-Response Cycle (Figure 14), where they are either resolved via the knowledge
resources of the agent (primarily the “Preparation” column of the documented
process in Table 9), or consigned to tier 2 (primarily the “Negotiation and
Performance” column of the documented process in Table 9). Other documented
processes, with variations on the general information procedure, include the straight
transfer of calls to nominated agency staff or services, the provision of specialist
advice/information, mail-out processes (caller-instigated), non-English speaking
caller processes, complaints, and an emergency response hotline process. The
performance levels relating to these processes are documented in a Service Level
Agreement (SLA).
Work, Organisation, and Processes
65
Figure 14 -Query-Response Cycle (Timbrell et al., 2005b, p. 548)
The SLA provides a framework for the development of a constructive and effective
relationship between CC-B and its Business Units. The framework has evolved from
an examination of the business needs of the call centre, and describes the levels of
service to be provided to meet those needs “and how they will be monitored,
evaluated, measured, and managed” (Service Level Agreement for Provision of
Services, 2004).19 Call-centre business process service levels are encompassed in the
general responsibilities of the call centre, business units, and specifically-nominated
roles within the agency. This reflects the symbiotic relationship between the call
centre and the wider agency, and specifies the type of personal dialogue to ensure the
maximum self-sufficiency of each. This apparent paradox balances the knowledge
needs of the call centre to reach its goal of resolution of 80 percent of calls without
referral, and the recognition that business units do not wish to be interrupted by
19 It is not intended that this agreement should have legal consequence; rather, that it serves the mutual benefit of both parties by providing a clear understanding of agreed operating arrangements and performance criteria. It is expected that it will evolve over time.
Knowledge
Problem
Query
Response
Query Reformulation
Search Strategy
Knowledge
Sources
Response Set
Pay off
Assessment
Knowledge Distance Adjustment
CONSIGN RE-QUERY
RESOLVE
ABANDON
Service Quality
Work, Organisation, and Processes
66
“trivial” issues, such as the dissemination of information able to be codified and
reused. Whilst it is the responsibility of the call-centre operator to communicate
accurate, complete, and up-to-date information to callers, it is the overall
responsibility of the business units to ensure the information within the relevant
databases is accurate, complete, and up-to-date. This is achieved through annual
reviews undertaken by the business units, and continual updates from clients. Figure
15 indicates how callers access the call centres and are managed in the initial phase
of the Q-R Cycle.
The operational aspects of both call centres studied, as documented below, were
presented to the 15th Australasian Conference on Information Systems in December
2004, and published in the conference proceedings (Schefe & Timbrell, 2004).
Research
and
Return call
CC-A Only
CC-A Only
The Operational Context
Caller
Integrated Voice Response
General agent
Specialist agent
Consign to other
Supervisory or
Specialist staff
Measure
Service
Quality}CC-B Only
Legend:
communication flow
is limited to the call-centres indicated.
communication flow
in both call-centres.
Figure 15 - Operational Context of the Call-Centres Studied
4.5. Operational aspects of CC-A
CC-A is a call centre in a Queensland government department that builds consumer
and business confidence in fair marketplace outcomes. CC-A is an integral
component of a Customer Service Centre (CSC), which also includes counter staff
both in the capital’s central business district and at regional centres. Calls to the
widely publicised 1300-number (which allows calls from anywhere in the state to be
Work, Organisation, and Processes
67
charged at local call rates) are routed to the nearest available reception point, which
will be either a regional office or the call centre located in the central business
district. The IVR system then routes the call based on whether it is a general query
or complaint (which will be more complex) or a transaction relating to the regulation
of business. Transactional functions are performed in real time, while client issues
may involve further research, be consigned to second-tier experts, or investigation’s
officers where legislative breaches have occurred.
Prior to a reorganisation in October 2002, the CSC comprised approximately 60
Brisbane-based staff. Regional staff generally processed over-the-counter enquiries
or took calls directly. The CSC consisted of teams with clearly defined roles
separating transaction processing (lower-level staff), complaints via the phones,
written complaints, and other business services, such as data entry and bank
reconciliation of payments. The situation was characterised by very little
communication between teams, limited written procedures, and no ability to load
balance (the process of distributing effort to areas currently experiencing the lowest
resource demand) during peak times. There was an imbalance in supervision loads
which generated some resentment. Training was minimal: it was on-the-job and ad-
hoc.
After October 2002, the organisation corrected the (both real and perceived)
inequities in work through the formation of work clusters and multi-skilling of all
staff in the CSC. The supporting information systems in the regional offices were
integrated with the Brisbane systems, and staff performing customer-service roles in
the regions were included in the CSC division. The PABX system commenced
routing callers to the nearest available service staff, both to decrease communications
costs and also to provide local recognition for the services offered by the
Department. Strategies included rotation of staff, production of procedure manuals,
targeted training (including regular meetings), and other technology upgrades.
The CSC consists of front counter staff members who predominantly serve face-to-
face customers, a call centre taking IVR-directed calls to areas of expertise, and the
business process area which data-enters payments and manages other finance-related
issues. This model is consistent with that of other service industries (Graumann,
Arnold, & Beltjes, 2003). The CSC established a matrix of teams across functional
Work, Organisation, and Processes
68
areas and regional offices. These teams are encouraged to share knowledge and
discuss decisions across their lines of responsibility.
The call-centre management software used in the central business district is
Symposium. This system displays call-queue information and generates a database
for performance statistics. It produces reports and offers real-time displays of current
status. The regional offices use a less sophisticated application, which does not route
calls based on skills and experience of the particular call-centre agent. It does,
however, collect statistics and give real-time displays of calls waiting and time-in-
queue. Management sees technology as the best option to drive efficiency while still
providing advice and management registers to support Queensland businesses and
the general public. All CSC staff has access to the Departmental Intranet, which is
based on Microsoft Team Services. The Intranet includes staff contact lists, self–
publishing capabilities, threaded discussions, and event-driven capabilities. Specific
work groups have also created sub-Webs for their individualised use.
Transaction systems include registers for the management of business details and
history of client complaints. These systems are accessible to all CSC staff. A static
database has been developed in-house to hold hierarchical lists of information,
categorised by functional area so as to allow fast access to the information.
Information is entered into this database by staff with a degree of expertise in a
particular area, and only contains statements of fact; no interpretive information is
provided. This information forms the basis for a scripted response, although it is not
kept in a dialogue format.
Reported performance measures include:
a) Response rates where the call-centre software displays and stores a range of
statistics on number of calls by type, time-in–queue, and abandonment rates.
Approximately half a million calls are processed per year.
b) Quality issues are addressed by “mystery shopping” exercises that have been
performed to test a range of criteria including the consistency and accuracy of
advice and service focus.
c) Staff satisfaction, although no measures were available. It was noted that as at
December 2002, the longest serving employee in one team had four months’
experience.
Work, Organisation, and Processes
69
An attempt to embed continuous performance improvement through training has led
to the appointment of a dedicated Training Officer. This role includes usual formal
training sessions, but also includes responsibility for maintenance of the information
database. Other strategies for performance improvement include regular team
meetings and externally facilitated workshops, to build trust and shared vision
through identification of improvement strategies by the teams. Action plans
developed and reviewed on a six-monthly basis include the identification of
knowledge gaps and the processes to address these. New staff operate in a “double
jacking” configuration (a dual mode where two agents are able to concurrently
participate in the query-response process) with an experienced operator for two
weeks prior to going solo, while each operator has a buddy nominated by
management to match experienced staff with the new staff for mentoring and
support. The CC-B call centre has a formal process for monitoring call quality. Team
leaders decide on the number of calls to be sampled, and assess the agent’s
performance against seven communication skills20 and sets of technical skills based
on the functional area21 of concern to the caller. Technical requirements are tailored
to the particular functional service area, and include:
a) Search the relevant business database correctly;
b) Perform the business transaction correctly;
c) Give the correct advice;
d) Explain forms and related matter accurately;
e) Use the information database to give consistent advice; and
f) Correct the referral process when required.
The results of these evaluations are used only to provide feedback to the agent, and
are not incorporated into a whole-of-unit performance regime. CC-A calls are not
20 Self-orientation, greeting, courtesy, listening, questioning, call termination, and dealing with difficult customers are each assessed against documented criteria which do not change. 21 Eight specific functional areas are possible. Only those areas of concern to the caller can be assessed. Some agents generally develop more expertise in one area than another through previous experience and the impact of the IVR system. Each call has similar assessment criteria, such as accurate advice, use of codified knowledge, and correctly identifying the database, and search methods employed.
Work, Organisation, and Processes
70
recorded, but double jacking22 is accepted as a standard practice both for training as
well as for evaluation and research.
4.6. Operational aspects of CC-B
CC-B operates within a government department whose principal business is to
maximise the economic potential for agricultural industries on a sustainable basis.
Assistance is provided to both domestic and commercial clients, with some services
being provided on a fee-for-service basis. Agents in CC-B address queries on a wide
range of scientifically-based topics. Any caller can contact CC-B for the cost of a
local call, either to a direct number or via a 1300-prefixed number.
Eight information officers (not scientists), a team leader, and the centre manager staff
the call centre. Laid out in cubicles, the call-centre operations room is quite
spacious, creating a relaxed environment for the workers. Calls are managed by a
call board and system (VU-ACD/100 for Windows), which continuously display a
range of status information on calling queues. All calls are recorded.
At the time of its establishment, CC-B’s management implemented a policy not to
use IVR technology to route calls. The policy that people, and not information
systems, will handle call diversions is designed to heighten callers’ perception of the
functional service quality (Brogowicz et al., 1990). The most significant investment
in technology is the infrastructure used to manage call-centre operations and
statistics. Continuous displays of calls in queue and average time-in queue monitor
centre performance in terms of call management. A range of databases and Intranet-
based information sets are used to support the agents. Scientists are encouraged to
provide information to the call centre to allow them to address “hot” issues.
Operators record a small amount of metadata about the conversation on the call
logging system (Figure 16), but minimal information about the resolution of the
problems raised is logged to assist future related queries.
22 Early in the research project the researcher was invited to spend as much time as necessary listening in to calls. The agents in CC-A were proud of their capabilities and seemed keen that time should be spent with them. A concern for the researcher was that such listening may influence agent response behaviours, but no evidence of changed behaviour has been identified over the three years of the study. The actual calls used as base data for the research were coded towards the end of the research period.
Work, Organisation, and Processes
71
Figure 16 - Call logging system – CC-B
The Queensland Government department which hosts CC-B makes a conscious
effort to maintain its Internet site with the most up-to-date information. However,
due to issues such as the lack of access to the World Wide Web and degree of
comfort with the technologies by the Department’s principal constituents (the
agricultural sector), querists prefer to access the call centre for advice and
information. Tier-2 support is provided by scientists, who are often located at
research stations outside the capital’s central business district.
Although responses to callers are not scripted, agents are instructed not to give
advice, but rather to provide only information sourced from their standard
information systems. They are supported by a range of fact sheets and databases
populated by the scientists. These resources continually evolve through interaction
between the scientists and the call-centre agents. Agents have responsibility for
meeting with scientists and business groups to identify future events, report on issues
being addressed by the call centre, satisfaction levels with the call centre, and to
request additional or updated fact sheets from scientists and other informants within
the organisation.
The scientists were initially reluctant to support the call centre. They felt that this
communication medium with their customer base was inappropriate; that the agents
would not have the expertise to answer the queries and could mislead their customers
Work, Organisation, and Processes
72
(Interview 2a - Annotated notes, May 27, 2003). When the CC-B call centre started
operations in 1995, however, scientists found it led to a decrease in “trivial” queries
normally attended to by them, allowing them to concentrate on research and other
higher value work. Today, generic calls managed by the call-centre account for up to
60 percent of calls, which would have previously been referred to scientists.
Owing to the technical and scientific nature of calls, substantial effort is placed on
recruitment and training of the call-centre staff. The key selection criteria for call-
centre agents contain the requirement to display superior functional service quality
attributes by using techniques such as active listening, patience, and “when and how
to speak” (Thompson, Warhurst & Callaghan, 2001). Superior technical knowledge
of the Department’s business minimises the time to reach acceptable levels of
competence (typically between six and eight months). Maintenance of the call-
centre staffs’ capabilities is achieved via weekly meetings at the close of operations
that allow discussion of important issues able to be identified from records of the
calls, future events, and other issues of concern.
The focus of the call centre is on client satisfaction and grade of service (in which 80
percent of calls are answered within 20 seconds). Normal days involve each operator
taking approximately 80 calls. In peak times caused by staff shortages or increased
activity from emergent events, however, CC-B can cope with up to 200 calls per
operator per day. All calls are recorded, with a library of calls dating back several
years.23These recordings are used primarily for confirmation of later queries and are
occasionally used to develop a profile of call-centre agent performance. The major
quality monitoring process, however, involves real-time evaluation of agents by the
team leader who “double jacks”24 to an operator in order to gauge the quality of
23 The recording of calls often raises the industrial issue of “spying” on staff for the purpose of management of diminished performance. It also has implications for privacy of caller information. The former issue is regarded by call-centre agents as a safety net against vexatious complaints – an indication of the maturity and confidence staff have in their capability and the trust in their peers. An often-cited example is one of a caller who complained to the relevant Minister that the call-centre agent was rude and unhelpful. A check of the recorded calls showed quite the opposite, where it was the caller who was rude and aggressive.
The second issue (privacy) is covered by advising callers that the call is being recorded, and supporting this with a strict policy of access to this recorded set. The researcher feels privileged to have been given access to this data. 24 Double jacking as explained in section 4.5 is the activity whereby a second agent can listen in to the call without being able to influence the outcome. For the purpose of the study, the researcher was only able to listen i.e. the microphone capability was turned off.
Work, Organisation, and Processes
73
responses via a sampling process. The process is documented and reflects
assessments over eight (8) calls on a monthly cycle. Evaluative criteria include the
accurate capture of data about the caller in the call-centre information system, a
range of nine categories in the area of communications skills (assessing greeting,
manner, determination of client’s needs, empathy, courtesy, professionalism,
verbalisation of actions, active listening, and confirmation strategies, which are not
too dissimilar to the Brogowicz et al. (1990) model), and five categories in call
management (call control, protocols, duration, resolution, and follow-up as the
effectiveness measures used). The process for the cycle ends in a face-to-face
feedback session where the previous month’s action plans are reviewed and new
ones are agreed for the next cycle. At this meeting, a graph of “true calls per hour”
(an efficiency measure which filters out calls incorrectly routed to the agent) versus
quality assessment is discussed.
This graph (Figure 17) shows the position of all agents (anonymously, except for the
individual being reviewed), with the objective being to identify best practitioners,
and to provide a visual personal goal of moving to the top right of the represented
area. These assessments form part of the broader (whole-of-centre) performance
report produced for management.
PQ MATRIX../../00 - ../../00
Agent 2
00.5
11.5
22.5
33.5
44.5
55.5
66.5
77.5
88.5
99.510
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
PRODUCTIVITY - True Calls Per Hour
QUA
LIT
Y
Figure 17 – Agent Performance Graph (Interview 2b - Full transcript, June 16, 2004)
Performance measures which are collected and used in reporting include response
rates (grade of service and occupancy statistics most used in reporting and decision
making), quality issues (from assessment of individual agent performance), and
client satisfaction. The vision for the future of CC-B is to have 80 percent of all calls
Work, Organisation, and Processes
74
completed by its staff (i.e., 80 percent knowledge self-sufficiency). In other words,
only 20 percent of calls will be consigned to second-tier support groups.
