Research design

This qualitative multiple-case study focuses on business model innovation of start-ups, and espe-cially the impact of external stakeholders on the BMI process. Case studies are appropriate to illustrate the real-life situation in a strictly limited context (Farquhar, p. 6). The main objective is to understand the success factors when innovating a circular economy business model. Research on BMI emphasizes cooperation to initiate, design and implement new business models with eco-system participants, and that is why it works as a framework for theory development. It concen-trates on explaining, running and developing business (Spieth et al. 2014). Research on BMI does not dive too deeply into the elements of business models, and instead centres on activities to develop the new business model. The main research questions of this study are:

 How to innovate successful circular economy-driven business models?

 How to embrace ecosystem collaboration in the business model innovation process of circular economy start-ups?

The first question addresses the process of innovating new business models. For this purpose, the dynamic capabilities framework was used to analyse case companies. It emphasizes issues which are important to ecosystem management. (Mezger 2014; Teece 2007) The main points consider external resources and dynamic capabilities which benefit start-up companies to configure their business models. The second question complements the main objective to understand external collaboration in the BMI process, and takes a broader perspective on ecosystems. The research questions support a case-study design, which usually answers ‘why’ and ‘how’ questions, and can be exploratory, explanatory, evaluative or descriptive. (Rowley 2002, p. 16) These questions are clearly exploratory, as this is a new field and there is no research about these kinds of companies.

Also, the BMI literature emphasizes that processes and elements most often employ an explora-tive approach (Schneider & Spieth 2013).

Figure 6: The research ‘onion’ (modified from Saunders et al. 2016, p. 124).

The methodology of this research is summarized in Figure 6. The outermost layer of the research

‘onion’ indicates which research philosophy is the most important in this thesis. This research adopts interpretivism, but it has other philosophical features as well, such as features of pragma-tism. For example, pragmatism features in the way the data are selected – by considering all kinds of available data which answer the research questions in the best possible manner (Denscombe 2010). Also, case sampling has pragmatic elements, because start-up companies were selected because of their potential to illustrate this innovation process. The interpretive approach considers reality as a complex and socially constructed environment through multiple perspectives. It acknowledges that theories are too simplistic, and focuses on different points of view and inter-pretations with in-depth investigations to gain access to sufficient knowledge. Researchers should be reflexive and contribute insightful interpretations, which can be subjective perceptions. The main objective is to create a new and profound understanding based on many social contexts and perspectives. (Saunders et al. 2016, p. 136–141)

The main issues of this research (the BMI process and ecosystem interaction) are based on earlier literature, and earlier studies function as deductions to the whole thesis. This approach is common in multiple-case studies, where the main themes and questions are predefined. (Eriksson & Ko-valainen 2008, p. 123) Of course, the theory development is not purely deductive, because it is formed iteratively from the literature and empirical material. The interviews started before the theory development was completed, to get a better understanding of the research objectives. Com-bining inductive and deductive approaches can be useful, and it is a common way to generate business research. (Saunders et al. 2016, p. 148–149) In this study, the abductive approach is dominant, as the tool for the analysis was based on the literature review, but it was completed by using empirical data.

This research follows a multiple-case strategy, and uses start-up companies as cases. Qualitative research is the only option to study this kind of new and broad theme. A case study strategy is appropriate to generate an in-depth understanding of the new and complex phenomenon (Saun-ders et al. 2016, p. 184–185). Cases comprehensively illustrate the real-life situation in the strictly limited context and with a strictly limited phenomenon (Farquhar, p. 6). The context of this re-search is strictly defined, as it considers circular economy start-ups in the construction sector in Finland. Cases are start-up companies, which comprehensively represent this context, but there are also other organizations which work as mini-cases and they complement the ecosystem re-view. The default of this research strategy is that the results will somehow be replicable between different cases. Multiple cases may provide more evidence by replication, instead of arguing only valid sampling logic. The robustness of the multiple-case design can be improved by replicating the results of earlier studies cited in the literature. (Zainal 2007, p. 2)

All cases in this research have a clear purpose – to illustrate how their business models were innovated through external collaboration with ecosystem actors. Purposive sampling is appropri-ate when there are only a few case samples, but they should be informative and convenient to answer the research questions (Saunders et al. 2016, p. 301). Information-rich cases enable learn-ing from issues which are of central importance to the study (Patton 1990, p. 169). The point of sampling is to find common patterns of different business model innovation processes, following the maximum variation sampling method. It is an appropriate sampling method to illustrate sim-ilar patterns in a variety of case companies. Maximum variation sampling can be useful with small samples, because then all emerging common patterns have greater significance. A maximum var-iation sample can be compiled by recognizing characteristics/criteria for sampling. This sampling strategy may provide unique and detailed data, as well as generalizable outcomes of common patterns. (Patton 1990, p. 172)

These case companies vary greatly and represent totally different kinds of business model inno-vations in the circular economy field in the construction sector. The type of business model was one of the selection criteria (detailed case descriptions in Chapter 6). They followed quite differ-ent strategies for their innovations, which influenced the level of external collaboration in the innovation process. There is also variation in how these companies exploit digitalization, how they take advantage of and contribute to the circular economy, how ambitious they are in building scalable and international businesses, and how financially successful they are. Case companies may provide a comprehensive and broad picture of circular economy start-ups in the Finnish con-struction industry. Availability and limited resources necessarily guided the case sampling.

The time horizon of this research is cross-sectional, because it focuses on studying a phenomenon at a particular time (Saunders et al. 2016, p. 200). A longitudinal study would not have been appropriate for this thesis work, because of limited time resources. Even when the main focus is

on the development ‘process’, the capabilities and resources (external and internal) which led to business model configurations are much more relevant than events and happenings at a particular point in time.