Data gathering

In case research the underlying principle of data gathering is the use and combination of different methods to study the selected phenomena (Voss, Tsikriktsis and Frohlich, 2002). The methods used in this study include interviews, expert discussions, presenta-tions, secondary data and internet sources. The use of multiple data sources on the same phenomenon increases the reliability of the data (Voss, Tsikriktsis and Frohlich, 2002). Majority of the data for the study was collected through interviews and this primary data was supplemented with secondary data sources, such as the companies’ and or-ganizations’ external material, including brochures, internet sources and recorded presentations as well as news articles. Especially in ecosystem research, interviewing is a recommended approach (Phillips and Ritala, 2019) A possible weakness of utilizing interviews as the primary data source is that as the interviews are verbal reports which might be affected with challenges such as bias, inaccurate articulation or defective memory (Yin, 2003 p.92). To address these issues, data gathered from interviews was supplemented and backed up with secondary data. Even though data triangulation was used, it is important to acknowledge that the secondary data could also be biased simi-larly to the primary data.

The primary data was gathered through interviews mostly between November 2019-Jan-uary 2020 and an additional two interviews were conducted in June 2020 to achieve a balanced coverage of both cases. A non-probability sampling approach was chosen for the data gathering through interviews to ensure that the most relevant persons in the studied ecosystems would be reached. The interviews were conducted as semi-struc-tured interviews. Semi-strucsemi-struc-tured interviews are likely to be the most advantageous ap-proach when collecting data on complex or open ended questions and in research situ-ations where the order and logic of questioning needs to be varied (Saunders, Lewis and

Thornhill, 2016). These circumstances are inherent in this study as the aim is to under-stand the relations of various actors in an ecosystem working towards a shared goal.

Interviews enable gaining deeper understanding on the research phenomena as inter-views allow asking follow-up questions and building on previously asked questions.

The interviewees were CEO or managerial level in their organizations in addition to rep-resenting their organization in the studied ecosystems. The interview audios were rec-orded with the interviewees permission and notes were taken during the interview, as recommended for researchers (Saunders, Lewis and Thornhill, 2016). The interview du-rations varied from approximately from 30 minutes to 90 minutes mainly due to the inter-viewees’ personal differences in self-expression. The interview records were transcribed by external transcribers. Irrelevant small talk and warm-up discussions not related to research were excluded from the exact interview durations mentioned in table 5 and table 6 as well as from the transcriptions. Majority of the interviews were conducted at the interviewees’ organization’s headquarters, while some of the interviews were con-ducted as remote interviews through digital systems (Skype and TEAMS). No significant difference between the quality of the data in face-to-face versus remote interviews was perceived. Some of the remote interviews were a little shorter in duration due to less small talk and disruptions between the questions and answers.

When conducting interviews it has to be kept in mind that the manner in which the inter-action is done and questions are asked will impact the collected data (Saunders, Lewis and Thornhill, 2016). Therefore, when conducting interviews for this study the objective was to maintain a similar relaxed and positive yet professional manner in all interviews to mitigate the influence of the interviewer’s behavior on the collected data. In addition, interviewer bias can cause data quality issues in semi-structured interviews (Saunders, Lewis and Thornhill, 2016), which underlines the importance of concentrating on objec-tivity when interviewing and interpreting responses.

The preliminary questions were given in advance for the interviewees to allow the inter-viewees to get acquainted with the interview topic. The structure of the interviews was similar in all interviews, but the questions had to be modified to suit the interviewees role in the ecosystem. Due to the inherent challenges of ecosystem research the exactly same interview can’t be used in different interviews as interviewees have varying roles in the ecosystem. Furthermore, the interviews for both cases were dissimilar due to the underlying differences across the cases regarding, for example, the value proposition

and objectives, which required case specific questions. Appendix B represents the struc-ture and question examples of the interviews for both cases in addition to the questions for the initial expert interview.

In ecosystem research it is advised to use multiple sources to understand relationships (Phillips and Ritala, 2019) as was done in this study when interviewing multiple actors in the ecosystems to understand the perspectives of different ecosystem participants.