4.7. Summary
CC-A has evolved due to business pressure to cope with increasing demand for
timely and accurate advice. Efficiency continues to be the major business driver in
CC-A. CC-B has developed from a planned strategy to provide accurate and timely
information as part of an integrated approach to client management. Based on a
personalised service and supported by training and coaching, the approach is to
protect clients from the technology driving the call centre. Table 10 summarises the
strategies employed in each of the call centres.
Strategy CC-A CC-B
Use of Integrated Voice Response / Call Routing
Yes No
Call Rate Up to 200 calls per person per day
80 calls per person per day
Information recorded on calls
Minimal – for reporting and return contact purposes
Minimal – for reporting and return contact purposes
Updating of information repository
Not involved Meetings with scientists / 2nd Tier operatives
Internet site Limited information available All information available
Consignment Policy Research often performed by agent prior to consignment
Consignment if query cannot be resolved with standard information
Level of remuneration for operators
Government clerk level 2 & 3 Government clerk level 4
Training 2 weeks 6-8 months
Table 10 - Summary of the Strategies Employed by Both Call Centres
The two call centres studied, although both within the Queensland public sector,
reflect their own unique business issues and underpinning service principles. Each
can be easily mapped onto the Q-R Cycle (Table 11), and provides clear performance
expectations of the call-centre agents. Customer access criteria contribute
significantly to decisions and strategies employed. The IVR system of CC-A has
been designed to direct callers to clusters of expertise within the call centre, and
hence reduces the need to refer on to another agent. Intelligent routing technology
has also been employed to allow regional callers to connect with local agents who
Work, Organisation, and Processes
75
may have a better understanding of local issues. CC-B has employed a standard-of-
service measure weighted strongly in favour of the reduction of waiting time-in-
queue. A premium has been allocated to this, owing to the need to allow callers to
develop a relationship with the agent, particularly when complex issues are being
addressed. Value-for-money at CC-B is reflected in the proportion of calls able to be
resolved without referral to (relatively more expensive) experts who are more
usefully occupied in other value-adding roles. CC-A had no benchmark available, but
as demand for their services has increased, the marginal costs of servicing this
increase have been able to be absorbed into the accepted funding model. A
framework to analyse these observations is developed further in the next chapter.
Phase of Q-R Cycle CC-A Strategy CC-B Strategy
Accept Query Caller directed to functional experts via IVR. Average time-in-queue 73 seconds (target of 60 seconds) – see Appendix 2.
All calls are accepted by generalist agents. Average time-in-queue < 20 seconds.
Query Reformulation Training provided in introduction, listening and paraphrasing.
Training provided in introduction, listening, and paraphrasing.
Search Strategy/ Knowledge Sources
Dedicated database with script support is the primary knowledge source.
Range of knowledge sources including the Internet, fact sheets, and specific databases.
Response Set Presented factually but embellished with context provided by the caller to create the link between the regulatory basis for the advice and the situation described by the caller – may be taken off-line for further research and more formal response.
Responses are highly tuned to client situation and expectations. Referral to experts is done with handover briefing to maintain client care.
Pay-off Assessment/ Knowledge Distance Adjustment
Unless extended by further caller queries, the call is terminated once advice has been given.
Significant proportion of time spent prior to termination in ensuring the outcome is understood and the caller is satisfied.
Table 11 - CC-A, CC-B, and the Q-R Cycle
A Framework for Analysis
76
Chapter 5
A Framework for Analysis
5.1. Introduction
The case studies introduced in the previous chapter have identified the work,
organisation, and processes used by the study’s call centres. Actual problem
resolution methods are not clear from these observations, and a further level of
analysis has been performed using Zack’s (2001) Four Knowledge Problem Model.
This two-dimensional model is simple to apply, but powerful in its ability to
categorise behaviours of call-centre agents. The analysis of these behaviours and
comparison with strategies recommended by Zack (2001) will allow both validation
of the model and the identification capability of gaps and redundant processes.
5.2. Zack’s (2001) Four Problem Model
The problem of “uncertainty … represents a lack of information, or factual
’knowledge about’ current and future states, preferences and appropriate actions”
(Zack, 2001, p. 20). CC-B recognises the high proportion of problem types which
fall into this category, and has developed processes to maintain up-to-date electronic
and paper-based information sources. These are designed and populated by experts
(scientists), so that callers with knowledge problems of uncertainty are generally able
to be satisfied quickly. Development of intellectual capabilities within CC-B to
predict, infer, estimate, and learn (as part of the training strategy), assists in the
management of uncertainty. CC-A has a similar strategy for maintenance of
electronic information systems (also populated by experts), but does not replicate this
information on the Internet. The IVR system also contributes to the efficiency of
CC-A, since it streams callers directly to functional experts who, through their
experience in dealing with specific business problems, are able to recall factual
details quite independently of the electronic systems.
Zack (2001, table 1) suggests the response to complexity is either to “increase a
firm’s capacity to process it or to reduce the level of complexity.” Both CC-A and
CC-B commence dialogue with the caller by following an initial scoping-questioning
technique. CC-A takes longer (due to the need to respond to the IVR prompts), but
directs callers with complex problems to the more capable staff within the team. The
A Framework for Analysis
77
superior, yet more complex, IVR programming provides this capability. The system
generates monthly call-time reports for CC-A staff. Their objective is to minimise
call time. The impact of this is that the time available to decompose complex
problems during the call is reduced. Once the CC-A agent decides time will limit
resolution, the problem is consigned either to tier-2 specialists, or it may be set aside
for further research by the agent at a later time. CC-B operators do not specialise in
particular functional areas, and are not considered to be fully trained until they have
undertaken several months of training and mentoring. The operator’s new-found
knowledge increases the call-centre’s capacity to deal with complex problems.
Higher levels of remuneration also increase the likelihood that more talented and
experienced operators will be both attracted to the call centre and will stay on longer.
In addition, since call time is not a critical success factor in CC-B, extra resources
can be committed in real time to decomposing complex problems. The option exists
for CC-B agents to consign to second-tier experts, but the goal to reduce referrals
encourages resolution within the call centre.
Provision of rich interactive conversation is the key capability in management of
ambiguity, according to Zack (2001). CC-B has addressed this in its strategy of
providing ample time for call-centre operators to discuss the problem with the caller,
and to attempt to reframe the problem into a scenario that is more meaningful. The
agent is able to repeat several cycles of interpretation and explanation until either the
problem becomes resolvable (complex or uncertain), or is consigned to an
appropriate scientist whose education and experience provide the skills to solve such
problems. CC-A agents do not have the time to perform this type of dialogue, and
the consignment policy of postponing it to resolution by off-line research does not
allow for reframing, interpretation, or explanation.
The multiple meanings and interpretations which are characteristic of problems of
equivocality require negotiation to converge on one meaning (Zack, 2001). With its
policy of generous time allocation for agent dialogue with callers, combined with
capable staff, CC-B is well placed to resolve such problems. The knowledge strategy
employed is not one of access to information, but rather one of process which
provides the capability to cycle interactive discussion, negotiation, and interpretation
until a resolution is achieved. The IVR system used in CC-A may become a liability
with problems of equivocality, since the confusion of meaning may even commence
A Framework for Analysis
78
with the initial options presented to the caller. This may end in an incorrect referral
to an agent who may have preconceived ideas on the context of the call. These
incorrectly directed calls are generally addressed quickly either through
abandonment or further referral to a more appropriate tier-2 agent.
CC-B has the goal of achieving 80 percent call-centre self-sufficiency. The reporting
and analysis of consigned calls serves as a trigger to identify strategies to reduce
them. Past strategies have included requests for fact sheets from scientists, briefings
by specialists, and incorporation of contemporary topics into formal training
sessions. CC-B also has sufficient redundancy in its resources to maintain a grade of
service which is interpreted as zero abandoned calls. In times of high load due to
staff illness or a significant event, calls can be lost. These lost calls are not analysed
due to the infrequent occurrence of such events. CC-A has a complex call-routing
system, which attempts to locate available local resources, and progressively skips to
the next available agent. This is also supported by an IVR system to stream callers
vertically into functional pools based on their business need. In this process, many
calls are abandoned. Reporting in CC-A includes (by business function) total
number of calls, number of referrals beyond the call centre, proportion of calls settled
with advice, and the percentage of all calls sampled for quality monitoring. The
analysis of this data is primarily used to identify the division of business functions
and the proportion of effort directed to this. It is also used to pre-empt future client
needs and educational or promotional requirements.
Within both call centres, the majority of problems fall within the
uncertainty/complexity problem type. This means that knowledge strategies based on
codified knowledge sources allow agents to converge quickly to a viable response
set. The effort put into the development of fact sheets and databases serves this
requirement well, since uncertainty is removed through access to current knowledge,
while complexity is reduced through online searching and supported by access to
experts via referral. CC-A, however, has a significant proportion of callers whose
problems fall within the ambiguity problem type. These generally emerge through
the IVR system as a complaint and are managed by the more experienced team
members. Such calls tend to take much longer to resolve and involve either referral
to experts or are managed off-line due to the requirement for further research.
A Framework for Analysis
79
Another issue of this dispersed agent model has been identified by Brooke (2002).
Workers not involved in the physical call centre (e.g., located in regional offices), but
who receive IVR-directed calls may not share the knowledge and meanings generally
agreed by the core group in the head-office environment. The continuous payoff
assessment by callers for these problem types generally demands much more
resource allocation, since the problems often have a personal/corporate financial
consequence for the caller. Staff working in this field do receive higher pay, but are
also subject to much more stress, both due to the intellectual challenges of the
problem, and also to the awareness of the need to reduce time per call (an efficiency
measure used both at an individual level as well as at a team level).
Analysis of CC-B calls where the researcher tracked the IVR path of a caller,
suggests that the IVR system effectively directs callers to the relevant area of
expertise. The message that greets the caller provides options for the major business
functions, which predominantly reflect uncertain problem types. The IVR system
routes the call to agents with expertise to resolve the problem, either through recall or
access to codified knowledge. Other options route the predominantly complex (due,
mainly, to legislative interpretation) and ambiguous problems to agents who employ
a range of resolution strategies in order to achieve successful problem resolution.
A F
ram
ework
for
Anal
ysi
s
Ta
ble
12
- S
um
ma
ry o
f K
now
led
ge
Str
ate
gie
s U
sed
in
th
e C
all
Cen
tres
by
Pro
ble
m T
yp
e
Zack
’s (
2001)
Kn
ow
led
ge
Pro
ble
m T
yp
es
Str
ate
gy
Call
C
entr
e U
nce
rtain
ty
Com
ple
xit
y
Am
big
uit
y
Eq
uiv
oca
lity
CC
-B
No
IV
R
No
IV
R
No
IV
R
No
IV
R
Use
of
Inte
gra
ted
V
oic
e R
esp
onse
/
Cal
l R
outi
ng
CC
-A
Po
siti
ve
imp
act
- in
itia
l cl
assi
fica
tio
n b
y c
alle
r p
rovid
es q
uic
k a
cces
s to
te
chnic
ally
cap
able
ag
ents
Pro
bab
le n
egat
ive
imp
act
due
to t
he
call
er
nee
din
g t
o s
elec
t fr
om
p
red
eter
min
ed o
pti
ons.
Pro
bab
le n
egat
ive
imp
act
due
to t
he
call
er n
eed
ing t
o
sele
ct f
rom
p
red
eter
min
ed o
pti
ons.
Pro
bab
le n
egat
ive
imp
act
due
to t
he
call
er n
eed
ing t
o
sele
ct f
rom
p
red
eter
min
ed o
pti
ons.
CC
-B
Po
siti
ve
imp
act
- al
low
s ti
me
for
clar
ific
atio
n b
y
acq
uir
ing a
dd
itio
nal
fa
cts
Po
siti
ve
imp
act
- al
low
s ti
me
for
pro
ble
m
dec
om
po
siti
on a
nd
re
buil
din
g
Po
siti
ve
imp
act
- al
low
s ti
me
to r
efra
me
a p
rob
lem
si
tuat
ion i
nto
so
met
hin
g
mo
re m
eanin
gfu
l
Cal
l R
ate
CC
-A
Po
siti
ve
imp
act
- as
sist
ed b
y
IVR
, ca
ller
s ar
e d
irec
ted
to
sit
uat
ional
kno
wle
dge
exp
erts
Neg
ativ
e im
pac
t -
pre
ssure
to
lim
it c
all
tim
e im
pac
ts o
n a
gen
t’s
abil
ity t
o i
den
tify
the
inte
rrel
ated
var
iab
les,
so
luti
ons
and
met
ho
ds
Neu
tral
sin
ce t
he
stra
teg
y t
o
carr
y o
ut
furt
her
re
sear
ch o
ff l
ine
is n
ot
fact
ore
d i
nto
cal
l ra
te
stat
isti
cs
Neg
ativ
e im
pac
t -
the
con
flic
ts
cause
d b
y m
ult
iple
in
terp
reta
tio
ns
req
uir
es
sig
nif
ican
t ti
me
to
conver
ge
on a
shar
ed
pro
ble
m s
tate
men
t
80
A F
ram
ework
for
Anal
ysi
s
Zack
’s (
2001)
Kn
ow
led
ge
Pro
ble
m T
yp
es
Str
ate
gy
Call
C
entr
e U
nce
rtain
ty
Com
ple
xit
y
Am
big
uit
y
Eq
uiv
oca
lity
CC
-B
Po
siti
ve
imp
act
- sm
all
imp
act
since
all
ow
s an
alysi
s o
f tr
end
s to
sup
po
rt o
ther
kno
wle
dge
stra
tegie
s
Nil
Po
siti
ve
imp
act
- ca
ller
det
ails
al
low
s ag
ents
to
re
vis
it/f
orw
ard
cl
arif
yin
g i
nfo
rmat
ion
Nil
Info
rmat
ion
reco
rded
on
call
s
CC
-A
Po
siti
ve
imp
act
- sm
all
imp
act
since
all
ow
s an
alysi
s o
f tr
end
s to
sup
po
rt o
ther
kno
wle
dge
stra
tegie
s
Nil
Po
siti
ve
imp
act
- gen
eral
ly
add
ress
ed o
ff l
ine
wit
h
form
al n
ota
tio
n o
f re
sear
ch i
ssues
, b
ut
no
t li
nked
to
the
call
-cen
tre
IS.
Nil
CC
-B
Po
siti
ve
imp
act
- ro
bust
and
cu
rren
t re
po
sito
ry e
asil
y
acce
ssed
by a
gen
ts
Po
siti
ve
imp
act
- sm
all
imp
act
since
so
me
fam
ilia
r ro
uti
nes
are
d
ocu
men
ted
Nil
N
il
Up
dat
ing o
f in
form
atio
n
rep
osi
tory
CC
-A
Po
siti
ve
imp
act
- ro
bust
and
cu
rren
t re
po
sito
ry e
asil
y
acce
ssed
by a
gen
ts
Po
siti
ve
imp
act
- b
oth
o
per
atin
g p
roce
dure
s an
d h
euri
stic
s ar
e co
ded
in t
he
call
-ce
ntr
e IS
.
