Research method for the mindset analysis 2007

The purpose of this empirical investigation was to identify and analyze the ‘dominant logic’ (Prahalad

& Bettis, 1986) of the device innovation area with regards to the management mindset on innovation – here termed ‘dominant innovation logic’.

Prahalad & Bettis (1986) in their implications for further research request ‘rigor approaches’ for identifying the dominant logic of a firm, beyond simple conversations with managers. The current empirical investigation builds such method for identifying the dominant logic of an organizational unit. Therefore, the research method in itself can be said to hold a novel research contribution at methodological level. Consequently, I have included a detailed description of the entire method in Appendix A. For reasons of business confidentiality, all information after 2008 had to be excluded.

Hence, the comparative analysis of the two investigations in 2007 and 2010 could not be included.

In this section, I therefore only describe the overall research method for the mindset analysis in 2007 (for details, see Appendix A). The findings from the 2007 analysis are presented in a separate section of Chapter 4.

Sample and data

The unit of analysis was the device innovation area of Novo Nordisk A/S. A sample of 10 managers was selected as informants. These 10 were chosen in order to represent the organizational functions as broadly as possible. First of all, the VPs of the three functional areas were chosen, plus a VP from a governance unit, who was a former VP from the device area. The other 6 were department managers.

The interviews were conducted as qualitative, semi-structured in-depth interviews throughout the summer of 2007. See total list of interviews, table III-1. – All interviews were recorded and

transcribed.

The interviews were based on lists of questions, which mainly included the themes

a) industry identification; b) value proposition; c) core capabilities; d) product innovation portfolio;

e) innovation barriers and enhancers. See table III-3 for the lists of core questions.

71

 What is the value, we produce for our customers? (we = device R&D)

 Which are our most essential capabilities?

 Could we be an independent company?

 Which industry would we then become a part of?

 What are the drivers of this industry?

 Who are the jokers within this industry?

 Could you mention 3 highlights from the history of device innovation at NN?

 Are we still creating history?

 What would it take to ensure that we have success in 2015?

 What is holding back and what is advancing our device innovation?

 Our current development projects: How innovative are they?

 Which are the most crucial factors, when deciding if an idea or a project is either supported or stopped?

 What are the external blocks and dilutors? – and the internal?

 What are our fears/concerns/doubts that prevent us from giving our best?

 Do we tend to stick to what we already know? – Why (not)? (Where have we settled in our comfort zones?)

 What’s the biggest learning we have made at device R&D?

 If time: Actant modeling (subject, goal, provider, receiver, supporter, opponent)

 If we were a car, which car would it be?

Table III-3. The list of key questions for the interviews in 2007.

Summaries of the interviews were made in a format resembling ‘cognitive maps’ (Eden, 1988; Fiol &

Huff, 1992; Barr, Stimpert & Huff, 1992; Bougon, 1992), containing condensed statements clustered in themes. Short follow-up interviews were conducted to get approval from the informants of the summaries. The follow-up interviews were not recorded, but the corrections by the informants were noted and approved at the interviews. These follow-up interviews typically lasted 15-30 minutes.

Thereafter, the data analysis was based on the summaries as data input. The summaries contained 680 condensed statements in total, corresponding to 0.9 condensed statement per interview minute.

Data analysis

The condensed statements were taken from the individual cognitive maps into a spreadsheet, reformulated in 107 yet more generic terms, termed ‘constructs’, so that each reformulated construct could cover several informants’ condensed statements. At the same time, the language was changed from Danish to English.

For the analysis, the constructs were clustered in three main spheres, Foundation-Innovation-Future (FIF), which basically formed a timeline of past, present and future. The FIF structure emerged out of the data analysis. Within each sphere, the constructs were clustered in groups. See table III-4 with the generic FIF structure.

72

FOUNDATION Value Proposition

Strategy - Business Drivers - Business Model Competition

Core Capabilities – strengths and weaknesses Organization – Structure, governance Identity - Culture

INNOVATION Innovation enhancers Innovation barriers Portfolio management

The current portfolio – balance radical/incremental FUTURE

Opportunities and threats Future competencies Future organization

Table III-4. The generic FIF structure which emerged out of the data analysis.

The FIF structure was applied for the data analysis in the spreadsheet of the total constructs. The spreadsheet was used as an instrument for quantification; each construct was correlated with the data to determine exactly which of the informants who supported the construct in question in their individual statements. This way I could quantify how widespread each of the constructs was amongst the 10 informants. For traceability, the exact timestamps from the interview transcriptions of the statements in support of each construct were also noted.

The ‘dominant logic’ is a worldview shared across the management team. Arguably, idiosyncratic individual viewpoints cannot belong to the ‘dominant’ logic. Therefore, to get a clearer picture of the management thinking, all constructs with less than 5 supporters (of 10 possible) were deleted. As references for this criterion, Tyler & Gnyawali (2002) similarly assess the concept of ‘shared cognition’ in a management team such that 50% must agree in a small group; 40% in a large group (where 30 managers make a large group).

The relatively few constructs left (28 of 107) were then ‘translated’ into full sentence narratives, which concluded the dominant innovation logic of the device management team. See the result in the specific section of Chapter 4.

Validation of findings

The only validation – after the informants’ approval of the condensed statements – were two workshops for the entire device management team held in 2008, where they were presented for the results from the 2007 interviews, including the ‘dominant logic’. The research findings served as input for discussions on the identity and strategy of the device area.

73

Chapter 4: The transformations of