An Efficient Customer Service: Finding a Balance between Self-Service Technology and Inter-Personal Interaction

Abstract

Nowadays customers are often offered the possibility to interact with companies through a mix of self-service technology (SST) and inter-personal contacts. Demands of both cost-effectiveness and good relationships with customers trigger the need to find a balance between these ways to interact. The purpose of this paper is to identify requirements to be put upon a computer-based tool for generating knowledge about interactions between customer service representatives (CSR) and customers that then can be used to develop new SST or improve existing systems. The empirical base consists of recorded conversations between customers and CSR situated in a Swedish subsidiary to a multinational industrial company in the building industry. Methodologically, conversation analysis complemented with pragmatic action concepts for understanding the role of information systems, are used for analysing the conversations. The results show that it is important to consider both intra-transactional and inter-transactional characteristics of conversations in order to derive a computer-based tool.

Full text

Page 1

Ågerfalk, Lind, Jacucci (Eds., 2008) Proceedings of ALOIS, Venice, Italy
53
An Efficient Customer Service: Finding a Balance between Self-Service
Technology and Inter-Personal Interaction

Mikael Lind1, 2, Nicklas Salomonson1

1University College of Borås, School of Business and Informatics, SE-501 90 Borås, Sweden
2Jönköping International Business School, P. O 1026, S-551 11 Jönköping, Sweden
Email: Mikael.Lind@hb.se, Nicklas.Salomonson@hb.se

Abstract
Nowadays customers are often offered the possibility to interact with companies through a mix of self-service
technology (SST) and inter-personal contacts. Demands of both cost-effectiveness and good relationships with
customers trigger the need to find a balance between these ways to interact. The purpose of this paper is to
identify requirements to be put upon a computer-based tool for generating knowledge about interactions between
customer service representatives (CSR) and customers that then can be used to develop new SST or improve
existing systems. The empirical base consists of recorded conversations between customers and CSR situated in
a Swedish subsidiary to a multinational industrial company in the building industry. Methodologically,
conversation analysis complemented with pragmatic action concepts for understanding the role of information
systems, are used for analysing the conversations. The results show that it is important to consider both intra-
transactional and inter-transactional characteristics of conversations in order to derive a computer-based tool.
Keywords: Self-service technology, inter-personal interaction, customer service, conversation, conversation
analysis, information systems actability
1 Introduction
To listen to the customers is a well used phrase in the corporate world and sometimes
something that is taken for granted. Why is this so important? Is it not enough to arrange
solutions of communication where the customers can serve themselves and hope that majority
choose that? More and more companies also introduce self-service technology (SST) in the
interface between companies and customers and can save money by reducing the number of
people in the front-line. The tendency that companies are reconfiguring the way they interact
with customers is called the “the front-office revolution” by Rayport et al. (2005). An
important factor that speaks in favour of a human-to-human contact is that customers’
evaluations and experiences often are based on specific interactions – service encounters with
the companies (Bitner, 1990). A service encounter is the time-frame during which a customer
directly interacts with service providers (Surprenant & Solomon, 1987). The interaction is a
critical determinant of customers’ satisfaction with the service (Czepiel et al., 1985) and the
contact employees can be the source of differentiation and competitive advantage for
companies (Pfeffer, 1994). There are also risks with removing the human contact, for example
a lack of human involvement if an error occurs, lack of personal relationships, reduced
opportunities of additional sales (c.f. Dabholkar, 2000) and even reduced customer loyalty
(c.f. Lindberg-Repo & Grönroos, 2003). There are also advantages with increased self-
service. It can be lower costs of information registration due to a reduction of the manual
work; increased security through fewer human errors; and added value to customers when
access to information and trade increases (c.f. Lancioni et al., 2000; Porter, 2001). A balance
between SST and inter-personal contacts seem to be appropriate. As Rayport et al. (2005, p.
68) mention: “The strategic question facing companies is how to effectively distribute
relationship building roles between humans and machines in a way that capitalizes on the
strengths of each.”

