A certain number of service failures are inevitable in the hotel business. In a scenario-based experiment, using an Internet-panel sample of 166 subjects, this study finds that: (1) those guests with a long transaction history are more satisfied with a recovery initiative and more willing to return than those with a short transaction history; (2) those guests who are thanked for voicing their complaints are more satisfied with the recovery initiative and more willing to return than those who do not receive a thank you statement; and (3) receiving a thank you statement while voicing a complaint might have a greater positive impact on the satisfaction of guests with short transaction histories than on guests with long transaction histories. The managerial and research implications of these findings are discussed.
Service failures are defined as any service-related mishaps or problems [real or perceived] that transpire during a customer's experience with a firm (Maxham, 2001). In the hotel industry, even top-rate employee training and well-orchestrated policies will not yield customer experiences with zero defects (Fisk et al., 1993 and Hart et al., 1990). Unlike manufacturers that can fine-tune systems until products are of uniform quality, hotel firms cannot circumvent variations. The high human component in the hotel business is the catalyst for many service-related problems. Moreover, the service experience relies on both the employee and the customer. For instance, in terms of the employee, no amount of training can make all front desk associates homogenous in their job performance. Also, as a human, an associate's mood can often be inconsistent throughout a shift. Service failures also transpire because guests actively participate in the service delivery process (Kelley et al., 1990 and Mills, 1986) and often they possess disparate expectations. For example, two guests could sleep on similar mattresses with one guest being satisfied and the other unsatisfied.
The simultaneous production and consumption that exists in the hotel industry inhibits advanced inspection of many aspects of the guest experience and consequently is also a catalyst of service mishaps (Fisk et al., 1993, Michel, 2001 and Hess et al., 2003). In other words, many ‘moment of truths’ can only happen after the guest arrives. This simultaneous production and consumption also permits unavoidable and often unpredictable extraneous factors to intervene in the customer transaction. For instance, even the most well-trained and personable hotel reservationist cannot accurately predict the occurrence of a natural disaster.
Since preventing all service failures is not a realistic goal, hotel firms must learn to effectively react to failures when they happen. This reaction is termed a service recovery and is defined as the process by which a firm attempts to rectify a service delivery failure (Grönroos, 1988 and Kelley and Davis, 1994). Increased comprehension of recovery initiative techniques is quintessentially significant because customers are often more emotionally involved in and observant of the recovery effort than in a routine service scenario (Berry and Parasuraman, 1991 and Bitner et al., 1990). In fact, poor recovery efforts can potentially have a double-deviation effect (Bitner et al., 1990 and Hart et al., 1990). Double deviation is the term used to describe a scenario in which the recovery was so poorly executed that it actually represented a separate service failure in the mind of the consumer. Furthermore, misguided service recoveries can also spawn ‘halo’ and ‘domino’ effects. A ‘halo’ effect entails a customer having a negative impression of all interactions with the provider, and a ‘domino’ effect refers to phenomenon in which a misguided failure spurs failures in other attributes or areas of the service process (Halstead et al., 1993).
Due to the importance of delivering effective service recoveries, a sizable body of research exists on the topic. Nevertheless, a review of the literature uncovers some understudied, yet theoretically and pragmatically relevant, service recovery factors. First, some studies conducted in non-hotel-settings (e.g. Ganesan, 1994 and Hess et al., 2003) found that the length of a customer's relationship with a firm intervenes in how a customer assesses a service recovery initiative. However, a more recent study (Magnini et al., 2007), using a hotel-setting, fails to find such results. These mixed findings, coupled with the modern proliferation of hotel loyalty programs, make this issue of transaction history a pertinent topic of inquiry. Hence, the first purpose of this research is to shed further light on the effect of a customer's past transaction history on his/her service recovery evaluations.
The second motivation of this research is to empirically examine the effects of an “employee thank you statement” on customer psychology. An “employee thank you statement” occurs when a customer voices a complaint and the employee hearing the complaint thanks the customer for notifying the firm of the problem. This is an important issue to investigate for two reasons: (1) hotel employees appear to inconsistently use thank you statements; and (2) thank you statements have never before been examined in the literature; yet, this research predicts that they have a significant influence on customer service recovery evaluations.
Lastly, this research studies the interaction between a customer's past transaction history and the presence of an employee thank you statement. This interaction has never before been studied despite the fact that these two elements interact on a daily basis in the hotel business. An enhanced understanding of how these two factors interplay could be of use to researchers and practitioners alike.
Manipulation check results indicate that the scenarios were perceived as intended. That is, those subjects in the long transaction history treatment groups (mean = 6.96) recorded higher scores on the question measuring perceptions of past transaction history than those in the short transaction history treatment groups (mean = 1.51). An independent samples t-test indicates that this mean difference is statistically significant (t = 17.42; d.f. = 164; p = .000). In addition, as can be seen in Table 1, all four scenarios were perceived by the respondents as being realistic.