Po
siti
ve
imp
act
- P
arti
cula
rly
hel
pfu
l in
reg
ula
tory
ar
eas
Nil
CC
-B
Po
siti
ve
imp
act
- ex
ten
sive
and
curr
ent
site
eas
ily
acce
ssed
by b
oth
cal
lers
an
d a
gen
ts
Nil
N
il
Nil
Inte
rnet
sit
e
CC
-A
Neu
tral
– n
ot
a m
ajo
r so
urc
e o
f use
ful
fact
ual
in
form
atio
n
Nil
N
il
Nil
81
A F
ram
ework
for
Anal
ysi
s
Zack
’s (
2001)
Kn
ow
led
ge
Pro
ble
m T
yp
es
Str
ate
gy
Call
C
entr
e U
nce
rtain
ty
Com
ple
xit
y
Am
big
uit
y
Eq
uiv
oca
lity
CC
-B
Po
siti
ve
imp
act
- m
ajo
rity
re
solv
ed i
n f
irst
cal
l w
itho
ut
refe
rral
; in
form
atio
n s
up
pli
ed b
y
exp
erts
do
cum
ente
d i
n
fact
shee
ts
Po
siti
ve
imp
act
- re
ferr
al
net
wo
rk t
o e
xp
erts
b
ased
on f
unct
ional
ar
ea
Po
siti
ve
imp
act
- re
ferr
al
net
wo
rk t
o e
xp
erts
bas
ed
on f
unct
ional
are
a
Po
siti
ve
imp
act
- re
ferr
al
net
wo
rk t
o e
xp
erts
bas
ed
on f
unct
ional
are
a
Co
nsi
gn
men
t P
oli
cy
CC
-A
Po
siti
ve
imp
act
- m
ajo
rity
re
solv
ed i
n f
irst
cal
l w
itho
ut
refe
rral
; ac
cess
to
exp
erts
thro
ugh
kno
wle
dge
of
org
anis
atio
nal
un
its
Po
siti
ve
imp
act
- re
ferr
als
to
mo
re e
xp
erie
nce
d s
taff
w
ithin
the
call
-cen
tre
Po
siti
ve
imp
act
- sp
ecif
ic u
nit
s es
tab
lish
ed t
o d
eal
wit
h
refe
rred
cli
ents
(b
ased
o
n c
onsi
gn
men
t ru
les)
Po
siti
ve
imp
act
- sp
ecif
ic u
nit
s es
tab
lish
ed t
o d
eal
wit
h
refe
rred
cli
ents
(b
ased
o
n c
onsi
gn
men
t ru
les)
CC
-B
Po
siti
ve
imp
act
- re
lati
vel
y
hig
h t
o a
ttra
ct a
nd
ret
ain
inte
llec
tual
res
ourc
es
Po
siti
ve
imp
act
- re
lati
vel
y
hig
h t
o a
ttra
ct a
nd
re
tain
in
tell
ectu
al
reso
urc
es
Nil
N
il
Lev
el o
f re
mu
ner
atio
n
for
op
erat
ors
CC
-A
Neg
ativ
e im
pac
t -
rela
tivel
y
low
; hig
h t
urn
over
m
eans
loss
of
inte
llec
tual
res
ourc
es
Neg
ativ
e im
pac
t -
rela
tivel
y
low
; hig
h t
urn
over
m
eans
loss
of
inte
llec
tual
res
ourc
es
Neg
ativ
e im
pac
t -
exp
ecta
tio
n
of
com
ple
tio
n o
f o
ff l
ine
rese
arch
, al
tho
ug
h
pro
vid
ing v
arie
ty,
also
cr
eate
s st
ress
Nil
82
A F
ram
ework
for
Anal
ysi
s
Zack
’s (
2001)
Kn
ow
led
ge
Pro
ble
m T
yp
es
Str
ate
gy
Call
C
entr
e U
nce
rtain
ty
Com
ple
xit
y
Am
big
uit
y
Eq
uiv
oca
lity
CC
-B
Po
siti
ve
imp
act
- fo
cuse
s o
n
kno
wle
dge
acq
uis
itio
n,
info
rmat
ion s
earc
h
met
ho
ds
and
ser
vic
e q
ual
ity c
rite
ria
Nil
N
il
Nil
Tra
inin
g &
m
eeti
ngs
CC
-A
Po
siti
ve
imp
act
- fo
cuse
s o
n
kno
wle
dge
acq
uis
itio
n,
info
rmat
ion s
earc
h
met
ho
ds
and
ser
vic
e q
ual
ity c
rite
ria
Po
siti
ve
imp
act
- ag
ents
en
coura
ged
to
uti
lise
m
etho
ds
to r
eso
lve
com
ple
x p
rob
lem
s
Nil
N
il
83
A Framework for Analysis
84
5.3. Summary
Neither of the call centres studied employs a knowledge strategy based on any overt
theoretical framework. The strategies employed have evolved due to the current
business drivers and continue to develop based on the historically bureaucratic
organisational structure designed to deal with complexity. However, the process of
participation in this study indicates that each organisation is prepared to develop a
capability to handle a broader set of knowledge problems. Table 12 provides a
summary of the impact of current knowledge strategies,25 and identifies the gaps that
need to be addressed if the model proposed by Zack (2001) applies. This summary
highlights the general practice of resolution of complex/uncertain problem types by
the call-centre agent, and a referral system for those aligning with Zack’s (2001)
ambiguous/equivocal types.
Zack’s (2001) four-problem framework has been used to analyse the policies and
actual practices of the call centres, CC-A and CC-B. Cross et al. (2005) suggest that
call centres use routine response social networks, since these are best suited to
situations where problems and solutions are fairly predictable and resolved with a
degree of collaboration. This is mirrored in the strategies espoused by CC-B and
CC-A, which have a robust tool set for resolving uncertain and complex problems.
Methods of attack for ambiguous and equivocal problems are not well supported
within the call centres, but consignment policies do exist to address them. In Chapter
Six, a closer investigation of call archetypes and their responses reveals patterns of
response strategies used by the agents.
25 A nil entry in any particular cell may mean either that the strategy is not appropriate for that problem type, or that the particular call centre does not employ that strategy.
Patterns from Observations, Interviews, and Responses
85
Chapter 6
Patterns from Observations, Interviews, and Responses
Qualitative research relies on patterns and commonalities extracted from rich case
data; however, the sensitive interpretations from this complex data are difficult to
represent in a numeric way. The seemingly infinite and unrelated data elements
originally encountered in the study of the two call centres have been used to develop
a new understanding of the organisational behaviours exhibited in this micro-
community of the Queensland public sector. A forms-driven database has been used
to encode calls, and a qualitative data analysis tool called NVivo has provided the
capability to link recurring concepts in interviews, observations, literature, and
individual call conversations. The outputs from these tools allowed the management
and synthesis of ideas into well-defined patterns that have emerged from the
relatively unstructured primary data sources. The relationships between components
of this research activity are shown in Figure 18.
Management Views
Literature Search
CC1 Management
CC2 Management
PS Call Centres
Quality Principles
KM principles
Data Collection
Ancilliary documents
Method
PhenomenographyYin
Analysis
Nvivo
Research Structure
Coded Calls by Zack Type
CC1Leader_Interv iew
CC2Leader1_Interview CC2Leader2_Interview
Figure 18 - Research components
Patterns from Observations, Interviews, and Responses
86
6.1. Executive Management Interviews
Data has been collected from a wide variety of sources. Early interviews with
Executive Managers were used to:
a) Identify the business drivers for the call centre;
b) Gauge the Executive Managers’ degree of familiarity with the principles
which underpin knowledge management; and
c) Obtain a commitment from the Executive Managers to support the research
program (by providing access to the staff, operations, and statistics).
Both the desired and the actual practices and procedures in operation were
documented during the cycles of meetings with call-centre managers, team leaders,
and staff that followed the interviews. During this period, calls were monitored in
real time,26 and retrieved from archival systems to develop a complete picture of the
interrelationships among management expectation, policy compliance, procedural
activity, and individual responses. The theories of knowledge reuse and their
application to repository design and intervention postulated by Markus (2001)
extended the work of Dixon (2000), who investigated criteria for successful
knowledge transfer. The roles of shared-work producers, shared-work practitioners,
expert-seeking novices, and secondary data miners, and their contributions and
interactions with repositories formed the basis of the structured interviews with
managers and team leaders.
6.2. Management and Team Leader Interviews
The role of repositories is of interest to contemporary knowledge managers
(Davenport & Prusak, 1998; Markus, 2001). Of interest to the later analysis of calls
is the resolution strategy employed by agents. In many cases this involves identifying
an appropriate information source, which leads to resolution of the particular query.
The postulate that knowledge reuse (the combined search strategy, information
source, and application in context) improves call-centre performance is central to this
study. In interviews with management and team leaders, at least 64 references were
26 Staff were quite comfortable with having the double jack arrangement, since it is used for training, performance monitoring, and the promotion of the work done by call-centre agents.
Patterns from Observations, Interviews, and Responses
87
made to knowledge repositories. These have been coded as External Knowledge,
Structured Internal Knowledge, or Informal Internal Knowledge repository types
(Davenport & Prusak, 1998, p. 146), as shown in Table 13.
Table 13 - Frequency of Repository References in Interviews
Call-centre agents and supervisors have identified the knowledge sources that
contributed to these repository types as including:
External: “Library; Web sites of other Government Departments; general Internet
sites; legislation; and referral indexes”;
Informal Internal: “palm notes; book where they’ve got common contacts and stuff;
email; daily news; just a matter of writing it down; we keep half folders of
information; tagged and marked so they can readily refer to things; some
will store things electronically”; and
Structured Internal: “Knowledge Base; Infotech; Brands database; telephone listing;
Call Centre information system; fact-sheets; flowchart procedures on how
to answer the calls; printed documents; Inform database system; Intranet;
training manuals; handbooks” (Interview 1a - Annotated notes, January 29,
2003; Interview 1b - Annotated notes, December 23, 2002; Interview 1c -
Full transcript, June 9, 2004; Interview 1d - Full transcript, June 11, 2004;
Interview 2a - Annotated notes, May 27, 2003; Interview 2b - Full
transcript, June 16, 2004).
The roles undertaken within the call centres have different responsibilities for the
development, access, and maintenance of these repositories. Team-leader roles
across both agencies have aligned extremely well with Markus’s (2001) shared-work
Patterns from Observations, Interviews, and Responses
88
producers, who generate knowledge for their own later reuse. Their title reflects their
primary purpose of the development of an effective operational environment. Team
performance has been mentioned regularly in the interviews, with clear operational
performance measures being defined, collected, and reported. The following extracts
from interviews with team leaders indicate their expectation that agents be aware of
performance targets:
“And they can see what they do for the day, for the week. They can
see what their peers do for the week. And in a way, it's sort of
y'know, encourages them to do that little bit better too.”
"If, see Symposium tells us, for individuals, it tells us how many calls
they’ve taken, what their talk time is. I actually had one girl in my
team that was taking like 50 calls a day, but her talk time was four
minutes."
"And our target is three minutes and under. So I actually sat with her
and I also sat at my desk and listened in on her calls. She knew I was
doing it. Listened in on her calls and then at the end of it, I went
through it with her and I said “Well y'know.” She had a problem, her
customer service was too much."
"And most staff actually know their stats by the end of the day
themselves."
"We refer to it [performance statistics] regularly and report on it. I
report on it weekly and monthly."
"And I give feedback weekly to the team and that’s usually on the
number of calls received per day, per week. The average wait-time."
This is consistent with findings in literature which indicate a predominance of
measurement against benchmarks by technology solutions (Anton & Gustin, 2000).
Easily measured and reported, strategies have been applied to ensure individual
performance is within the acceptable boundaries for the agency. The mentoring role,
although not referred to in the interviews, was a significant component of call-centre
staff’s daily work noted in real-time observation. The team leaders have been
generally regarded as a source of general and contextual knowledge that relates to the
contemporary issues of the time. This reuse situation comes from the historical
Patterns from Observations, Interviews, and Responses
89
experiences of the team leaders, who have referred to the development of the call-
centre practices and procedures and reasons behind such actions.
The following passages reflect the characteristics of shared-work producers
described by Markus (2001, p. 65) that, “Users need to know what is done, how it
was done, why it was done and how it can be done better”.
An understanding of the single view of Queensland Government illustrates why
things are done in a particular way:
"And it's not only knowledge outsiders who ask what happened within
[the host agency], we have to have a broad knowledge of what’s
happening in the other Government Departments as well."
Also the intuition exists to maintain historical information for personal use, but
shared with other agents including both what and why a decision was made:
"And that’s what we do. Cause I mean a couple of years ago we had
something, …. I kept the number because..."
Practices are established to identify how things are done and where
improvement opportunities exist:
"Team leaders perform coaching mentoring role and are the
information expert. They also produce performance statistics and are
able to report by subject by day."
How things were done in the past is passed on by the experienced staff to later
recruits:
"Some of us might know more information, cause we’ve been here
longer. So yes, we’re going to have more information, but you don’t
keep that information to yourself."
"We’ve got people who build up expertise over time. Some of the
people have information that goes back a fair way, more so than
someone who’s just joined us. And they’re often relied on as the
knowledge experts and the information is gleaned from them."
Although these examples support the Markus (2001) model of shared-work
producers, some characteristics of shared-work practitioners are also displayed. This
Patterns from Observations, Interviews, and Responses
90
is to be expected, since shared-work practitioners produce knowledge for sharing
with the team and leaders drop into and out of the agent role from time to time. They
have a strong affinity with the team, and hence do not see their role as being outside
that team. Call-centre agents generally fall into the shared-work practitioner
category, and “acquire new knowledge that others have generated,” and “get advice
about how to handle a … situation that is new to the team” (Markus, 2001, p. 64).
The only overt strategy used to support this predominant role in both call centres has
been the regular staff meeting/training activity. However, this has not significantly
contributed to the production of knowledge for the use of other agents. Since these
small call centres have been designed in a way that allows personal interaction
between agents, having access to staff with tacit knowledge is an efficient process
(when compared to codification strategies) (Hansen et al., 1999). This strategy has
simply evolved, and has been taken for granted. CC-B managers have referred,
however, to a redundancy strategy27 in their scheduling, which arguably is a
necessary precondition for knowledge transfer (Davenport & Prusak, 1998).
Emphasis has been placed on team harmony, with the consequence that agents noted
the mutual benefits associated with knowledge sharing. Markus (2001) refers to the
producer, intermediary, and consumer roles in the knowledge reuse process and notes
that each role “can be performed by the same individual or group, different
individual or group, or some combination” (2001, p. 61). The contribution of both
individuals and groups within the call-centre team are valued on the assumption that
it provides benefit to call-centre performance, as seen in the comments below:
“It's the culture of the call centre. The culture that’s been fostered here is just
like a family.”
“If you don’t have it, you know for a certainty that somebody in the call
centre will.”
“Oh yeah, I had this fellow was talking about this and y'know and I found this
out. And then they just, a lot of it is word of mouth between themselves.”
27 Redundancy implies a strategy that allows staff time-out from the routine of work to communicate with their peers, allowing knowledge transfer to occur due to a shared/common view of that knowledge. The value of conversation around the water fountain is an example.
Patterns from Observations, Interviews, and Responses
91
“I always not only tell them the answer, I show them where in the legislation
they can find that answer. And next time they don’t have to come back or
next time they can show somebody elsewhere to find it as well.”
“… it's still shared initially anyway. So if [agent named] finds a really good
site or something, she’ll share it with the team and then it's up to the team
what they do with it.”
“… [An agent] will just stand up and say, ‘Hey, can anybody, Does
everybody know this, that and this?’”
