Theoretical implications

The academics have discussed service recovery from various views, yet this research is mainly based on the service recovery process model introduced by Gonzalez et al. (2005) which is supplemented with the findings from existing service recovery literature especially in the context of social media. Moreover, the research adopts the theories related to service quality and failure (e.g. Grönroos 2007, 74), customer complaints (e.g. Tax et al. 1998), customer complaining behavior on social media (e.g. Grégoire et al. 2015), eWOM (e.g.

Balaji et al. 2016) and service recovery on both offline (e.g. Gonzalez et al. 2005) and online

(Einwiller & Stelein 2015) environments. As this research not only aims to increase the knowledge on how to manage service recovery on social media but also to compare previous literature to empirical findings, this chapter discusses whether findings are supported by the literature and how new findings fit in. However, as customer behavior, service failure and service recovery are relatively old subjects of study, this research did not aim to provide significant contributions on the literature, yet both customer behavior and service recovery on social media are less studied phenomenon providing the possibility for theoretical contributions.

This study integrates service failure, service quality, customer complaining behavior and service recovery literature by examining concepts in the context of social media from service provider’s point of view. In addition, this research provides evidence that in order to recover customers successfully on social media managers have to adopt new practices due public nature of social media, even though same strategies are used on both offline and online.

Thus, this research not only enrichens the current knowledge on both customer complaining behavior and managerial practices on social sites but also insights on the behavior of social media audience (i.e. third parties).

The main theoretical implication on this study is a process model for social media service recovery for especially larger companies which is built on the findings of Gonzalez et al.

(2005) who created a recovery process that first analyses the failure followed by recovery efforts. As most of the recovery practices could not be identified due confidentiality issues there is a limited research among the subject, which does not provide new information on the processes introduced in academic literature. However, this study supports to some degree procedural service recovery attributes including fast response (e.g. Boshoff 1997;

Gonzalez et al. 2005), thanking for pointing out the failure (Blodgett et al. 1993), apologizing (Gonzalez et al. 2005; Roschk & Kaiser 2013) and paraphrasing the original issue (Balaji et al. 2016). In other words, the importance of these attributes is identified, yet their practical use varies due to employee’s personal skills and attitudes.

Nonetheless, the updated framework is also adoptable to offline environment as it seems that previous literature provides guidelines mainly for smaller businesses without significant importance on the amount of incoming messages, the extent of internal processes or acknowledging the possibility to employ internal expertise interdisciplinary (e.g. bi-level problem solving or employing communications department) not only to increase efficiency but also to relieve customer service’s pressure in frontline. This perspective provides a

slightly differing view on employee empowerment (e.g. Hocutt & Stone 1998, 128; Smith et al. 2010, 448) as based on empirical research employees should be empowered yet more complex issues are forwarded to customer claim handling which may be due to the size of the company and the amount of incoming customer claims. Moreover, service recovery culture is placed into the process model to describe its impact on both managerial and employee practices within the organization as it is the primus motor for successful service recovery practices. Surprisingly, quality management did not have that significant of a role on the research. However, it is included to the daily processes of the company, yet was not highlighted during the interviews even though for example ensuring internal quality is part of common practices. It is possible that quality dimension would have appeared more clearly if more interviews were conducted.

Second contribution to the existing literature is the extended knowledge on customer complaining behavior on social media, as Balaji et al. (2016, 537) concluded previous literature does not offer a holistic perspective on customer complaining behavior on social network sites. The base for social media complaining behavior is built on framework of Tax et al. (1998, 61), which is modified to meet actions of social customers as social media allows not only to complain directly to the company but also enables third-party actions.

This study indicates that customers who have adopted social media into their lifestyle have not only embraced norms of the society but also the line between private and public conversations is blurred due active involvement in the community as conversations are vivid even when discussion takes place using their real names providing an opposite view on research of Fu et al. (2015) who stated that customers are concerned about their reputation on social media which may affect their voice online. Moreover, even though various researchers (e.g. Bell & Zemke 1987, 32; Hart et al. 1990, 151; Balaji et al. 2015, 647) have concluded that complaining customers are the exception as dissatisfied customers usually remain silent, may be valid on offline environments but there clearly is a customer type who wishes to complain publicly on social media.

Even though customer complaints are studied by various researchers including comparing objective and subjective ingredients of complaints (Halstead et al. 1996), studying e-complaining and company responses (Harrison-Walker et al. 2001), investigating relationship between pre-recovery emotions, severity of the service failure and post-recovery emotions (Ozgen & Kurt 2012), examining the effects of social influence on virtual presence (Schaefers & Shamari 2015) and exploring the role of customer emotions on service failures and recovery implementation (Smith & Bolton 2002), none of the studies

identified customer complaining styles. To address this gap, this research provides new determinants of complaining communication as complainer either asks a question or places a statement when voicing the occurrence of service failure on social media, whereas statements are mostly used when a failure has occurred more than once, customer has faced a process failure or double deviation needing immediate recovery. The view is partly similar to the research of Fan & Niu (2016, 1032) where they concluded that if the failure is of a great magnitude, slow responses on social media have more negative effect on customer satisfaction compared to the less severe ones. Thus, by observing customer complaining style may help to evaluate the magnitude of service failures.

Triggers for social media complaining vary between original complainers and third parties, as personal and company benefits are identified among original complainers whereas third parties are looking for personal and social benefits, yet all the aspects previous literature (Amine 1998; Harrison-Walker 2001; Hennig-Thurau et al. 2004; Ward & Ostrom 2006;

Zaugg 2006; Rimé 2009; Grégoire et al. 2010; Chamboux et al. 2012; Bronner & de Hoog 2011; Verhagen et al. 2013; Balaji et al. 2015; Xu et al. 2015) suggested are identified on this study. Moreover, third parties are mainly attracted on writing styles that are either story-like or ironic, providing new knowledge on literature as none of the studies used in this research has examined the subject earlier. However, it should be noted that both complaining styles and triggers for either complaining or joining the conversation represented relatively small sample from both Facebook and Twitter users making the results possibly non-generalized.

The elements of successful service recoveries are studied widely, yet information is disassembled providing the third contribution on this study. Both existing literature on service recovery and results of this study indicate that service recovery culture is the most important factor to recover customers successfully. Even though earlier literature is focused on managerial skills and knowledge (e.g. Maxham & Netemeyer 2003, 46; Gonzalez et al.

2010, 224), service recovery strategies and supporting employees (e.g. Robinson 2011, 96), yet new components identified on this study relate to internal support including the importance of peer support and coaching, bi-level problem solving and interdisciplinary information exchange along with dialogue within the organization.

Alongside service recovery culture, service recovery implementation is relatively widely studied area in offline context (e.g. Kelley et al. 1993; McCollough et al. 2000; Andreassen 2001; DeWitt & Brady 2003; Guo et al. 2016), yet there is very little research on service

recovery implementation on social media. The fourth contribution applies to practical performance on social media by creating three phases to communication including ex ante, during action and ex post phases of reactive social media service recovery which is a synthesis of previous literature and empirical research. To add new knowledge, empirical results indicate that it is recommendable to create social media ground rules to provide guidelines for social media communication. Social media could be employed to both external and internal recoveries and the existence of service designers on frontline may prevent future failures.

Moreover, earlier literature does not recognize either the assistance of communications department or customer education on social media. They both seem to provide effective results by benefiting both customer and company with more comprehensive reply for the customer also supporting the company image providing a new aspect on social media service recovery literature. However, this study supports conclusion of Einwiller & Steinlen (2015) as it seems that interaction with stakeholders on social media is not fully adopted due inefficient service recovery implementation on social media.