How can companies find this balance between contacts that could be handled through
technology and contacts that should be handled on an inter-personal level? One way is to ask

Page 2

Ågerfalk, Lind, Jacucci (Eds., 2008) Proceedings of ALOIS, Venice, Italy
54
customers how they perceive the service and what kind of interaction they want. Research
about how customers perceive the interaction, the service and the company is quite common
(c.f. Brown et al., 1994; Svensson, 2006). In this study we suggest another approach and that
is to study the actual interaction patterns and especially those in the interface between
customer service representatives (CSR) and customers. Salomonson and Lind (2006a) argue
that a detailed study of what actually occurs in these conversations is a way to increase the
understanding of which interactions that should remain personalised and which interactions
that could be transferred to self-service. Such a detailed study is however complicated to
perform resource-efficiently in a company. Even so CSR engage in conversations with
customers on a daily basis and have a unique opportunity to receive all different kind of
signals during these interactions. Through the direct interaction the CSR also have an
opportunity to adapt to what the customers say and make clarifications to avoid
misunderstandings. Over time knowledge about the customers is generated. This knowledge
needs to be transferred to the organisation in order make improvements that lead to more
satisfied customers that return for further purchases. The introduction of SST’s also indicate
that there is an understanding in companies that customers should have the opportunity to
serve themselves through some system that the company supplies. The knowledge about what
occurs in interactions between CSR and customers could also help these companies to find a
balance between SST and inter-personal interaction and improve existing systems.

In this paper we suggest some requirements for a tool that can generate knowledge about
interactions between CSR and the customers. The rational behind such tool is to have a
resource-efficient way for deriving knowledge of contemporary interactions. The tool
therefore needs to be quite simple to use in order to generate knowledge in a cost-effective
way. It should reveal important characteristics of ongoing conversations between CSR and
customers. Some companies use CRM-systems to keep track of previous encounters with
customers and thereby identify what customers need and want. The purpose of this paper is
instead to identify requirements to be put upon a computer-based tool for generating
knowledge about interactions between CSR and customers that then can be used to develop
new SST or improve existing systems. This paper builds upon insights generated by
Salomonson and Lind (2006b) in which the potential role of an IT-system as an instrument for
supporting the interaction between two parties in ongoing conversations was determined. In
this paper we take this analysis a step forward by suggesting requirements to be put upon such
a tool for capturing and disseminating knowledge about contemporary conversations. This
research is driven from the question of how to derive and categorise knowledge about
interaction between CRS and customers by using a computer-based tool.

Methodologically, conversation analysis (CA) complemented with action concepts for
understanding the role of information systems, are used (c.f. also Salomonson & Lind, 2006b)
for analysing the conversations. More specifically the notion of information systems actability
(ISAT) is used for understanding which potential that a computerised information system
could have in making interactions more efficient. ISAT uses socio-instrumental pragmatism
(SIP) as the foundational ontology. SIP is to be interpreted as the chosen ontology for
understanding socio-instrumental action performed between involved actors/agents. These
concepts are important in order to understand IT-systems as action and communication
systems supporting organisational action. The term SIP/ISAT is used in order to stress the
ontological foundation for analysing conversations described by CA. Other scholars (c.f.
Goldkuhl, 2003) also acknowledge the potential in combining analytical constructs from CA
and pragmatic action theories, such as speech act theory. As empirical base, recorded
conversations between customers and CSR in a multinational industrial company in the

Page 3

Ågerfalk, Lind, Jacucci (Eds., 2008) Proceedings of ALOIS, Venice, Italy
55
building industry (BIAB), is used. The CSR are situated in a Swedish subsidiary and the
customers have long-term contracts with BIAB and usually call to place sub-orders based on
their contracts; request information; make changes or check status of on-going orders; and
sometimes complain about something. We focus on conversations that according to
Salomonson and Lind (2006a) can be classified as “typing” or “talking/typing”. These
categories and a third one classified as “talking” are described in section 2.4. The
conversations analysed are from a larger corpus of data (Salomonson, 2005) that were
recorded during a three month period. The conversations chosen for this paper are
representative for other calls in the larger of corpus where customers order products or get
information for future orders and that can be classified as “typing” or “talking/typing”. In this
paper essential action characteristics constituting conversations are focused. This means that
we delimit ourselves from financial and ROI aspects of computerisation. This is rather
identified as an issue for further studies.