“Or if we get something come through at the counter that people aren’t aware
of then we send it by email to all staff.”
“Staff talk all the time amongst themselves and luckily, we’re very lucky that
we have a lot of staff that really get on well together.”
“So we have a meeting with them once a fortnight and any issues that are
raised by Team Leaders from their teams at team meetings.”
“It’s just a matter of writing it down and if it's something new they think
needs recording or discussing with colleagues, they’ll do that. And perhaps
do that at the end of the day.”
“A lot of people often go to her because those people who store things
electronically may not know where it is, but [agent named] can put her finger
on it with the hard copy.”
These examples illustrate the impact of social networks. Cross et al. (2005) provide a
framework for the identification and nurturing of informal networks in a productive
fashion. Such a knowledge strategy which aligns technology, processes, and human
capability, already fundamentally, yet informally, exists in the call centres studied.
Management intervention that encourages these collaborative networks to flourish
would improve call-centre performance in areas of self-sufficiency.
The goal of 80 percent self-sufficiency in CC-B and the strategy of taking tier-2
work off-line to research stretch the shared-work practitioner role to one of expertise-
seeking novice for call-centre agents, particularly with the goal of minimising the
need to access the more expensive tier-2 experts. The majority of actors in the call
centre defined as expertise-seeking novices are, however, the callers who are seeking
Patterns from Observations, Interviews, and Responses
92
to develop new knowledge about unfamiliar problems, or are unable to contextualise
correctly the issues they are investigating. Calls made to each of the call centres have
been monitored to understand both the processes and interactions of the agents, as
well as the nature of the incoming queries.
Calls examined early in the process28 have been used to test the coding methodology
and to train a research assistant in its application. Archived calls for the final analysis
were randomly selected on the basis of 25 calls for each of the six-month periods
from 1 January 2000 to 30 December 2004, giving a total of 200 calls. They were
retrieved by call-centre staff from recordings and reconstructed to enable them to be
copied to compact disc.
The ability to take them off site for analysis has had several advantages, including:
a) No interruption to the operations of the call centre;
b) Ability to double-code the calls to ensure consistency of approach; and
c) The easy revisitation of individual calls for further detailed analysis.
The major disadvantage encountered in this approach has been the error rate in the
reconstruction process. This has occurred as a result of the recording system used to
archive calls using a time-slicing algorithm, and hence has recorded calls as a block.
Calls that were incomplete at the end of that time-slice have in some cases been
difficult to re-link, and hence have had to be discarded from the sample.
Of the calls selected, 143 have been incorporated into the final analysis for this
research. Rejected calls generally fall into categories of incomplete29 or incorrectly
routed, often from within the host agency. The latter category of calls falls outside
the scope of this research. Based on strategies developed by Di Gregorio (2000), the
output from the database of calls was automatically coded using NVivo, a tool with
search and coding functionality based on predetermined criteria.30 Such coding has
allowed information collection relating to the problem type, resolution, and quality
themes to be explored. This is primarily due to the structured nature of the output
28 These calls did not form part of the final analysis. 29 Due to a quirk in the call-centre recording software, some calls became fragmented in the process of copying them to CD ROM for off-line analysis. 30 Future research could utilise these processes to encode much larger information sets giving a higher degree of confidence and a capability to perform more rigorous statistical validation.
Patterns from Observations, Interviews, and Responses
93
from the original data-collection tool. Each line refers to a particular call which
already has had the predefined codes for problem type, resolution strategy, and
quality. The algorithm used to code problem types automatically has involved
searching for a particular code and allocating the associated node to the whole line
which due to the coding process equates to a paragraph for the purpose of NVivo
analysis.
For response and quality codes, only the particular code was associated with the
relevant node, since the unique codes for these could be used not more than once for
each call. This has allowed frequency analysis to be presented more easily, since it
ties the quality and resolution strategies to the problem type, rather than to the unique
call identifier.
Transcripts of interviews and case notes from direct observation have also been
coded using this tool. Each document has been imported into NVivo and manually
coded to ensure that the context of the relevant node is captured. NVivo's search tool
has been used to identify combinations of themes (such as knowledge reuse and
strategic position of the call centre) during interviews. Simple searches referring to
the predetermined nodes have revealed references to areas of consideration. The
Boolean matrix intersection and difference searches provide the composite tables
(Table 14) used below to identify the patterns of interest for the actual calls.
Pat
tern
s fr
om
Obse
rvat
ions,
Inte
rvie
ws,
and R
esponse
s
94
R
esp
on
se S
trate
gie
s f
rom
Co
ded
Call
s
Zack T
ypes
Cyc
lic
al
refr
am
ing
D
e-
co
mp
ose
Me
an
ing
th
rou
gh
d
isc
uss
ion
R
eq
uir
es
re
se
arc
h
So
cia
l c
on
str
uc
tio
n
Te
rmin
ate
C
all
To
ok
P
ers
on
al
De
tail
s
Tra
ns
fer
to
us
er
Tra
ns
fer
to
Fu
nc
. A
rea
Us
e
co
dif
ied
in
form
ati
on
T
OT
AL
C
all
s
Am
big
uit
y
7
0
14
1
2
1
1
7
2
4
15
Co
mp
lex
ity
5
9
11
2
0
0
8
8
3
7
17
Eq
uiv
oc
al
0
0
0
0
0
0
0
0
0
0
0
Un
ce
rta
inty
9
1
63
1
3
2
52
28
37
63
111
Am
big
uit
y
46.7
%
0.0
%
93.3
%
6.7
%
13.3
%
6.7
%
6.7
%
46.7
%
13.3
%
26.7
%
10.5
%
Co
mp
lex
ity
29.4
%
52.9
%
64.7
%
11.8
%
0.0
%
0.0
%
47.1
%
47.1
%
17.6
%
41.2
%
11.9
%
Eq
uiv
oc
al
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n
/a
Un
ce
rta
inty
8.1
%
0.9
%
56.8
%
0.9
%
2.7
%
1.8
%
46.8
%
25.2
%
33.3
%
56.8
%
77.6
%
Q
uality
Cri
teri
a f
rom
Co
ded
Call
s
Zack T
ypes
Clo
su
re
Sa
tis
fac
tory
D
iffi
cu
lt t
o
ac
ce
ss
E
as
e o
f a
cc
ess
Em
pa
thy
wit
h
cu
sto
me
r
Ex
ce
ss
Tim
e-i
n-
qu
eu
e
No
Em
pa
thy
wit
h
cu
sto
me
r C
all
Ba
ck
Te
ch
nic
al
res
ou
rce
c
ap
ab
ilit
y
Te
ch
nic
all
y
co
mp
ete
nt
Un
-s
ati
sfa
cto
ry
clo
su
re
TO
TA
L
Ca
lls
Am
big
uit
y
10
0
0
14
0
0
0
0
3
1
15
Co
mp
lex
ity
11
1
0
16
1
0
1
6
7
0
17
Eq
uiv
oc
al
0
0
0
0
0
0
0
0
0
0
0
Un
ce
rta
inty
71
0
3
90
0
2
12
31
36
2
111
143
Am
big
uit
y
66.7
%
0.0
%
0.0
%
93.3
%
0.0
%
0.0
%
0.0
%
0.0
%
20.0
%
6.7
%
10.5
%
Co
mp
lex
ity
64.7
%
5.9
%
0.0
%
94.1
%
5.9
%
0.0
%
5.9
%
35.3
%
41.2
%
0.0
%
11.9
%
Eq
uiv
oc
al
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n
/a
Un
ce
rta
inty
64.0
%
0.0
%
2.7
%
81.1
%
0.0
%
1.8
%
10.8
%
27.9
%
32.4
%
1.8
%
77.6
%
Ta
ble
14
- F
req
uen
cy o
f R
esp
on
se S
tra
teg
ies
Patterns from Observations, Interviews, and Responses
95
6.3. Responses to Calls
Uncertainty
The predominant problem type presented to the call centres falls into the category of
uncertainty, as defined by Zack (2001). Such problems are characterised by lack of
knowledge about current and future states. Often the caller simply states, “I am
looking for information about XYZ” at the commencement of the call. This generally
indicates an uncertain problem type. Not all calls are so easily categorised, however,
and in some cases this simple lead-in statement points to another problem not so
easily resolved. A typical uncertain call follows the following pattern of this
transcript of an actual conversation:
Caller: “We’re on a property and we want to get into wild flowers growing. Is
there any information that you can send me that will help me out?”
Agent: “I do have some information that’s on wild flowers – it will depend on
what species you are talking about. Any in particular that interests you?”
Caller: “Oh – No, nothing really, we just sort of rung because we want to make
it sort of work for us – that’s all …. and somebody suggested wild flowers you
know they seem to be quite in the export you know side of it….”
The sample of calls contains 111 with uncertain problem types. Significant strategies
employed by call-centre agents in answering these calls are enumerated below:
a) Meaning through discussion (57 percent) is used primarily to clarify the real
needs of the caller (Weick, 1995). This strategy also links to the quality
characteristic of empathy. The consequence of this apparent display of
interest in the problem by the agent is that it provides a perception of quality
to the caller as well as confirmation of the real problem to be solved.
b) The agents access codified information (57 percent) in the resolution of a
majority of calls, an expected tactic based on interview data, strategies
recommended by Zack (2001), and the investment in structured internal
knowledge repositories. These repositories not only provide the necessary
information, but also the architecture to ensure (routine) successful search
outcomes, and hence ensure consistency of advice.
Patterns from Observations, Interviews, and Responses
96
c) Personal details are recorded (47 percent) as a matter of policy in CC-B to
satisfy the requirements of the call-centre management software. In CC-A,
although not formally recorded, notes are regularly made both to prompt the
use of the caller’s name in conversation, and also to fulfil a contract such as
the mail-out of fact sheets if required. This again supports the quality criteria,
but is also used by CC-B to identify the nature of previous calls when a
follow-up call is received and referenced by the caller.
d) Transfer to functional area (33 percent) occurs when the required knowledge
lies outside the expertise of the agent and not apparently available from the
structured internal knowledge repositories. Agents who identify a need for
knowledge on recurring topics are able to request fact sheets, briefings, or
other information to reduce the need to rely on the transfer call strategy.
e) Transfer to user (25 percent) normally indicates a lack of necessary
information able to be provided by the caller. These calls are recorded as
successfully closed since the call centre has lost control of the process, but is
able to provide direction on the next steps towards obtaining a satisfactory
solution (such as how to locate a vehicle identification number prior to
providing advice on the status of the vehicle).
f) Cyclical reframing (8 percent) reflects a trial–and- error approach where the
initial discussions indicate a lack of clarity for the agent (Weick, 1995).
Although not a widely used strategy for handling problems of uncertainty, it
is useful for the alignment of perceptions of both the caller and agent.
g) Other documented strategies have not been employed to the degree that they
are of significance in the analysis of uncertain problems.
Complexity
Problems of complexity are characterised by a diversity of issues, a number of
variables, and their interrelationships. This implies a potential multiple–method
analysis, producing many (differing) solutions. Often characterised by statements
such as, “I just don’t know where to start,” these problem types are surprisingly easy
to identify (based on the 100 percent correlation between researcher and assistant in
the coding process). A typical complex problem call follows the following pattern of
this transcript of an actual conversation:
Patterns from Observations, Interviews, and Responses
97
Caller: “I’m trying to identify two types of grasses that grow near my house
but I have no idea who I would sort of ring or how to find out this
information.”
Agent: “OK Now these grasses are a problem or...”
Caller: “No we’re actually trying to identify them because my dog has an
allergic reaction to either one of them and I’m trying to find information…”
This has become more than just a simple grass-identification exercise which would
have resulted in a reference to a taxonomy for grasses. The multiple problems and
the relationship between these led to a resolution based on a range of branched
decisions, such as, “If the initial reaction started during a particular time of the year,
then investigate this, else ...”
The sample of calls contained 17 having complex problem types. Significant
strategies employed by call-centre agents in addressing these problems are specified
below:
a) Meaning through discussion (65 percent) has been used primarily to identify
the major issues involved in the problem. This strategy also links to the
quality characteristic of empathy. The consequence of this apparent display of
interest in the problem by the agent is that it provides a perception of quality
to the caller as well as confirmation of the real problem to be solved.
b) Decomposition (53 percent) is the strategy recommended by Zack (2001).
This strategy requires the larger problem to be broken down into smaller
ones, with the reintegration of the simplified elements directed towards
achieving a satisfactory solution.
c) Personal details are recorded (47 percent) as a matter of policy in CC-B, to
satisfy the requirements of the call-centre management software. In CC-A,
although not formally recorded, notes have been regularly made both to
prompt the use of the caller’s name in conversation, and also to fulfil a
contract, such as mail-out of fact sheets if required. This again supports the
quality criteria, but is also used by CC-B to identify the nature of previous
calls when a follow-up call is received and referenced by the caller.
Patterns from Observations, Interviews, and Responses
98
d) Transfer to user (47 percent) normally indicated a lack of necessary
information able to be provided by the caller. These calls have been recorded
as successfully closed, since the call centre has lost control of the process, but
has been able to provide direction on the next steps to obtaining a satisfactory
solution. This proportion initially appears to be high, but considering the
nature of the decomposition process, it is likely that the resolution of some
sub-problem is likely to require further investigation by the caller.
e) The agents access codified information (41 percent), since once simplified (to
the degree that they become familiar and solvable), the sub-problems often
present as being uncertain. The strategies then reflect those which have been
discussed above. These findings indicate a hierarchical relationship between
problem types, as per Figure 19, where the application of resolution strategies
to a complex problem may either provide a resolution directly, or produce a
problem of uncertainty, which in turn requires an appropriate strategy.
Figure 19 - Complex Problem Devolution
f) Cyclical reframing (29 percent) is used more widely in solving problems of
complexity than those of uncertainty, since the method of attack in
decomposition of a complex problem often requires a “trial-and-error”
strategy (at least in the initial stages). As suggested by Zack (2001), this
convergent process leads to identification of sub-problems which are either
familiar or able to be solved by a prescribed method.
g) Transfer to functional area (18 percent) occurs when the required knowledge
lies outside the expertise of the agent, and is not apparently available from the
structured internal knowledge repositories. In terms of transfer-to-user
strategy, it is likely that a sub-problem will require knowledge from an expert
who has experience in the area being addressed.
Complex Problem
Complex Problem Strategy Uncertain
Problem
Uncertain Problem Strategy Problem
Resolved
Patterns from Observations, Interviews, and Responses
99
h) Queries requiring further research (12 percent) are not considered closed until
the outcome of the research is relayed back to the caller. These calls reflect a
knowledge gap on the part of the agent, but the tier-2 support available is
unclear. Rather than commencing a referral cycle,31 the agent takes the
problem off-line, having a personal commitment to resolve it and to
communicate the outcome to the caller. Although the response addresses the
quality criteria of empathy and competence, it is not registered as
satisfactorily closed at the time of the call.
Ambiguity
The final problem type encountered in the call centres is one of ambiguity (Zack,
2001). These problems present in calls which tend to skip around a range of issues
without any apparent clear request from the caller. Often many questions, seemingly
unrelated either to each other or to the functions of the organisation, are asked,
without the opportunity to provide a logical solution. The following actual transcript
of a conversation illustrates the ambiguous problem type.
Caller: “My...my son….I’m I’m divorced….my son has been taken .. he sort
of ...he’s unaware it is illegal. He’s only 7 years of age but um when I report
it .. if I ask the Boating and Fisheries... um if this guy is caught …because
my son’s involved in it. Would they notify me?”
