Department of Technology Management and Economics 
Division of Innovation Engineering and Management 
CHALMERS UNIVERSITY OF TECHNOLOGY 
Göteborg, Sverige 2015 
Report No. E 2015:006 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Entrepreneurial learning  
An Exploratory Study on the Learning Styles of 
Venture Insiders and Outsiders 
Master of Science Thesis 
in the Management and Economics of Innovation Programme 
 

MATHILDA LUNDQVIST 
SOFIE JOHANSSON-HEROU 
 
 
 



 

 

 

 



 

 

 

 

 

 

MASTER’S THESIS E 2015:006 
 
 
 
 
 
 
 
 
 
 

Entrepreneurial learning 
 

An Exploratory Study on the Learning Styles of Venture 
Insiders and Outsiders 

 
 
 

MATHILDA LUNDQVIST 
SOFIE JOHANSSON-HEROU 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Tutor, Chalmers:  Henrik Berglund 
 
 
 
 
 

Department of Technology Management and Economics 
Division of Innovation Engineering and Management 

CHALMERS UNIVERSITY OF TECHNOLOGY 

Göteborg, Sweden 2015



 

 

 
 
 
 
 
 
 
 
 
 
 
Entrepreneurial learning 
An Exploratory Study on the Learning Styles of Venture Insiders and Outsiders 
 
MATHILDA LUNDQVIST 
SOFIE JOHANSSON-HEROU 

 

 

 
©  MATHILDA LUNDQVIST & SOFIE JOHANSSON-HEROU, 2015. 

 

 
Master’s Thesis E 2015:006 
 

 

 

Department of Technology Management and Economics 

Division of Innovation Engineering and Management 

Chalmers University of Technology 

SE-412 96 Göteborg, Sweden  
Telephone: + 46 (0)31-772 1000 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Chalmers Reproservice  
Göteborg, Sweden 2015



 

 

Abstract 
Behind every venture is a visionary with an idea of a product or service to provide value to its 

users. The process of developing that idea into a product and starting a new venture exposes the 

founder to great uncertainty, not knowing exactly who will experience the greatest benefits when 

using the product. In the search for a repeatable, and scalable business model different types of 

customer interaction are suggested. According to Blank and Dorf (2012), it is crucial that 

founders themselves “get out of the building” to validate hypotheses before settling on a specific 

path.  

 

However, today many efforts are carried out in startup teams. Previous research on the 

entrepreneurial process and entrepreneurial learning mainly focuses on understanding the 

individual or an established organization (Wang & Chugh 2014), and therefore lacks appropriate 

understanding of how team efforts can be explained. 

 

This research explores the process of learning within a startup team. This is done by examining 

the ways in which the team collaborates, and how different individuals each collect, interpret, 

and make use of new information. In order to analyze the differences between insider and 

outsider collection of new information, a demarcated study was conducted.  

 

The results from the study suggest that, contrary to previous literature, startups may benefit from 

including external researchers in the early phases of the entrepreneurial process. It was shown 

that the insider used established perception of the situation and potential problems when asking 

questions and probing, therefore the approach applied by the person with domain expertise is 

likely to generate more narrow and focused information. Thus, the learnings of the expert were 

shown to generate more area specific and detailed information, as it was generated by expanding 

from the individual’s existing cognitive frames. 

 

The outsider included more explorative input, resulting in greater variance in terms of what type 

of information is gathered from each respondent. However, the lack of heuristics to quickly 

develop follow-ups meant that instead of developing insights related to the product a novice’s 

information gathering resulted in insights on widely spanning opportunities, generated from 

exploration of the particular respondent’s context. 

 

Startup team learning, seeking opportunities while allowing for radical changes to existing 

assumptions and refinement of the product, therefore includes reviewing individual as well as 

group level learning. As it is important that the startup’s learning is process of reviewing 

explorative and specific information gathered by multiple individuals, fruitful startup learning 

efforts can be explained by the extent to which diverse teams are active in the earliest stages of 

collecting information about the uncertain environment. 



 

 

Acknowledgements 

This master thesis was conducted within the master’s program Management and Economics of 

Innovation at Chalmers University of Technology. The thesis was carried out jointly with the 

founder of the startup ProposalsFactory in Kalmar, Sweden. 