At first in the paper, pragmatic action concepts for understanding IT-mediated business
interaction and conversation analysis (CA) for understanding conversations are introduced.
The notion of information systems actability (ISAT) forms together with conversation
analysis the basis for the analysis of the conversations. In the same section we also discuss
earlier studies on how IT-systems could support ongoing interactions. Then three typical
telephone conversations between CSR and customers that has occurred at BIAB are put
forward and analysed. This analysis is followed by deriving requirements of a computer-
based tool for generating knowledge about interactions based on important characteristics
from the conversations as well as theoretical constructs. The paper is concluded by
summarising findings that can be a basis for making well founded decisions about which
interactions between CSR and customers that could be transferred to a self-service system.
2 Pragmatic action concepts and conversation analysis for understanding
business interaction
2.1 Social actions in business interaction
Several scholars within the information systems field put attention towards human action
when conceptualising information systems (c.f. e.g. Goles & Hirschheim, 2000; Baskerville &
Myers, 2004). Goldkuhl (2005, p. 1) claims that “[…] technical artefacts should be a
prominent phenomena to study together with the humans surrounding them. There are things
and subjects to study. What is also important is what the human subjects do with the artefacts;
i.e. the actions of development and usage“. In an organisational setting it is thus important to
understand the notion of organisational action as the basis for deriving a notion of IT-systems.
Goldkuhl and Röstlinger (2003) have developed a notion of socio-instrumental action (SIP)
for understanding actions performed by different types of agents, as for example human
beings and IT-based artefacts.

A human being intervenes in the world, by performing actions, in order to create some
differences. An important distinction is made between the result and the effects of the action
(von Wright, 1971). The action result lies within the range of the actor and the action effects
may arise as consequences outside the control of the actor. An action is performed in the
present based on a history and aims for the future (Goldkuhl & Röstlinger, 2003). A social
action is an action oriented towards other persons (Weber, 1978), and such action can be a
communicative act, e.g. someone saying something to another person, or material (Goldkuhl,
2001; Goldkuhl & Röstlinger, 2003). Actor relationships between the intervening actor and
the recipient are established through social actions (Habermas, 1984). Humans (often

Page 4

Ågerfalk, Lind, Jacucci (Eds., 2008) Proceedings of ALOIS, Venice, Italy
56
supported by artefacts) perform action in the name of the organisation (Ahrne, 1994; Taylor,
1993). Actions are performed within the organisation – internal acts - and there are also
external acts towards other organisations (e.g. customers or suppliers). Humans act in order to
achieve ends (von Wright, 1971).

A generic model of social action including both communicative (Austin, 1962; Searle, 1969)
and material acts is presented by Goldkuhl (2001) and Goldkuhl and Röstlinger (2003). E.g.
an order from a customer to a supplier is a communicative act. The delivery of goods from the
supplier to the customer is a material act. Since these are actions directed from one actor
towards another actor they must both be considered as social actions. In this perspective we
are mainly interested in communicative dimensions (as the illocutionary force with
appurtenant propositional content according to Searle (1969)) of social actions in order to
determine action patterns. This means that we delimit ourselves from instrumental and
strategic actions according to Habermas (1984).

During a business transaction several business acts (c.f Lind & Goldkuhl, 2003) are issued
from one party to the other. These business acts are grouped together in patterns of
interaction. In the beginning of the eighties Winograd and Flores (1986) introduced the
conversation-for-action schema (CFA-schema). Building upon that schema a number of
scholars have propagated for different models for interpreting patterns of interaction during
business transactions (c.f Dietz, 1999; Medina-Mora et al., 1992). The underlying meaning of
these different transaction models is that communication implies action and that the
performance of fulfilment acts (many times material acts) should be preceded by
communicative acts. Such communicative acts are about setting the expectations, in terms of
e.g. proposals and commitments, between the two parties.
2.2 IT-mediated actions
For the performance of most actions people need instruments of different kinds. The language
is used as one instrument when performing business communicative acts. Such social and
organisational issues are handled within linguistic (Dietz & Widdershoven, 1991; Goldkuhl &
Lyytinen, 1982; Winograd & Flores, 1986) and semiotic perspectives (Stamper, 2000) for
understanding information systems. Lyytinen (1981) claims that a substantial part of a
practice is the business language, which includes vocabulary as well as rules for
communicative action. For performing material acts there is often a need for an external
instrument, which then extends the ability of an actor. Different media could be used to
support the “transfer” of the communicative act between the parties involved in the
communication. One such media is inter-organisational IT-system aimed to support (instances
of) business interaction.