The example above relates to an anonymous caller trying to get the partner of his ex-
wife into trouble with the Boating and Fisheries authorities, but without implicating
his son.
Problems of ambiguity are generally associated with a degree of anxiety as well. This
reflects findings by Weick (1995) that such problems cause concerns about the
rationality and orderly nature of their environment. From the sample of calls, 15 have
been classified as being ambiguous, with one of these calls being terminated early
(i.e., undergoing unsatisfactory closure) owing to the inability of the agent to be able
to engage the caller in a rational discussion. In these problem types, the ability to
31 An issue regularly referred to in discussions with staff has been the desire to remove “telephone tag”, where callers may wait in a queue for long periods only to be told they are in the wrong area and be transferred to another holding area awaiting response. Both management and agents have expressed a desire to minimise referrals predominantly for this reason.
Patterns from Observations, Interviews, and Responses
100
make sense from what is being sought by the caller is of more importance than the
actual knowledge demands. Once agreement is reached on what the actual problem
is, appropriate knowledge strategies are able to be employed as per the Q-R cycle
(Timbrell et al., 2005b) The resolution methods require cycles of interpretation
without the introduction of new materials (Weick, 1995; Zack, 2001), and are
reflected in this case study. Below is an analysis of the ambiguous problems raised
in the call set.
a) Meaning through discussion (93 percent) has been used primarily to identify
the major issues involved in the problem. This strategy also links to the
quality characteristic of empathy, which is essential in removing the feelings
of irrationality of the caller. These discussions then provide the opportunity to
clarify the issues in the mind of the caller.
b) Cyclical reframing (47 percent) follows the initial discussions. The aim of
this strategy is to converge what the caller actually required to a single
meaning or understanding, and then to bring about an acceptable solution.
c) Transfer to user (47 percent) indicates a lack of necessary information able to
be provided by the caller. This strategy reflects the expected behaviour of
callers who are, as a result of the clarification process, able to attack the
problem with their own resources supported by a more rational decision-
making context.
d) The agent’s access codified information (27 percent) predominantly from
external sources, rather than structured internal knowledge repositories used
for complex and uncertain problems. This often reflects confusion on the part
of the caller about the responsibilities of the agency. Once the problem has
been clarified, however, agents take the opportunity to access publicly
available information (e.g., via other public-sector Web sites) in order to
deliver a satisfactory closure. The frequency of use of this strategy again
reinforces the hierarchical nature of the problem-solving strategies used
where ambiguous problems may be translated into either complex or
uncertain problems which are then resolved with appropriate strategies. This
is illustrated in Figure 20.
Patterns from Observations, Interviews, and Responses
101
Figure 20 - Problem Devolution Hierarchy
e) Social construction (13 percent) is significant in that the resolution of
ambiguity may involve convincing the caller that the agent is pursuing the
right objectives in the right way. This active process utilises the caller’s
senses of perception, attention, and memory to construct “new” knowledge
through collaboration and cooperation.
f) Other documented strategies have not been employed to the degree that they
are of significance in the analysis of ambiguous problems.
Problems of equivocality have not been encountered in the study. This phenomenon
is a consequence of the nature of business in public-sector bureaucracies, where rules
and policy-based activities are thoroughly researched – precisely to remove
ambiguity and equivocality. While such multiple meanings or conflicting
interpretations may exist at the political level, the departmental delivery of services
aligned with public policy is generally clearly articulated in business rules. Hence,
presentation of equivocal problems will likely be met with emphatic, “Our policy
position is that XYZ is the case.” This is a strategy that raises issues of service
quality for the enquirer.
6.4. Service Quality
Table 15 is a composite representation of the frequency of the identified quality
criteria for each problem type derived from the NVivo matrix search of coded calls.
Figure 21 graphically represents these frequencies to identify the variations in their
application by agents.
Complex Problem
Complex Problem Strategy
Uncertain Problem
Uncertain Problem Strategy
Problem Resolved
Ambiguous Problem
Ambiguous Problem Strategy
Patterns from Observations, Interviews, and Responses
102
Quality Criteria from Coded Calls
Zack Types Satisfactory Closure
Empathy with
Customer
Needs more Info - Call
Back
Technical Resource Capability
Technically Competent TOTAL
Ambiguity 10 14 0 0 3 15
Complexity 11 16 1 6 7 17
Uncertainty 71 90 12 31 36 111
143
Satisfactory
Closure
Empathy with
Customer
Needs more Info - Call
Back
Technical Resource Capability
Technically Competent TOTAL
Ambiguity 66.7% 93.3% 0.0% 0.0% 20.0% 10.5%
Complexity 64.7% 94.1% 5.9% 35.3% 41.2% 11.9%
Uncertainty 64.0% 81.1% 10.8% 27.9% 32.4% 77.6%
Table 15 - Frequency of Quality Criteria
Strategies to develop empathy with the caller directly reflect the emphasis placed on
training in call-centre procedures. The development of a service culture has been
observed to be very strong in both call centres, with scripted introductions followed
by ad-lib discussions clearly in evidence in the strategies used. The early
establishment of rapport with callers with problems coded as uncertain (81 percent)
satisfies the “empathy quality” criterion. The problem types that require rich
interactive conversations (complexity, 94 percent; and ambiguity, 93 percent) reflect
a high correlation between the response strategy of developing meaning through
discussion, and the development of empathy with the caller.
Patterns from Observations, Interviews, and Responses
103
Figure 21 - Graphical Representation of Frequency of Quality Criteria
Technical resource capability32 has contributed to the quality of responses to
problems of uncertainty (28 percent). This is displayed through well-defined
search/navigation processes surrounding the familiar structured internal knowledge
repositories which relate to agencies’ core business. For complex problems, this
capability (35 percent) has been used relatively more often. It has extended to
broader knowledge sets due to the lesser degree of familiarity with the issues raised
by the callers.
Technical competence33 is reflected in the ability of an agent to recall facts and past
experiences relating to the query. This ability is seen to be involved in the resolution
of problems of uncertainty (32 percent), and is aligned with the business of the
agency.
32 Technical competence is the accumulated impact of the machines, materials, facilities, technologies, and databases available for use in resolving the problem. 33 The knowledge and ability to attack the range of problems directed at the agent.
Quality Criteria from Coded Calls
Satisfactory closure
Empathy with customer
Needs more information - Call back Technical resource capability
Technically competent
Ambiguity
Complexity
Uncertainty
Patterns from Observations, Interviews, and Responses
104
A further aspect of technical competence is the ability to apply rules and heuristics to
the ambiguous (20 percent) and complex (41 percent) problems in order to realise the
necessary preconditions (clarification and simplification) for resolving the query. All
agents in CC-B demonstrated this competence, while the IVR technology of CC-A
effectively routed the complex and ambiguous calls to specialist sections within the
call centre.
Satisfactory closure (approximately 65 percent) is coded as a quality action only if an
active statement is used by the agent or caller to indicate a satisfactory resolution has
been obtained. Agents often specifically ask if they can be of further assistance, or if
the advice has met their client’s requirements, while callers have often thanked the
agent for their assistance. The noted low frequency of unsatisfactory calls represents
the number of times either the agent or the caller terminated the process prematurely.
This coding rule has also been used for other negative criteria; unless specifically
encountered, however, these criteria are not encoded.
6.5. Summary
Although much time has been spent in building a rich description of the
organisations, their business drivers and management expectations, the actual calls to
the call centres CC-A and CC-B provide the best insight into the ability of agencies
to provide quality advice in an efficient way. The nature of the business of the
public-sector departments studied has determined the types of calls received, while
the different technical and organisational structures within the departments have
influenced management responses to problem solving within their call centres.
Although having different underpinning philosophies and policies in order to provide
satisfactory responses to callers almost 100 percent of the time, both call centres
have systems and procedures backed by a range of knowledge repositories. The
theoretical models and resolution strategies of Zack (2001), Weick (1995), Pentland
(1991), Markus (2001), and Brogowicz et al. (1990), have all been displayed in the
activities of the call centres, indicating at least at an intuitive level that the
management and staff are pursuing practices which align with their business needs.
The technical elements have been coded and reported with predictable or justifiable
frequencies, while the social aspects of the interactions are not so well described by
the models and taxonomies used in the data-collection processes.
Patterns from Observations, Interviews, and Responses
105
The relationship between calls and responses is documented in Table 14 - Frequency
of Response Strategies and Table 15 - Frequency of Quality Criteria. The application
of the Zack (2001) problem categorisation to caller problems, and analysis of
response strategies by agents provide a description of work practices of these public-
sector call centres. This view of the organisation is transparent to both callers and
agents, but when both groups are specifically questioned on strategies employed to
deal with the problem types encountered, an almost subconscious response is
triggered. The agents display a natural ability to recognise the method of attack in an
instant, which has a close alignment with the findings of Zack (2001) in terms of
resolution, and with Brogowicz et al. (1990) for quality issues. Although the coding
reflects activity at a quite general level, the results reflect the critical elements
necessary to support the business of the department in taking calls and providing
responses.
A Pragmatic Approach to Call-Centre Strategies
106
Chapter 7
A Pragmatic Approach to Call-Centre Strategies
A traditional model for call-centre performance, represented in Figure 22 below,
suggests a simple independent relationship between quality and call-time. This study
has revealed a variety of both documented and enacted performance goals dependent
on the role of the performer. The managers of CC-B have indicated that call-time
has not been an issue, but have relied heavily on the grade of service reporting which
is primarily based on time-in-queue. They considered quality of service to be the
major determinant of individual performance, but monthly reviews were aimed at
moving all agent performance to the fast resolution and high quality quadrant in
Figure 2234.
Figure 22 - Resolution Rate - Quality Matrix
CC-A management placed an emphasis on measurable activities displayed on the
call-centre display boards (time-in-queue, talk-time, abandonment rate), and seemed
much more concerned with throughput and efficiency. However, all observations
place agent activity in the right half of the matrix meaning that they were striving to
deliver a high quality service. The study has found that no evidence of time-
watching existed in the behaviours of either call-centre agents, suggesting that for
their purpose, the concern has simply been to deliver a quality service with an
34 The Resolution Rate - Quality Matrix was used by one of the call centre supervisors to record and report on relative performance of agents. The origin of the tool is unknown.
Quality (as per Quality Monitoring Process
Fast, but lacking quality
Slow and lacking quality
Fast and high quality
Slow, but high quality
Ca
ll R
es
olu
tio
n R
ate
A Pragmatic Approach to Call-Centre Strategies
107
expectation that management would address any unit-wide issues around hard
metrics.
The incorporation of time-based components in call-centre performance cannot be
ignored; it is important in the drive for efficiency as the stated business driver for the
formation of CC-A. Timbrell et al. (2005b) posit a model that links service quality to
pay-off assessment on the part of the caller. The pay-off assessment is the critical
decision to determine if the query has been sufficiently addressed (successful
response); has not been adequately answered, but wishes to continue with the
interaction (the Query-Response Cycle continues another loop); or does not appear to
be converging to a solution (unsatisfactory conclusion to the call). Hence, it is the
caller who indirectly determines the number of cycles traversed, and consequently,
the time spent on an issue. This time has a cost, both to the caller and the agency. It
is this cost which determines the inertia of the interaction, and has been quantised in
customer loyalty literature as the perceived investment in time, money, and effort by
the client (Ranaweera & Neely, 2003). Given that the services provided by the call
centres studied are primarily free of charge, the formula for pay-off assessment is
reduced to demands on the time and the effort of the caller. Quality measures used as
a basis for improvement in call-centre performance in the public sector need to
account for such demands on callers.
7.1. Service Value
A positive pay-off assessment does not necessarily imply a satisfied customer.
Service value (as perceived by the caller), a combination of service quality, and
client satisfaction is the initiator of desired client behaviours (Dedeke, 2003).35 One
of the objectives of this study has been to identify improvement strategies without
the need to rely on exit satisfaction-type surveys. Service value better equates to pay-
off assessment, since it incorporates the cost perceptions of callers through their
behaviours (to terminate or continue with a call). The proposition, then, is that an
improvement in service value will deliver an associated improved call-centre
performance.
35 These include displays of trust, assurance, and either acceptance of the advice or permission to
continue.
A Pragmatic Approach to Call-Centre Strategies
108
Personal qualities
Service value is derived from a range of characteristics. The personal qualities of
call-centre agents - empathy, reliability, responsiveness, assurance, and tangibles
(such as a clear voice) - directly contribute to caller satisfaction (Parasuraman,
Zeithaml, & Berry, 1988). In their 1999 study, De Ruyter et al. identify drivers of
customer behavioural intent, finding that they differed markedly across six different
service industries, as listed in Table 16 (Dedeke, 2003, p. 284). The strategy implied
is to deliver on these drivers in the same proportion as expected by the caller, and to
improve on this marginally to exceed that expectation. Below is an assessment of
both the service criteria used by De Ruyter et al. (1999) and their impact in the
Queensland Government call-centres being studied.
a) At an individual level, the ability to establish an empathetic relationship with
callers quickly was evidenced across both call centres. To some degree this
has been a consequence of the problem-resolution process, but its pervasive
existence represents a group norm which is either identified as part of the
recruitment process, or established during the pre-service induction. De
Ruyter et al. (1999) have found expectations of consumers of government
services in establishing empathy occurred in only 5 percent of responses. This
is quite a low benchmark: one which is well exceeded by staff in CC-B and
CC-A. This implies no anticipated improvement from strategies to improve
the call-centre staff’s empathy with callers.
b) Reliability of government services (rating 17 percent in the 1999 De Ruyter
study) ranked second to public transport in criticality. This issue has not been
investigated in this call-centre study, since the continued existence of the
service is only threatened by a technical catastrophe (for which business-
continuity plans exist), or a political decision to alter the service.
c) The grade of service measures in place directly addresses responsiveness
characteristics (rating 27 percent in the 1999 De Ruyter study). In general,
Government has the highest rank of all services when measured against this
criterion. Any improvement strategy aimed at this area should deliver
improvement in client satisfaction. Time-in-queue and abandonment rates are
the hard metrics to be minimised, while the ability to address concerns of
callers once connected to an agent provides scope for enhancement. Such
A Pragmatic Approach to Call-Centre Strategies
109
strategies are closely aligned with the technical determinants addressed
below.
d) Assurance (rating 15 percent in the 1999 De Ruyter study) is ranked as
second lowest of the six services. The decision to omit this attribute from this
study is supported by the unique business position of each department. The
two agencies have no competitors in the marketplace, and each has evolved to
a point where the organisations are respected, particularly in key
representative stakeholder groups. Supporting this position is the lack of any
negative responses to the quality criteria used in the data-collection process.
No real benefit is likely to be achieved through investment in strategies to
improve assurance.
e) The nature of call centres means that clients interact remotely, and hence do
not experience the ambience of the built environment. Improvement in
tangibles (rating 10 percent in the 1999 De Ruyter study), outside the
generally-accepted standards of telephone communications and published
materials sent to callers, would serve little purpose in increasing perceived
service value for the caller.
