Background
Study 1 captured what people say they do when involved in difficult service encounters. The study did not, however, determine whether these self-reports of service encounter behaviour are matched by reality. Our second study examined this issue by observing and analysing naturally-occurring, problematic service encounters. As with Study 1, the aim was to describe and classify behaviours within these encounters, and identify underlying needs and possible correlates of specific behaviours.
Method
Principles from two approaches, grounded theory and conversation analysis, were applied in this study. A theoretical sampling approach was used to select for analysis a sample of pre-recorded, real-life service encounters that had occurred over the telephone. Organizations from four sectors (utility, local government, accommodation, and finance) were given a briefing sheet detailing requirements for the selection of suitable telephone call interactions. With the assistance of a senior research assistant (independent of the research analysis stage), the organizations screened their databases for calls that reflected difficult encounters. A final screening resulted in the selection of 157 suitable calls, all of which involved disagreement or conflict between employee and customer.
Using NVivo 8, a comprehensive coding structure was developed in vivo from the data and then further complemented with coding categories derived from the academic literature. Once the open coding process was completed, the research team met and discussed a pattern coding process the aim of which was to bring the finely coded data together into a coherent set of propositions (Corbin and Strauss, 2008; Saldana, 2009). Using conversation analysis principles, the dyadic interactions were further coded by closely examining verbal and non-verbal elements of the two-way conversation (Silverman, 2006). We adopted several procedures to increase the credibility of the data analysis and interpretation process. Specifically, we (a) maintained a journal of critical decisions undertaken throughout the analysis phase, (b) kept a detailed codebook describing the key nodes and sub-nodes associated with the analysis, (c) wrote memos about possible theoretical links to the data and attached these to nodes, and (d) undertook group analysis sessions that involved researchers from two disciplines (psychology and marketing) and that used both ‘live’ audio material and the transcribed verbatim records.
Illustrative Findings
This rich data source generated many interesting findings. The following three are illustrative. First, when dealing with difficult (high emotion) interactions, service providers frequently adopted the role of "controller", in an apparent attempt to better manage the interaction (and the customer). However, use of controlling behaviours too early in the encounter seldom succeeded in containing the customer’s high emotion, and instead often led to an escalation of conflict and negativity. Second, while most difficult interactions were triggered by core service problems, the handling of the ensuing recovery process by the employee required emotion management prior to a focus on problem-solving or resolving "concrete" service problems. Employee validation of the customer's concerns seemed to be particular important in facilitating later progress towards problem resolution. Third, for customers, many emotionally-charged interactions stemmed from an unfulfilled need to make sense of the situation and from perceptions of low procedural fairness. In many cases, the employee failed to recognize the customer’s need to understand and make sense of the manner in which their complaint was being addressed. Rather than trying to help the customer assimilate company procedure into a broader sense - making framework, employees often merely restated rules and regulations. Thus, customers were left having heard the policies, but not having understood why these polices apply. These and other findings demonstrate the relevance to service encounter success of meeting participants' psychosocial needs for control, respect, cognition, justice, and others.
Conclusions
We concluded from this study that difficult service encounters require the management of complex human needs. The development of a "skilled provider" typology was proposed. We argue that better service encounter outcomes are likely to occur when employees (a) possess a wide repertoire of service recovery and interaction management strategies, and (b) are able to engage the correct tactic, at the correct time, based on outcomes from previous strategies and appraisals of ongoing customer needs.