The traditional evaluation criteria of validity and reliability have their roots in quantitative research. In qualitative studies, especially when research relies on relativist ontology and subjectivist epistemology (see, 3.1), it is recommended to substitute validity and reliability for the evaluation criteria that better accommodate the philosophical premises of qualitative approaches (Eriksson and Kovalainen, 2008, p. 294). Lincoln & Guba (1985) replaced validity and reliability with a concept of trustworthiness, which covers four aspects: credibility, transferability, dependability, and confirmability. In the following, each of these issues is elaborated in relation to the present study.
Credibility refers to the idea of how plausible are the findings and conclusions of the research, for example, whether practitioners and readers find the study credible and does it generally make sense (Miles, et al., 2014, p. 312). In the current study, credibility was ensured especially through familiarization of the researcher(s) with the case contexts and through member checking (Stake, 1995, p. 115), thus making sure that the understanding gained by the researcher makes sense (see, 3.4). This was done in three stages during the research. First, workshops were organized with the focal firms to introduce the case context to the researcher(s) before data collection.
Second, the preliminary findings were discussed in workshops with focal firm
169
personnel. Finally, each focal firm was allowed to review and comment the final findings and conclusions. Moreover, quotations were abundantly used for presenting the findings to offer the reader a good sense of the data in the form of “thick description” to demonstrate the credibility of the data and the reasoning for meeting the research objectives (see, Geertz, 1973).
Transferability refers to the wider applicability of the findings with other contexts and connection with prior research and theories. It is not about generalizing or replicating the findings to other settings as this is not the goal of qualitative case studies; it is about showing the similarities with other research contexts (Eriksson and Kovalainen, 2008, p. 294; Miles, et al., 2014, p. 314). In this research, transferability was first addressed by describing the sample and its characteristics as well as data collection and analysis to enable comparison with other settings (see, Miles, et al., 2014, p. 314). Second, the conceptual framework of the study was derived from the prior research, and the research findings on customer involvement from other settings were reviewed in the theoretical background chapter. The present study also demonstrated that the findings have relevance to prior research (see, 5 and 6.1). Finally, the implications for further research are discussed to encourage the replication of the findings in other contexts (see, 6.4).
Dependability is concerned with the quality of the research process in terms of logic, consistency, traceability, and documentation (Eriksson and Kovalainen, 2008, p. 294; Miles, et al., 2014, p. 312). To ensure dependability and transparency, the present research has followed the guidelines typically recommended for qualitative case studies and interviews (e.g., Gibbert and Ruigrok, 2010; Piekkari, et al., 2010;
Stake, 1995; Yin, 2003). For example, all interviews were conducted face-to-face, recorded, and transcribed in verbatim; the sample and the basic characteristics of the interviewees were documented and introduced; the data collection and analysis process were described in detail; the applied coding schemes and interview outlines were presented; and illustrative quotations of the data were provided when presenting the findings.
Finally, confirmability refers to the idea that the findings and conclusions are not imagined by the researcher but have been derived from the data in a traceable manner that is understandable to others (Eriksson and Kovalainen, 2008, p. 294). In the present study, confirmability was secured by providing a detailed description of how the chain of reasoning proceeded from data collection to drawing conclusions.
Moreover, the research process was documented during all steps of the research, and the data and other documentation was stored according to good research practices to permit audit or reanalysis if needed (see, Miles, et al., 2014, p. 312).
170
It is admitted, however, that this study, as is likely the situation with any other case studies, cannot completely satisfy all trustworthiness criteria (Gibbert and Ruigrok, 2010; Miles, Huberman and Saldaña, 2014, p. 311). As pointed out by Miles et al. (2014, p. 311), the problem of quality and authenticity of the findings and conclusions is always present in qualitative research. In essence, the current study has followed the generally recommended case study practices, as discussed above, and ensured transparency in reporting to show how the research was conducted in practice in order to ensure a sufficient level of trustworthiness.
Nevertheless, some limitations in relation to the study and its research design remain. Some of these limitations could also provide a starting point for future research. First, the findings of this study are based on the exploration of three cases and as such, only represent a small sample of the vast realm of B2B services and industrial service business. For example, the study only covered some industrial service categories, and services such as administrative services, financial services, and recycling services were not included in the service portfolios studied (see, Rabetino, et al., 2015). Moreover, the selected cases were mainly centered on Finland, and despite having a few participating customer firms from other countries (Belgium, Poland, and the Netherlands), the used sample represented only a narrow snapshot of the extensive, international, and rather heterogeneous B2B service business realm.
The findings likely have some resonance with similar contexts, for example, with other industrial service firms, services, or countries. Nevertheless, the implications for the study are not claimed to hold in other settings outside the scope of the cases studied.
Second, the views of the interviewees, especially on the customer side, only represent individual views and not necessarily the views of the entire organization, even though the objective was to select highly knowledgeable interviewees (see, Eisenhardt and Graebner, 2007). The participating customer organizations were also suggested by the focal firms. Although the selected customer firms were chosen to represent typical customers of the focal firms, it is possible that the sample does not present a balanced view of the focal firms’ customers. For example, it is possible that the selected customers had a closer than average relationship with the focal firms or that their views were more positive toward the focal firm than the other customers that would have been randomly selected.
Third, the research design only enabled the exploration of service portfolio development in a cross-sectional setting. A longitudinal research approach would likely provide a more comprehensive picture, especially as “development” is fundamentally a dynamic phenomenon. Fourth, owing to the explorative nature of
171
the research, the conceptual framework focused on the rather generic categories of both customer involvement forms and offering development modes. Admittedly, a more fine-grained classification would be possible and could likely provide complementary results. As an example, there are specific customer involvement tools and methods available, especially in the B2C context (Edvardsson, et al., 2012), that were excluded from the study (see, 1.2).
Fifth, the research process was conducted following an inductive approach, and access to data was limited to the scope of the S4Fleet research program (see, 1.3).
This meant that some issues only emerged from the data during the analysis stage.
At that time, however, it was not possible to change data collection and the used interview outline, for example, by addressing some issues more in-depth with the interviewees. Therefore, such issues remain to be addressed by the future research.
For instance, the role of the key customers in encouraging service providers to develop partnership-based collaboration was one such issue that inductively emerged in the analysis stage.
Finally, this research theoretically draws from the service growth, NSD, and customer involvement literature. The extensive volume of research conducted on these research fields limits the possibilities of taking into account all earlier research that could be relevant to the present research topic. As for the service growth literature, a recent bibliometric study alone identified over 1000 related scientific articles in the field (Rabetino, et al., 2018). This posed a challenge in reviewing the earlier research and getting hold of the theoretical background in this research. In contrast, it also opens up the possibilities to alternative theoretical framing and thus warrants complementary research that can further increase the understanding of the research topic. A partly obscure and excessive theoretical background may also be a potential explanation of why earlier research has not combined service growth, NSD, and customer involvement in a similar fashion.