Tangibles Reliability Responsiveness Assurance Empathy Total
Shops 12 (13.2%) 13 (11.1%) 40 (20.4%) 38 (16.8%) 28 (17.5%) 131
Restaurants 48 (52.7%) 13 (11.1%) 28 (14.3%) 52 (23%) 8 (5%) 149
Government 9 (9.9%) 20 (17.1%) 53 (27%) 33 (14.6%) 8 (5%) 123
Public transportation 13 (14.3%) 41 (35%) 20 (10.2%) 13 (5.8%) 30 (18.8%) 117
Banking 7 (7.7%) 16 (13.7%) 27 (13.8%) 44 (19.5%) 39 (24.4%) 133
Health care 2 (2.2%) 14 (12%) 28 (14.3%) 46 (20.4%) 47 (29.4%) 137
Total 91 (100%) 117 (100%) 196 (100%) 226 (100%) 160 (100%) 790
Table 16 - Analysis of critical incidents (De Ruyter, Wetzels, & Van Birgelen, 1999, p. 1139)
Technical determinants
Technical determinants provide the largest opportunity for improvement in the
delivery of service value. The ability to resolve issues more quickly will improve
responsiveness (a critical quality contributor to caller satisfaction for government
clients), reduce cost to the caller (time and effort), and reduce demand on call-centre
resources (efficiency principles). Hence, although a practical limit exists on the
ability to effect efficiencies in the problem resolution process, knowledge-based
A Pragmatic Approach to Call-Centre Strategies
110
strategies provide opportunities to improve call-centre performance when applied to
the following technical determinants.
a) Competence surrounding problem resolution has not been considered as a
specific component of any of the in-service strategies employed by the call-
centre managers, and is instead a by-product of the exposure to the
experience of solving problems. The heuristic knowledge which results from
the creative processes of interpretation is a direct outcome of that process
(Tsoukas & Vladimirou, 2001). This is also reflected in the findings of
Pentland (1991) where the software support group in the call centre were
found to “use rules but do not merely follow rules” (Pentland, 1991, p. 216).
A recursive relationship exists between structure and behaviour, which in turn
determines the rituals, systems, and competence reflected in organisational
knowledge (Timbrell, Delaney, Chan, Yue & Gable, 2005a). The work of
Zack (2001) and the findings of this study suggest the problem-solving
competencies are able to be developed, and that active management
intervention is required to target these competencies.
The formal pre-service strategies employed by CC-B are directed towards
learning the policies, procedures, and standards (which are further expanded
later in this section), the business of the department (facts, contacts, and
responsibilities), and the development of a service culture (again, further
addressed later this chapter). The large percentage of problems of uncertainty
encountered in CC-B indicates the emphasis on knowing facts about the
organisation in pre-service programs is an effort well spent. Given the
boutique nature of this relatively small organisation, such an in-depth
knowledge of the Department is able to be achieved within the extended
period of pre-service and mentored initial period of service. Agents develop
their competence in addressing complex or ambiguous problems as a
consequence of the mentoring program, rather than as a result of a planned
management intervention. The group norm based on engendering empathy
with the caller is primarily responsible for addressing ambiguous problems
through active dialogue (“meaning through discussion”). The system of
referral to experts allows novice agents to refer complex problems based on
their knowledge of the organisational structure; the referral decision may be
A Pragmatic Approach to Call-Centre Strategies
111
an uncertain problem type for the agent. With experience, agents develop an
ability to recognise partial solutions (based on factual knowledge) to complex
problems, which in turn encourages the problem decomposition process.
Higher rates of pay and investment in the emotional intelligence of the team
is aimed at maintaining a stable group of agents, which is a necessity given
the strategy employed for problem-solving competency development (Urch
Druskat & Wolff, 2004).
CC-A receives a relatively higher proportion of ambiguous and complex
problem types compared to those of uncertainty. The short pre-service
training process reflects a strategy of becoming familiar with systems and
processes and an in-depth knowledge of the primary information system. This
system contains the information basis for scripted responses, which are based
on a hierarchical structure of functions and related facts, legislation, and
policies. The recruitment process of CC-A does favour applicants with a prior
knowledge of the organisation, but interpersonal experiences are reported as
being more important to selection. The department serviced by CC-A has a
history of regular and significant structural change. The strategies employed
by call-centre management (who also regularly change) mirror the
characteristics of an agile organisation. Very little investment is placed in the
development of the emotional intelligence of the group, which is reflected in
the high turn-over rates of staff. However, the centre manages to operate
effectively. The problems of uncertainty are able to be addressed through the
reasonably static, but highly robust information systems. Complex and
ambiguous problems are able to be referred to agency experts, who contrast
the demographic of the call centre in that they have been working in the area
for a long time (see Figure 13 - Age Demographics of staff employed by the
agency hosting CC-A). They may also result in the agent taking the call off-
line in order to perform further research, which is an option not available in
CC-B.
b) An important factor in the success of knowledge-reuse strategies is the
development and maintenance of knowledge repositories (Markus, 2001).
Generally, consideration is only given to formal, relatively static knowledge
sources such as Web sites, functionally specific databases and fact sheets,
A Pragmatic Approach to Call-Centre Strategies
112
libraries, and archival records. Search strategies associated with these sources
are well-defined, and mitigate against concerns of information overload.
Development of a successful knowledge repository strategy requires
significant effort, since it may be based on a creator documenting for their
own needs only, for others in the call centre (having similar needs), or for a
different audience altogether. Motivation of call-centre agents to contribute to
knowledge repositories may also be a limiting factor. Pressures on time, ease
of data capture, confidence in personal abilities, and pay-back for the effort
will all impact on the rate of contribution to shared repositories by authors. In
both call centres studied, agents had effective methods of capturing
knowledge for their own reuse (often showing as paper-based processes). The
knowledge sources produced for dissimilar audiences have primarily been the
responsibility of other organisational units with management processes in
place for their maintenance. Markus (2001) provides direction in the
development of a knowledge repository for capture and reuse by agents for
agents in their roles as:
i. Shared-work producers (maintain context, support searching, document
the rationale);
ii. Shared-work practitioners (publish context along with content, package
expertise, push knowledge to recipients, provide incentives for
contribution and reuse);
iii. Expert-seeking novices (support re-contextualisation of knowledge in
the local context, capture knowledge using understandable terminology,
and training); and
iv. Secondary knowledge miners (metadata stored for secondary use,
structure the knowledge base, develop analytical and synthesis skills,
verify results).
The dilemma faced when considering the development of a knowledge
repository is the degree to which it will contribute to knowledge reuse. CC-B
agents developed individual approaches to documenting responses to new
problems. The ability to stand up and call out “Does anyone know anything
about XYZ?” and expect a positive response suggests that investment in
A Pragmatic Approach to Call-Centre Strategies
113
technology to capture and make available such knowledge may deliver little
benefit. However, team leaders have suggested that many issues are cyclical
(season-dependent), and currently rely on the agent’s ability to recall past
experiences in order to respond to calls successfully. As a trade-off, the
implementation of a policy aimed at the capture of relevant responses to
(identifiable) cyclical issues would assist in obtaining consistency of advice
on matters as they recur over time. No new technology would be needed,
since CC-B has call registering software (log the caller, issue, etc.). It does
not, however, record resolution strategies. Current analysis is limited to
frequency reporting on particular issues, which may initiate a management
response, such as seeking further input from departmental experts. The nature
of the evolution of CC-B means that no single (integrated) knowledge
repository (or linking technology) exists. Hence, the agent’s knowledge of the
organisation is an essential pre-condition of successful knowledge repository
use in CC-B. The requirement to open many databases simultaneously has
been a source of nuisance to agents, both because of the cluttered desktop, but
also because of its impact on performance. The specification of an
information architecture capable of identifying subject matter across
databases should be the goal of CC-B management. This would allow the
incremental implementation of technologies to provide a consolidated view of
the codified knowledge held in the agency.
The maintenance of the single knowledge base for functional matters in
CC-A is the responsibility of the Training Officer. Changes to legislation,
feedback from agents, and pro-active planning by the CC-A team initiate
updates to this information source. However, this system is not capable of
capturing the organisational logic underpinning its use. Neither actual data
nor metadata on the advice given to a complainant is captured for future
reference. This is consistent with the current procedure documented in the
knowledge base. Formal (written) complaints are addressed with more rigour,
but the electronic information systems supporting this process do not make
such information easily accessible to call-centre agents. This provides the
potential for the provision of conflicting advice to similar queries, and loses
the opportunity to provide assistance for similar queries in the future.
A Pragmatic Approach to Call-Centre Strategies
114
Changes to incorporate such a strategy in systems would be expensive, but
the dynamic nature of CC-A indicates a cost and quality benefit would
accrue.
c) Davenport and Prusak (1998) advise that a process perspective be adopted in
the creation of a knowledge environment. This approach incorporates the
formalisation of routine procedures embedded in normative activities
specified by rules and standards, as well as core team knowledge processes of
knowledge auditing, development, mutual updating, briefing, and reviewing
(Eppler & Sukowski, 2000).
Documented procedures exist in CC-B in the form of flowcharts to assist
agents in working through a generic problem-resolution process. Although
these procedures show the steps involved, they are unable to provide
information on the significance of the intellectual activities encompassed in
each step. The culture of trust and sharing displayed by the agents in CC-B
helps fill this void. Consequently, a consistent standard of service is
displayed, and rituals reinforcing such behaviour are common-place. No
formal process exists for knowledge auditing. However, given the relatively
small size of the unit and the high amount of interaction between staff, a
broad knowledge of agents’ skills, experiences, expectations, and motivations
is shared among team members. Team knowledge development is achieved
through adherence to formal responsibilities to maintain dialogue with
external reference groups. The early-warning system for campaigns and
responses to requests for fact sheets on topical issues ensures CC-B agents’
knowledge and practice are current. Such interaction with external groups is
supplemented by regular and frequent team meetings designed to provide the
opportunity for mutual updating and briefing. The one area able to deliver
benefit to CC-B is the systematic reflection on its performance, critical
incidents, and recurrent activities.
A comprehensive list of training manuals provided by CC-A evinces a
systematic approach to call-centre (and other service-centre) activities.
However, the approach to work in the call centre is more on an individual
basis, with each agent developing strategies for their own use. Work is
assigned to CC-A staff in a way that attempts to maximise individual agent
A Pragmatic Approach to Call-Centre Strategies
115
effort utilisation. This reduces any opportunity for development of emotional
intelligence (Urch Druskat & Wolff, 2004). The stress associated with the
constant demands of the telephones is somewhat offset by the opportunity to
take work off-line to be completed at a more leisurely pace. The general
impact of the high work rate is that very little time is given to social activities
able to develop a sense of group efficacy. Knowledge development and
mutual updating are facilitated by regular formal team meetings; however,
knowledge auditing and reviewing do not exist. Eppler and Sukowski (2000)
refer to a range of tools supporting a systems approach that would benefit
team knowledge management in CC-A. These include:
i. Team Knowledge Auditing – A team matrix provides a framework to
determine the present and required knowledge state of an organisation,
while expert web provides transparent access to experts and their
external contacts.
ii. Team Knowledge Development – Coordination of knowledge
development and alignment with business goals is achieved by applying
the pyramid principle which decomposes particular business hypotheses
into manageable knowledge development tasks. The systematic
mapping of knowledge present in team discussions is captured using a
Toulmin map (Toulmin, 1958) which disaggregates arguments into
explicit functional components of claim, grounds, and rebuttal etcetera.
iii. Mutual Updating and Briefing – Visual protocols assist in visually
arranging the main issues addressed in meetings. Team members are
able to be kept up-to-date using flight plans which tabulate situational
tasks against the size/impact of the task and chart the status of all team
tasks in a single place.
iv. Reviewing – Case study and the lessons learned inventory are effective
tools for team review.
d) Communications infrastructure is also identified as a technical determinant
for service quality. Most organisations have technologies such as email and
Intranets that allow users to interact with the changing contextual knowledge
of the organisation. These tools encourage performance improvement
A Pragmatic Approach to Call-Centre Strategies
116
through collaboration by capturing, representing, and interpreting their
knowledge resources (Fensel, 2001).
Communication behaviours in CC-B are characterised by the development of
a culture of mutual support. In this environment, informal communications
activities account for a significant proportion of staff interaction.
Conversations over the partitions and in meetings are supplemented by email.
This personalisation strategy reflects the findings of Hansen et al. (1999), in
that the computer is used to “help people communicate knowledge, not store
it”(1999, p. 107). The databases and other published materials are
supplementary to the main organisational modus operandi.
The reliance on the knowledge base within CC-A reflects a strategy of
codification. Resources are applied to ensuring the knowledge base is current
and complete, with an officer assigned to the role of ensuring a consistent and
structured approach to both the insertion of new material and maintenance of
current details. The systems are regarded as the primary source of
information, and are applied where scripted responses are available to the
problems at hand at many levels. The information used to populate these
knowledge bases is also replicated on the Internet,36 but in a different format.
The challenge is thus to keep these sources consistent. The trade-off in the
drive for efficiency is the effort required to build, maintain, and develop
efficient search and navigation methods.
Neither of the knowledge strategies employed is any better than the other.
The goal is that CEOs understand the strengths and weaknesses of both
strategies, thereby choosing the one that best suits their decisions about
knowledge management, and will persist with that approach. Both CC-A and
CC-B reflect the advice of Hansen et al. (1999), in that they predominantly
employ either a codification or personalisation strategy, but also direct some
effort to the subservient strategy. The codification strategy of CC-A relies
36 Customers are encouraged to find the information on the Internet without reference to the call centre. Although the use of the Internet is increasing, the complexity of interpretation of this information means the call-centre agents are needed in many instances.
A Pragmatic Approach to Call-Centre Strategies
117
heavily on IVR technology, a tightly controlled information database, and the
relevant transaction processing systems. There is a much smaller reliance on
referrals based on personal expertise. CC-B, on the other hand, applies
significant resources to training new staff in the business of the agency.
Although documents are gathered on specific topical issues, a far greater
reliance is placed on personal knowledge of the situation, or on the ability to
refer to specific expertise.
Functional service offerings
Functional service offerings provide the cohesion between individual personal
qualities and the system-imposed technical determinants. This allows the customer
to synthesise a perspective of the call centre. The importance of empathy and
responsiveness to client satisfaction, and the problem-solving ability’s contribution
to service value are realised in the degree to which the organisational culture is
attuned to a service culture. This occurs through:
− Software tools, techniques, methodologies, routines, and best-practice support
call-centre activities, and the information required to develop the sorts of
indicators and reports that would be as meaningful for agents as they are for
management;
− Contextual awareness to include a consideration of agents’ aptitudes and skills
of willingness to document information, attention to detail, analytical ability;
and
− Willingness of the agent to share personal caller intelligence, and to engage in
dialogue with others about ways to optimize the use of the encultured norms and
practices (Thompson & Walsham, 2004).
The combination of the attributes above contributes to an organisational culture
overtly attuned to service delivery. The following provides an analysis of the two call
centres and their functional service capability.
A Pragmatic Approach to Call-Centre Strategies
118
a) Unlike organisation-wide measures37 such as financial performance, gross
transaction counts, and quantum of service, process measures provide detail
on tasks and activities at the work-unit level. Meyer (2004) suggests a
guiding principle underpinning effectiveness of teams is that measurement
systems are to allow them, rather than management, to gauge their
performance. By implication, this requires the team to be intricately involved
in the design and implementation of the system. The focus of these measures
is on the identification of strategies by team members to improve their own
performance.
Performance measurement in CC-B has a hierarchical structure. At the work-
unit level, the grade of service levels are reported on a regular basis.