According to Goldkuhl and Röstlinger (2003) a computerised system is an action system. It is
both an instrument for the performance of action and a support tool for humans to perform
their actions. Information systems should be actable. IS actability (ISAT) is defined as “an
information system’s ability to perform actions, and to permit, promote and facilitate the
performance of actions by users both through the system and based on information from the
system, in some business context” (ibid, p. 6). SIP (c.f. section 2.1) is the ontological
foundation for ISAT. The theory of information systems actability has two essential
ingredients. The first one is the distinction between three type of IS usage situations;
Interactive usage situation (where users performs actions interactively together with and
through the system), Automatic usage situations (where the system performs actions by itself
based on predefined rules), and Consequential usage situations (where users performs actions

Page 5

Ågerfalk, Lind, Jacucci (Eds., 2008) Proceedings of ALOIS, Venice, Italy
57
based on the information from the system). The second ingredient is the interpretation of an
IS as consisting of (Goldkuhl & Röstlinger, 2003) an action potential (a predefined and
regulated repertoire of actions), actions performed through and by the systems, an action
memory (a memory of earlier performed actions including prerequisites for actions), and
messages and document (where some documents are action media for user’s interactive
actions).
2.3 Conversation analysis
Since conversation analysis (CA) has been an important tool to analyse conversations in
previous studies (Salomonson & Lind 2006ab) we feel that it is essential to describe the basic
assumptions and concepts in this methodological approach. CA is an attempt to describe
people’s methods for producing orderly social interaction (Silverman, 2001). As Boden
(1994, p. 15) describes it can contribute with a detailed insight about interaction: “[…]
through a turn-by-turn analysis of organizational talk, it is possible to gain insight not only
into how everyday business gets done at the level of talk, but also the interactional and
organizational business that is accomplished through that talk”. Hutchby and Woffitt (1998, p.
13) describes CA as a systematic analysis of talk produced in everyday situations of human
interactions. CA is closely related to ethnomethodology. This is a naturalistic view with a
purpose to understand how social order is created and shaped through conversation and
interaction (Gubrium & Holstein, 1997). Garfinkel (1972) describes ethnomethodology as the
study of the methods people use in their everyday life in order to shape and maintain social
order. Reality is thus seen as something the members continuously have to recreate and
uphold which is done through social interaction. When using CA researchers do not speculate
about the conversationalists’ intentions (Heritage, 1984). As Heritage (1984, p. 243) describe
it: “[…] there is a strong bias against a priori speculation about the orientations and motives
of speakers and in favour of detailed examination of conversationalists’ actual actions.

In this paper the focus is directed towards an institutional and work related context where
customers and CSR interact. Institutional interaction normally involves the participants in
specific goal orientations which are tied to their institution relevant identity. According to
Heritage (1997) institutional interaction also has constrains on what will be treated as
allowable contributions to the business at hand and is associated with inferential frameworks
and procedures that are particular to specific institutional contexts. Drew and Sorjonen (1997,
p. 94) concludes that: “Analysing institutional dialogue involves investigating how their
orientation to and engagement in their institutional roles and identities is manifest in the
details of participants’ language, and their use of language to pursue institutional goals”.

Two important concepts in CA are turn taking and adjacency pairs. Turn-taking is how the
conversationalists take turns in talking and each turn can be seen as the basic unit in the
conversation. A turn can be very short as a yes or no or a longer statement as a full sentence.
When combined these turns lead to a conversation. Each turn is also adapted to the situation
and the specific context. Lepper (2000) describes that institutional relevance is sustained and
worked upon on a turn-by-turn basis through the recognizability of the procedurally
consequential actions of individual speakers who organize their turns within a framework of
locally relevant rules of exchange. Drew and Sorjonen (1997) describes that an institutional
turn taking system thus can be seen as a product of the participants’ orientation towards their
task related roles. An adjacency pair is two communicative actions, a first and a second, that
together represent an exchange of words produced by two speakers, for example question-
answer, offer-acceptance/rejection, invitation-acceptance/rejection and complaint-response
(e.g. Schegloff & Sacks, 1973). After an utterance, for example a question, an answer is