Through interviewing it was possible to establish personal contacts (Saunders, Lewis and Thornhill, 2016) which gained access to relevant contacts in the studied ecosystems as interviewees were asked to recommend their contacts in the ecosystem for further interviews. In ecosystem research identifying potential interviewees through expert dis-cussions and then snowballing interviews is an example of well-suited research ap-proach (Phillips and Ritala, 2019), which was used in this study. Table 5 present an overview of the data sources utilized for both cases.

Table 5. Overview of the data sources for both cases.

Data type CASE: Circular Economy Ser-vice Platform

CASE: Beverage Package Recycling System

Interviews with key actors in the ecosys-tem

2019-2020: 7 interviews 2019-2020: 6 interviews

Presentation materi-als; company data

Presentations (3) Presentations (2)

Media data, market-ing materials, e.g.

brochures

Brochures and booklets (1), webpages (5)

Brochures and booklets (3), webpages (6), news articles (4)

Other Report (1), research article (1)

Interviews

Interviews conducted for case Circular Economy Service Platform are presented in table 6. The interviews took on average half an hour to one hour. The duration of the interview

was affected for example on the extent and comprehensivity of the interviewee’s answers as well as the personal communication style of the interviewee.

Table 6. Interview data on case CEP.

INTER-VIEWEE

DATE ORGANIZATION, LEVEL OF INTER-VIEWEE

THE ROLE OF THE OR-GANIZATION IN THE ECO-SYSTEM

DU- RA-TION

I1 12.11.2019 CLIC Innovation, Head of CE

Coordination of the project 50min

I2 26.11.2019 The Chemical Indus-try Federation of Fin-land, Chief Advisor

Industry association 73min

I3 26.11.2019 UPM, Director Forestry company and pro-ject leader

98min

I4 3.12.2019 Evianet Solutions, CEO

Technology provider 33min

I5 13.12.2019 Metsä Group, Man-ager

Forestry company 37min

I6 8.1.2020 Stora Enso, Man-ager

Forestry company 20min

I7 8.1.2020 LUT, Associate Pro-fessor

Knowledge provider 49min

To achieve deep understanding on the relationships and the alignment in the ecosystem formed for Circular Economy Service Platform development, representatives from multi-ple key organizations in the ecosystem were interviewed. The objective was to acquire data from different viewpoints in the ecosystem. The interviewees chosen from the eco-system’s organizations were the persons who were in charge of the Circular Economy

Service Platform project in their own organizations. The interview structure and question examples for the case CEP interviewees can be found in appendix B.

Table 7 presents the interviews conducted for the beverage package recycling system case. The durations of the interviews were similar to the interviews done for the case Circular Economy Service Platform. Two final interviews for the case were conducted in June 2020 due to issues with reaching the interviewees because of the global situation.

As the time frame of the beverage package recycling system is substantial, a few months between the interviews doesn’t affect the quality of the data gathered through interviews.

Table 7. Interview data on case beverage package recycling system.

I8 & I9 12.11.2019 Motiva, Senior Expert

& Senior Expert

I10 28.11.2019 Suomen Palautus-pakkaus, CEO

Recycling system admin-istration

57min

I11 21.1.2020 The Federation of the Brewing and Soft Drinks Industry, CEO

Industry association 29min

I12 5.6.2020 Sinebrychoff, Man-ager

Brewery and soft drinks company

49min

I13 30.6.2020 Kesko, Director Retail company 68min

The objective for selecting the interviewees was to gain understanding about the eco-system from various actors in the recycling eco-system. Interviewees included

representa-tives from an expert organization, industry associations, central organization and com-panies producing and selling beverages. The interviewees were chosen based on their presumptive knowledge and experience of the beverage package recycling system.

Long experience and involvement in the system’s operations gives valuable insights on the ecosystem. Access to the ecosystem was achieved by snowballing. Interviewees gave their recommendations and contacts for further interviews. The structure and ques-tion examples for the interviews on case beverage package recycling system are pre-sented in appendix B.