Performance below the nominated benchmark is investigated, but normally
reflects an unplanned event which either dramatically increases the call-
centre load, or reduces the available pool of agents. The policies and practices
of CC-B, which include sufficient resource redundancy to allow for
predictable fluctuations in demand, ensure that the desired grade of service is
effectively achieved. At an individual level, the call rate and quality of call
are measured against a standard criteria set, and plotted against (anonymous)
others to provide a graphic of relative performance compared to all agents.
The protocol-based processes are assessed, and reports and
corrective/improvement actions are negotiated for the next period. In line
with the recommendations of Meyer (2004), the development and
maintenance of call protocols is carried out in team meetings.
The critical reporting criteria in CC-A are call-resolution rates, time-in-queue
and abandonment rates. Agents are scheduled between on-line and off-line
tasks, having the ability to load-balance by moving staff between these
modes. The relatively high expected call rates of CC-A leads to tension to
meet the benchmark, and has the consequence of increasing stress in the call
centre. Individual performance is assessed primarily in terms of average talk-
time (with the goal of reducing this), but is supplemented by a comprehensive
37 These measures are predominantly of interest to senior management. In a public-sector organisation, staying within budget is important, as is reporting on the number of services delivered (e.g., number of operations performed in a health district).
A Pragmatic Approach to Call-Centre Strategies
119
list of performance criteria, which cover skills as well as knowledge of
particular functional issues. Feedback on performance and an associated
action plan are provided to the agent after each appraisal.
b) Agent aptitudes and willingness to document information, their attention to
detail, and their analytical ability contribute to the functional service quality
of an organisation (and hence contribute to perceived service value). Agents
will display various degrees of commitment towards publishing their
knowledge for the use of their peers, owing to time, quality, and social
pressures. Successful knowledge reuse requires the provision of proper
incentives for knowledge producers in the development of quality repositories
(Markus, 2001). Given a policy to publish is actively encouraged, the
robustness of the resulting document will reflect the authors’ moods,
demonstrating that they are certain that the content is true, believe that the
content is true, are indifferent to the truth value of the content as either from a
second-hand and indefinite source, or are uninformed about the truth value
content (Gotel & Finkelstein, 1995). Hence, an ability to identify implicit and
derived group contributions in addition to the explicit ones is an analytic skill
required by all agents.
The only knowledge collected by CC-B operators is basic identifying
information and the call’s topic to facilitate call-back if required. Referrals to
previous calls only serve to identify frequency and cyclical timing issues that
are also highlighted in reporting cycles. Agents within CC-A are not required
to formally collect or publish any call-related information, unless it is the
subject of a particular marketing campaign. An opportunity exists in both call
centres to expand the agent’s capability to capture and publish resolution
strategies and outcomes in order to exploit knowledge transfer and reuse.
7.2. Summary
The common elements of this study are limited to call centres and the Queensland
public-sector environment. The differences in goals (effectiveness versus efficiency)
and organisational components (workload, performance management, technology)
are reflected in predominantly evolutionary knowledge strategies (personalisation
versus codification). In each situation, the business drivers have quite rightly
A Pragmatic Approach to Call-Centre Strategies
120
determined the strategies implemented. Personal and technical determinants have
contributed most significantly to each call centre’s performance, with problem
resolution reflecting the methods proposed by Zack (2001) in relation to uncertain,
complex, and ambiguous problem types. Functional service offerings, although
supported by performance management systems, are able to be improved through the
untapped knowledge reuse opportunity to capture and share resolution strategies.
Conclusion
121
Chapter 8
Conclusion
8.1. Introduction
This final chapter is a discussion of the overall learnings from the study. A
conceptual link runs from the presentation of a problem in the first instance, through
a search of the literature to identify related theory and practice, which is followed by
the adherence to a defensible research methodology, to the development of
appropriate research protocols. This case study of two Queensland public-sector call
centres is directed by this protocol (incorporating access to sources of data, sampling
strategies and ethical standards), and by an analytical framework applied to converge
the sources of evidence into patterns and explanations. Certain implications,
limitations, and opportunities for further research are also discussed.
The genesis of the study is based in seeking the improvement of performance of
advice-giving in the Queensland public sector, with call centres being the study’s
focus, since they provide a stable and manageable set of operational variables. A
review of the literature has indicated that:
a) Very little research existed to inform call-centre managers on performance
improvement using principles of knowledge management;
b) Public-sector service delivery has different business drivers from the private
sector;
c) Theories, models, and frameworks existed for organisational problem
solving, role definition, performance measurement, and quality improvement,
but no model existed for their integration into a call-centre environment; and
d) Principles underpinning knowledge management provided fundamental
guidelines in the search for improvement strategies in service delivery.
The structured case study methodology has provided an opportunity to evaluate
evolving theories and knowledge. A cyclical process of collection of evidence and
analysis and reflection allows further development of these themes until a resolution
of the research problem is achieved. Essential to the evidence collection process has
been the Query-Response Cycle, as proposed by Timbrell et al. (2005b), which has
Conclusion
122
provided a model for problem resolution within a call centre. A taxonomy of
problem types and resolution strategies provided by Zack (2001), and quality criteria
defined by Brogowicz et al. (1990) forms a fundamental framework for the study,
with Pentland (1991) providing a coding method to capture the detailed behaviours
in the resolution of individual calls. Finally, Markus (2001) provides the model used
to distinguish roles within a call centre and strategies for performance improvement
based on the knowledge reuse of agents as they perform any or all of these roles.
A broad set of sources has been used in the collection of the data. Guidance from Yin
(1994) has been used to design the protocols and case study activities to ensure tests
for construct, internal and external validity, as well as reliability have been met.
These sources have included direct observation, open-ended interviews, structured
interviews, archival documents, and recorded phone calls. Data from these sources
has been coded according to problem type, resolution strategy, and quality criteria. It
has been analysed through documents generated using relationship identification
capabilities of NVivo qualitative research software.
An analysis of the data in Chapters Six and Seven has indicated that, although each
call centre has differing underpinning philosophies and policies, they both have
systems and procedures to provide satisfactory responses to callers almost 100
percent of the time. The theoretical models and resolution strategies of Zack (2001),
Markus (2001), and Brogowicz et al. (1990), have all been displayed in the activities
of the call centres, indicating, at least at an intuitive level, that the management and
staff are pursuing practices which align with their business needs.
8.2. Research Question
“How well does Zack’s (2001) framework represent problem types, and
hence impact on strategies utilised in response to queries encountered in
a public-sector call-centre environment?”
The process used to answer this question has been to:
a) Identify the business drivers of the organisations from a management
perspective;
b) Determine the degree of alignment of the call centre with those drivers,
through observation and interviews with call-centre agents; and
Conclusion
123
c) Collect and organise evidence to describe the processes used by agents in
their daily work.
CC-A is primarily organised to deliver efficiency benefits to the host department
through a low-cost service-delivery model. Technology plays a large part in CC-A’s
operations, with IVR being used to direct calls to agents with experience in those
matters. CC-B is primarily organised to deliver service quality benefits through the
personalised provision of information. Performance targets require calls to be
answered promptly. Since no pre-screening has occurred, any agent may receive a
query on any matter. Uncertain, complex, and ambiguous problem types have been
encountered during the study; their resolution methods are summarised in Table 14.
The sample of calls contains 111 with uncertain problem types. Significant strategies
employed by call-centre agents in their resolution are as follows:
a) Meaning through discussion (57 percent) has been used primarily to clarify
the real needs of the caller (Weick, 1995);
b) The agents accessed codified information (57 percent), which is an expected
tactic based on interview data, drawing on strategies recommended by Zack
(2001); and
c) Transfer to functional area (33 percent) has occurred when the required
knowledge lies outside the expertise of the agent.
The sample of calls contains 17 with complex problem types. Significant strategies
employed by call-centre agents are identified as follows:
a) Meaning through discussion (65 percent) has been used primarily to identify
the major issues involved in the problem;
b) Decomposition (53 percent), as the strategy recommended by Zack (2001);
c) The agents accessed codified information (41 percent), since once simplified
(to the degree that they become familiar and solvable), the sub-problems
often present as being uncertain; and
d) Cyclical reframing (29 percent) is used more widely in solving problems of
complexity than uncertainty, since, as suggested by Zack (2001), this
convergent process leads to the identification of sub-problems, which are
either familiar or able to be solved through a prescribed method.
Conclusion
124
From the sample of calls, 15 are classified as being ambiguous, and are resolved by
the following strategies:
a) Meaning through discussion (93 percent) has been used primarily to identify
the major issues involved in the problem;
b) Cyclical reframing (47 percent) has been used to converge to a single
meaning or understanding of what the caller actually required, and then to
achieve an acceptable solution; and
c) Transfer to user (47 percent) reflects the expected behaviour of callers, who,
as a result of the clarification process, are able to attack the problem with
their own resources, supported by a more rational decision-making context.
Although the sample is relatively small, the results confirm both the efficacy of
Zack’s (2001) taxonomy, both in its ability to describe problem types, and in the
problem resolution strategies associated with the uncertain, complex, and ambiguous
problem types. The study has also found that technical competence of the public-
sector call-centre agents aligns closely with this taxonomy, even though they have
had no prior knowledge of it.
Problems of equivocality have not been encountered in the study, so no claims are
able to be made with respect to their place in the framework.
8.3. Implications for Theory
How well the theory of problem resolution in call centres is informed by this study is
limited to the consequential observations of the range of practices being performed.
The existence of heuristics employed by the range of agents without specific training
targeted at that resolution strategy implies a social phenomenon exists to support
learning. No reference has been made to resolution-specific coaching or training by
either call-centre manager. Emphasis has been placed on engaging the caller, use of
the script by CC-A agents, and organisation-specific knowledge by CC-B agents.
The aim of this study has not been, however, to develop new theories. Rather it has
been to validate and test a model which integrates a set of pre-existing theories,
models, and principles in their application to a public-sector call centre. To this
extent, the Query Response Cycle (Timbrell et al., 2005b) provides a robust
description of call-centre activity, and is able to be used to investigate the technical
Conclusion
125
and functional determinants (Brogowicz et al., 1990) of the quality problem
resolution in public-sector call centres.
8.4. Implications for Practice
Zack’s (2001) Four Knowledge Problem Model provides a powerful lens through
which to analyse knowledge-based management strategy within a call centre.
Combined with the Query-Response Cycle proposed by Timbrell et al. (2005b), the
technical capability of two public-sector call centres has been documented and
analysed. Although the strategies associated with problem resolution have not been
overtly managed, they have evolved to reflect the framework proposed by Zack
(2001). Even with the relatively small sample of calls analysed, it is clear that the
problem types encountered have reflected the nature of the business of the agency. In
this regard, one of the call centres studied has a complaints-management role, a
factor which has contributed to higher proportions of complex and ambiguous
problem types and the development of strategies to resolve these in an efficient way.
The other, having a policy of providing information, not advice, has developed
significant skills in the solution of problems of uncertainty.
Neither of the managers of the call centres investigated has seen the need to address
their work-day issues explicitly from a knowledge perspective. However, the systems
and processes in place reflect an intuitive appreciation of knowledge management
principles, and management have been keenly interested in the outcomes of the
study. In each case, the reuse strategies relate only to static (i.e., factual and process)
knowledge. Investment in technology to allow the easy capture and location of
successful problem resolution would improve the performance of the call centres.
Markus (2001) recommends the inclusion of specialist intermediaries in both the
design and codification of knowledge processes, since they are able to maintain the
repositories and facilitate the reuse process, and to address problems associated with
incomplete records and indexing.
The government-wide drive for efficiency is effective from the point of view of each
call centre; however, the organisations’ more valuable knowledge resources (tier-2)
are being used more frequently than needed, thus leading to an overall higher cost of
service. By managing the service more closely from a knowledge self-sufficiency
viewpoint, the overall cost of this service to the organisation could be reduced.
Conclusion
126
Examples of this achievement include extending the time allowed for problem-
resolution in CC-A, improving the knowledge resources available to agents,
providing better training in the technical areas, improving agents’ information
literacy skills (i.e., to improve their search capabilities), and developing better
internal processes for handling problems of complexity and ambiguity. These
strategies would all improve the performance of call-centre CC-A. Such a
rebalancing of workforce and technical strategies could improve not only the
efficiency of the call centres’ function, but also improve their effectiveness, thus
leading to higher service quality.
8.5. Further Research Opportunities
a) The research methodology - a structured case research method (Carroll et al.,
1998) and an extension of the coding work of (Pentland & Rueter, 1994) - is
able to be applied to other similar environments. Extending this research to
include private-sector organisations and other service-centre interactions
(such as counter service) would demonstrate the generalisation of the
frameworks used.
b) By extending the tools developed to capture and code the call data to include
the ability to interrogate calls in real time using voice recognition technology,
broader research questions could be investigated. Examples of these include a
time analysis of strategies, and how they evolve from the perspective of new
agents, changes to call-centre policies, or other organisational developmental
instances.
c) The existence of heuristics employed by a range of agents without specific
training targeted at that resolution strategy implies a social phenomenon
exists to support agent learning. An investigation of this may well inform
both academics and practitioners alike on the activities that enforce transfer
of critical knowledge (socialisation, according to Nonaka and Takeuchi
(1995)).
Conclusion
127
8.6. Summary
The research described in this thesis has used a range of models to investigate the
operations of public-sector call centres. The models have been entirely congruent
with the behaviours and practices of callers and responding agents. The applied
nature of this study means it is easily repeatable in any similar agency wishing to
obtain a better understanding of its activities and its problem-resolution capabilities,
as defined in the taxonomy of Zack (2001). At the same time, the thesis provides
recommendations on ways both to measure and improve call-centre performance
through the implementation of repositories to facilitate knowledge reuse (Markus,
2001).
References
128
References
Alavi, M., & Leidner, D. E. (2001). Review: Knowledge management and knowledge management systems: Conceptual foundations and research issues. MIS Quarterly, 25(1), 107-136.
Anton, J. (2000). The past, present and future of customer access centers. International Journal of Service Industry Management, 11(2), 120-130.
Anton, J., & Gustin, D. (2000). Call Centre Benchmarking: How Good is "Good
Enough". West Lafayette: Ichor Business Books/Purdue University Press.
Arzbacher, S., Holtgrewe, U., & Kerst, C. (2002). Call Centers: Constructing Flexibility. In U. Holtgrewe, C. Kerst & K. Shire (Eds.), Reorganizing
Service Work: Call Centres in Germany and Britain. Burlington: Ashgate.
Baker, K., & Badamshina, G. (2002). Knowledge Management. In E. L. Malone (Ed.), Managing Science as a Public Good: Overseeing Publicly Funded
Science: United States Department of Energy.
Becker, M. C. (2001). Managing dispersed knowledge: Organizational problems, managerial strategies, and their effectiveness. Journal of Management
Studies, 38(7), 1037-1051.
Blackler, F. (1995). Knowledge, knowledge work and organizations: An overview and interpretation., 16(6), 1020-1046.
Blue is the colour. (1998, 1998/06/06/). Economist, p. 65.
Bradbury, A. (2000). Develop Your NLP Skills (2nd ed.). London: Kogan Page Limited.
Brogowicz, A., Delene, L., & Lyth, D. (1990). A Synthesised Service Quality Model with Managerial Implications. International Journal Of Service Industry
Management, 1(1), p 27-45.
Brooke, C. (2002). Information Systems at callcentre. International Journal of
Information Management, 22(5), 389-401.