 

First, we would like to thank our supervisor Henrik Berglund, associate professor at the Center 

for Business Innovation at Chalmers, for his support throughout the thesis. Without his sharing of 

ideas and knowledge, this thesis would not have been possible.  

 

Second, we would like to thank Carl Piva, the founder of ProposalsFactory, for giving us the 

opportunity to conduct the research together with his startup.  

 

Finally, we would like to thank all respondents for taking part in the interviews and teaching us 

about effective and efficient purchasing and procurement processes. 

 

Gothenburg, March 2014 

Mathilda Lundqvist & Sofie Johansson Herou 

 



 

 

 

 

 



 

 

Table of contents 

1 Introduction ............................................................................................................................................... 1 
1.1 Empirical context ............................................................................................................................................ 2 
1.2 Purpose ............................................................................................................................................................... 3 
1.3 Report structure ............................................................................................................................................... 3 

2 Theory ......................................................................................................................................................... 4 
2.1 Entrepreneurship and startup .................................................................................................................... 4 
2.2 The composition of the startup team ........................................................................................................ 5 
2.3 Collaborative research ................................................................................................................................... 6 
2.4 Exploring opportunities in uncertain environments ........................................................................... 7 
2.5 Knowledge acquisition in uncertain environments ............................................................................. 9 
2.6 Interviews as learning tools ....................................................................................................................... 10 

2.6.1 Deciding on the type of interview .................................................................................................................... 10 
2.6.2 Designing questions to explore the environment ...................................................................................... 10 

2.7 Barriers to learning ...................................................................................................................................... 12 
2.7.1 Individual level barriers to learning ................................................................................................................ 13 
2.7.2 Different types of bias ........................................................................................................................................... 13 
2.7.4 Overcoming barriers due to biases .................................................................................................................. 15 

3 Methodology ............................................................................................................................................ 17 
3.1 The initial study ............................................................................................................................................. 17 

3.1.1 Results from the initial study ............................................................................................................................ ..17 
3.2 The demarcated study .................................................................................................................................. 19 

3.2.1 Research design ....................................................................................................................................................... 19 
3.2.2 Research methods................................................................................................................................................... 20 
3.2.3 Data analysis ............................................................................................................................................................. 21 

3.3 Validity and reliability .................................................................................................................................. 22 
3.3.1 Construct validity .................................................................................................................................................... 22 
3.3.2 Internal validity ....................................................................................................................................................... 22 
3.3.3 External validity ...................................................................................................................................................... 23 
3.2.4 Reliability ................................................................................................................................................................... 23 

3.4 Self-criticism ................................................................................................................................................... 23 

4 Findings .................................................................................................................................................... 25 
4.1 Quantitative ..................................................................................................................................................... 25 

4.1.1 Results from static assessment .......................................................................................................................... 25 
4.1.2 Results from dynamic assessment ................................................................................................................... 26 

4.2 Qualitative ........................................................................................................................................................ 27 
4.2.1 Categories .................................................................................................................................................................. 27 
4.2.2 Focus ............................................................................................................................................................................ 28 
4.2.3 Input types and nuances ...................................................................................................................................... 28 
4.2.4 Presentation of qualitative static assessment .............................................................................................. 31 
4.2.5 Presentation of qualitative dynamic assessment ....................................................................................... 35 



 

 

5 Analysis ..................................................................................................................................................... 40 
5.1 Proportion of speech .................................................................................................................................... 40 
5.2 Initiating new topics during the interview ........................................................................................... 41 
5.3 Following up on responses ......................................................................................................................... 42 
5.4 Use of statements and interjections ........................................................................................................ 43 
5.5 Tendencies to formulate explorative input ........................................................................................... 43 
5.6 Tendencies to explore context versus testing assumptions ............................................................ 44 
5.7 Tendencies to adapt to the information received during interviews .......................................... 45 
5.8 Summary .......................................................................................................................................................... 47 

6 Discussion ................................................................................................................................................ 50 
6.1 Key differences between insiders and outsiders ................................................................................. 50 
6.2 Group level learning of the startup .......................................................................................................... 52 
6.3 Implications for future entrepreneurial efforts ................................................................................. 53 
6.4 Summary .......................................................................................................................................................... 54 
6.6 Further studies ............................................................................................................................................... 55 