Brown, J., & Duguid, P.(2001) Knowledge and organization: A social-practice perspective. Organization Science, 12(2); 198-214
Bryar, J. (2001). Taxonomies: The Value of Organised Business Knowledge. Retrieved 29/07/2002, 2002, from http://www.newsedge.com/materials/whitepapers/taxonomies.pdf
Caldwell, F. (2004). KM Adoption Model Highlights Critical Success Factors
Research (Research Article): Gartner Research.
Carlsson, S. (2001). Knowledge Management in Network Contexts. European
Conference on Information Systems ECIS 2001, Bled, 616-627.
Carroll, J., Dawson, L., & Swatman, P. (1998). Using Case Studies to Build Theory:
Structure and Rigour. Proceedings of the 9th Australasian Conference on Information Systems, ACIS'98, Sydney.
Choo, C. W. (1996). The knowing organization: How organizations use information to construct meaning, create knowledge and make decisions. International
Journal of Information Management, 16(5), 329-340.
References
129
Close, W. (2003). Contact Centre Quality Management User Survey Results: Gartner.
Cross, R., Liedtka, J., & Weiss, L. (2005). A Practical Guide to Social Networks. Harvard Business Review. 83(3), 124-132.
Davenport, T., Eccles, R., & Prusak, L. (1992). Information politics. Sloan
Management Review, 34(1), 53-65.
Davenport, T., & Prusak, L. (1998). Working Knowledge: How Organisations
Manage What They Know. Boston: Harvard Business School Press.
Davis, G., & Weller, P. (2001). Are You Being Served?: State, Citizens and
Governance. Crows Nest, N.S.W. Allen & Unwin,.
De Ruyter, K., Wetzels, M., & Van Birgelen, M. (1999). How do customers react to critical service encounters?: A cross-sectional perspective. Total Quality
Management 10, 1131-1145.
Dedeke, A. (2003). Service quality: a fulfilment-oriented and interactions-centred approach. Managing Service Quality, 13(4), 276-289.
Di Gregorio, S. (2000). Using NVivo for your literature review. Proceedings of the Strategies in Qualitative Research: Issues and results from analysis using QSR NVivo and NUD*IST, September, London.
Dixon, N. (2000). Common Knowledge: How companies Thrive by Sharing What
They Know. Boston: Harvard Business School Press.
Duder, J. C., & Rosenwein, M. B. (2001). Towards "zero abandonments" in call center performance. European Journal of Operational Research, 135(1), 50-56.
Eppler, M., & Sukowski, O. (2000). Managing Team Knowledge. European
Management Journal, 18(3), 334-341.
Estelami, H., & Maeyer, P. D. (2002). Customer reactions to service provider overgenerosity. Journal of Service Research, 4(3), 205 - 216.
Feinberg, R. A., Kim, I.-S., Hokama, L., de Ruyter, K., & Keen, C. (2000). Operational determinants of caller satisfaction in the call center. International
Journal of Service Industry Management, 11(2), 131-141.
Fensel, D. (2001). Ontologies: Silver Bullet for Knowledge Management and
Electronic Commerce. Berlin: Springer-Verlag.
Gaballa, A., & Pearce, W. (1979). Telephone sales manpower planning at Qantas. Interfaces, 9(3), 1-9.
Gable, G. (1994). Integrating case study and survey research methods: an example in information systems. European Journal of Information Systems, 3(2), pp. 112-126.
Goh, S. (2002). Managing effective knowledge transfer: an integrative framework and some practice implications. Journal of Knowledge Management, 6(1), 23 - 30.
Gotel, O., & Finkelstein, A. (1995). Contribution structures. Proceedings of the Second IEEE International Symposium on Requirements Engineering,, York, England.
References
130
Graumann, M., Arnold, H., & Beltjes, N. (2003). Call centers: A case study on the interplay between organization and information technology. The Geneva
Papers on Risk and Insurance, 28(1), 111-126.
Gronroos, C. (1984). A service quality model and its marketing implications. European Journal of Marketing, 18(4), 36-44.
Gualtieri, J. (2004). When the whistle blows. CA Magazine, 137, 43-45.
Handzic, M., & Chaimungkalanont, M. (2004). Enhancing organisational creativity through socialisation. Electronic Journal of Knowledge Management, 2(1), 57-64.
Hansen, M. T., Nohria, N., & Tierney, T. (1999). What's your strategy for managing knowledge? Harvard Business Review, 77(2), 106-116.
Harris, C. M., Hoffman, K. L., & Saunders, P. B. (1987). Modeling the IRS telephone taxpayer information system. Operations Research, 35(4), 504-523.
Harvey, N., Harries, C., & Fischer, I. (2000). Using advice and assessing its quality. Organisational Behaviour and Human Decision Processes, 81(2), 252-273.
Humphrey, W. (1989). Managing the Software Process. Reading, Massachusetts: Addison-Wesley Publishing Company.
Interview 1a - Annotated notes, January 29, 2003.
Interview 1b - Annotated notes, December 23, 2002.
Interview 1c - Full transcript, June 9, 2004.
Interview 1d - Full transcript, June 11, 2004.
Interview 2a - Annotated notes, May 27, 2003.
Interview 2b - Full transcript, June 16, 2004.
Jonassen, D. H. (2000). Toward a design theory of problem solving. Educational
Technology, Research and Development, 48(4), 63.
Jungermann, H. (1999). Advice Giving and Taking. Proceedings of the Proceedings of the 32nd Hawaii International Conference on System Sciences - 1999, Hawaii.
Kaplan, B., & Duchon, D. (1988). Combining Qualitative and Quantitative Methods in Information Systems Research: A Case Study., MIS Quarterly,12, 570-586.
Kjellerup, N. (2004). Call Centre Know How Essays.. Productivity, Measurements &
Benchmarks. Retrieved 18/06/2004, 2004, from http://www.callcentres.com.au/resource.htm
Kreizman, G., Kotsikopoulos, D., & Shah, J. (2005). How to Build a Business Case
for Service-Oriented Development of Applications in Government: Gartner Research.
Kucza, T. (2001, 9/01/2002). Knowledge Management Process Model. Retrieved 6/08/2002, 2002, from http://www.inf.vtt.fi/pdf/publications/2001/P455.pdf
Leonard, D. (1995). Wellsprings of knowledge: building and sustaining the sources
of innovation. Boston: Harvard Business School Press.
References
131
Ling, T. (2002) Delivering joined–up government in the UK: dimensions, issues and problems, Public Administration, 80(4), 615-642
Mahon, J. (2003). Social issues in management conference symposia abstracts. Academy of Management Proceedings, pp. 1-14.
Majchrzak, A., Neece, O. E., & Cooper, L. P. (2001). Knowledge reuse for
innovation-The missing focus in knowledge management: Results of a case
analysis at the Jet Propulsion Laboratory. Proceedings of the Academy of Management 2001, Washington, D.C.
Managing for Outcomes - Strategic Governance Principles & Indicators. (2003). Retrieved 2/02/2005, from http://www.treasury.qld.gov.au/office/knowledge/docs/manage-for-outcomes/mfo-governance.pdf
Markus, M. (2001). Toward a theory of reuse: Types of knowledge reuse situations and factors in reuse success. Journal of Management Information Systems,
18(1), 57-93.
Meyer, C. (2004). How the Right Measures Help Teams Excel. Harvard Business
Review on Teams That Succeed, 27-51. Boston: Harvard Business School Press.
Milner, E. M. (2000). Managing Information and Knowledge in the Public Sector: New York: Routledge.
Nonaka, I. (1994) A dynamic theory of organisational knowledge creation. Organisation Science, 5(1).
Nonaka, I., & Takeuchi, H. (1995). The Knowledge Creating Company: How
Japanese Companies Create The Dynamics Of Innovation. New York: Oxford University Press.
Parasuraman, A., Zeithaml, V., & Berry, L. (1988). SERVQUAL: a multiple-item scale for measuring customer perceptions of service quality. Journal of
Retailing, 64(Spring), 12-40.
Pentland, B. T. (1991). Making the Right Moves: Toward a Grammar of Software
Support Hot Lines. Massachusetts Institute of Technology, Massachusett.
Pentland, B. T., & Rueter, H. H. (1994). Organizational routines as grammars of action. Administrative Science Quarterly, 39(3), 484-510.
Performance Management Framework. (2002). Retrieved 2/02/2005, from http://www.treasury.qld.gov.au/office/knowledge/docs/manage-for-outcomes/mfo-performance.pdf
Peters, T. J., & Austin, N. (1985). A Passion For Excellence : The Leadership
Difference (Fontana paperback ed.). London: Fontana/Collins.
Polanyi, M. (1967) The tacit dimension. Garden City, New York: Anchor.
Ranaweera, C., & Neely, A. (2003). Some moderating effects on the service quality-customer retention link. International Journal of Operations & Production
Management, 23(2), 230-248.
Riggs, B. & Thyfault, M. (1999) The Modern Call Center. Retrieved 6 February 2005, from http://www.informationweek.com/755/call.htm
References
132
Rogers, A., Hassell, K., Noyce, P., & Harris, J. (1998). Advice-giving in community pharmacy: variations between pharmacies in different locations. Health and
Place, 4(4), 365-373.
Rus, I., Lindvall, M., & Sinha, S. (2001). A State of the Art Report: Knowledge
Management in Software Engineering. Retrieved 6 February 2005, from http://www.dacs.dtic.mil/techs/kmse/kmse.pdf
Schefe, N., & Timbrell, G. (2004). A Tale of Two City Call Centres. Proceedings of the 15th Australasian Conference on Information Systems, Hobart TAS.
Senge, P. M. (1997). Communities of leaders and learners. Harvard Business
Review, 75(5), 30-32.
Service Level Agreement for Provision of Services. (2004). Brisbane.
Standards Australia (2003) Interim Standard for Knowledge Management (AS 5037(Int)-2003), Retrieved 10 October 2004, from Standards Australia Online database.
Teicher, J., Hughes, O., & Dow, N. (2002). E-government: a new route to public sector quality. Managing Service Quality, 12(6), 384-393.
Thompson, M., & Walsham, G. (2004). Placing knowledge management in context. J Management Studies, 41(5), 725-747.
Thompson, P., Warhurst, C., & Callaghan, G. (2001). Ignorant theory and knowledgeable workers: Interrogating the connections between knowledge, skills and services. Journal of Management Studies, 38(7), 923-942.
Timbrell, G., Andrews, N., & Gable, G. (2001). Impediments to Inter-firm transfer of best practice: in an Enterprise Systems Context. Australian Journal of
Information Systems, Special Edition, 116-125.
Timbrell, G., Delaney, P., Chan, T., Yue, A., & Gable, G. (2005a). A Structurationist
Review of Knowledge Management Theories. Proceedings of the International Conference on Information Systems, Dec 11-14, Las Vegas, USA.
Timbrell, G., Koller, S., Schefe, N., & Lindstaedt, S. (2005b). A knowledge process infrastructure hierarchy model for call-centre processes. Journal of Universal
Computer Science, 11(4), pp. 546-564.
Timbrell, G., Nelson, K., & Jewels, T. (2003). Knowledge Reuse in an Aplication Service Provider. In Knowledge Management: Current Issues and
Challenges. Hershey PA: IDEA Publishing.
Timbrell, G., & Shepperd, B. (2002). Improving Advice and Support Services using
Applied Knowledge Management Strategies. Proceedings of the Enabling the Information Future Conference, 16-17 October, 2002, Brisbane.
Todd, R., & Southon, G. (2001). Educating for knowledge management: Perceptions of library and information professionals. Australian Library Journal, 50(4).
Toulmin, S. (1958) The Uses of Argument. Cambridge University Press, Cambridge.
Tsoukas, H., & Vladimirou, E. (2001). What is organizational knowledge? Journal of
Management Studies, 38(7), 973-993.
References
133
Tuten, T. L., & Neidermeyer, P. E. (2004). Performance, satisfaction and turnover in call centers; The effects of stress and optimism. Journal of Business
Research, 57(1), 26-34.
Urch Druskat, V., & Wolff, S. (2004). Building the emotional intelligence of teams. In Harvard Business Review on Teams That Succeed (pp. 27-51). Boston: Harvard Business School Press.
Van de Walle, S., & Bouckaert , G. (2003). Public service performance and trust in government: The problem of causality. International Journal of Public
Administration, 26(8-9), 891-913.
von Krogh, G., Nonaka, I., & Aben, M. (2001) Making the most of your company’s knowledge: A strategic framework. Long Range Planning, 34, 421-439
Weick, K. (1969). The Social Psychology of Organizing. Reading, Mass: Addison-Wesley.
Weick, K. (1995). Sensemaking in Organizations. Thousand Oaks: Sage Publications.
Wiig, K. (1999). Comprehensive Knowledge Management - Working Paper. Retrieved 6 February, 2005, from http://www.knowledgeresearch.com/downloads/compreh_km.pdf
Wilde, C. (2000). Web Call Centers Benefit Customers and Businesses - Linking web
sites and call centers can improve service and customer retention. Retrieved 23 March, 2004, from http://www.informationweek.com/809/calls.htm
Yin, R. K. (1994). Case Study Research, Design and Methods (2nd ed. Vol. 5). Newbury Park: Sage Publications.
Zack, M. H. (1999). Developing a knowledge strategy. California Management
Review, 41(3), 125-145.
Zack, M. H. (2001). If Managing Knowledge is the Solution, then What's the Problem? In Y. Malhotra (Ed.), Knowledge management and business model
innovation (pp. 16-36). Hershey, PA: Idea Group Publishing.
Appendices
134
Appendices
Appendix 1 - General Information Processes in CC-B
GENERAL INFORMATION PROCESS
RESPONSIBILITY PREPARATION NEGOTIATION AND
PERFORMANCE
ACCEPTANCE
CLIENT
CALL CENTRE
BUSINESSAREA
Client calls
Call Centre
Does
operator
have enough
information toanswer the
inquiry
Log call
details
Determine General
Information is
required
Identify the correct
contact number
Explain call details
to Business Area
and transfer call
Identify the
location of the
information
Answer inquiry
using all information
available in CCIS
Cut and paste
call details into
email and send
Business Area
provides general
information
Random call
back to verify
client
satisfaction
Provide feedback
to Call Centre
END
N
Y
Identify the
subject nature of
the call
Is the
Business Area
available
Business Area
documents inquiry
for CCIS
N
Y
Information
provided to
client
Log call
details
Appen
dic
es
135
A`
pp
end
ix 2
- A
ver
age
Ad
vic
e L
ine
Wait
Tim
e – C
C-A
38
Bri
sb
an
e C
usto
mer
Serv
ice C
en
tre
0
20
40
60
80
100
120
140
Jul-04
Aug-0
4S
ep-0
4O
ct-
04
Nov-
04
Dec-0
4Ja
n-0
5F
eb-0
5M
ar-
05
Apr-
05
May-0
5Ju
n-0
5
Seconds
BC
SC
(R
EV
S)
BC
SC
(C
ON
S)
BC
SC
(B
US
)B
CS
C A
VG
MP
S S
TD
BC
SC
(B
US
)B
CS
C A
VG
YTD
38 A
bb
revia
tio
ns
are
no
t ex
pan
ded
sin
ce t
hey
sim
ply
rep
rese
nt
dif
feri
ng f
unct
ional
gro
up
s w
ithin
the
agen
cy.
MP
S S
TD
– t
he
org
anis
atio
nal
tar
get
AV
G Y
TD
- t
he
aver
age
for
yea
r-to
-
dat
e per
form
ance
.
135
136