7 Conclusion ................................................................................................................................................ 57 

References ................................................................................................................................................... 58 
 
 

Appendices 
Appendix A: Interview lists and guides 

Appendix B: Tables illustrating quantitative data analysis 

Appendix C: Tables illustrating qualitative data analysis 

 



 

 

List of figures 

Figure 1 – The roles represented in the research team 
Figure 2 - Customer Development process (Blank & Dorf, 2012) 

Figure 3 – Classification of procurement projects 

Figure 4 – Categorization of input 

Figure 5 – Insider, focus of total input 

Figure 6 – Outsider, focus of total input 

Figure 7 – Insider, focus of new question input during different phases 

Figure 8 – Outsider, focus of new question input during different phases 

Figure 9 – The most significant differences between insider and outsider interviewer 
 



 

1 
 

1 Introduction 
In this chapter the background of the study is explained by introducing the concepts related 

to founding a company such as searching for a scalable business model under uncertain 

conditions. In addition, the entrepreneurial process and the tools available to the founder 

and/or startup team during that process are presented. Finally, the purpose of the research is 

developed, and the report structure is presented.  

 

Behind every new venture is a visionary with an idea of a product or service that will be 

valuable to a consumer or an organization. Traditionally, the initiator behind the vision is 

portrayed to be a lone agent, a visionary who sets out to commercialize the discoveries and 

develop their ideas into a concept that is offered through a real company (Cooney, 2005; 

Allen, 2010, ch. 1). The opportunity-focused, innovative, and growth-oriented mindset is 

what characterizes entrepreneurship and the person driving the processes, the entrepreneur 

(Allen, 2010, ch. 1).  

 

The entrepreneurial process exposes the founder to great uncertainty, not knowing exactly 

who will perceive the greatest benefits from using the product or exactly what about the 

product will generate the most customer satisfaction (McGrath & MacMillan, 2000).  In 

order to deal with the uncertainty in the early phases of the startup history, different ways of 

collecting information can support decisions and reduce the costs involved in launching a 

new product (Lean Startup, 2014; Blank & Dorf, 2012, ch. 1; Furr & Ahlstrom, ch. 1).  

 

A startup, or a startup company, is a temporary organization designed to search for a 

repeatable, and scalable business model (Blank & Dorf, 2012, intro XVII). In the case of 

developing products that differ greatly from existing products customer interaction is 

suggested to deal with uncertainties. Such uncertainties may involve e.g. what features of the 

product will customers value the most, what market segment will have the greatest demand 

etc . In order to cope with such uncertainties, different methods are suggested e.g. market 

analysis, lead-user studies, and other types of customer interaction (von Hippel, 1986, Ulrich 

& Eppinger, 2008, Ch. 4). 

 

Many customer interaction processes to be implemented by entrepreneurs are inspired by the 

Lean Manufacturing method and includes Genchi Genbutsu which means to rely on the 

source for facts, which enables correct decision-making, building of consensus, and goal 

achievement (Wikipedia, 2014; McGrath & Macmillan, 2000, ch. 1). The processes are to be 

iterative in terms of turning customer interaction into some action or iteration of the existing 

product or startup organization. However, during the entrepreneurial process, being lean 

means both getting things done with very little resources, and about integrating learning 

processes into the startup from the very beginning (Lean Startup, 2014). If learning is 

integrated from the beginning the organization has built agility into its foundation and 

therefore is more likely to be able to adjust to necessary changes in the future.   



 

2 
 

When not knowing what customers may value it is suggested that entrepreneurs develop 

their plans in collaboration with people or organizations that they wish to have as customers. 

As the founder gets out and talks to potential customers the ability for the intended solution 

to solve important customer problems is identified, and preferably validated. This means that 

the founder can define the problem-solution fit, finding customers that place great value in 

the product or service. Thus, the entrepreneur needs to implement a structured process for 

investigating potential markets and customers segments in situations of uncertainty as well 

as tools to discard the parts that do not generate value for the customer (Lean Startup, 2014; 

McGrath & Macmillan, 2000, ch. 1). 

 

According to Blank and Dorf (2012), it is crucial that founders themselves “get out of the 

building” (Blank & Dorf, 2012, p. 24) to validate hypotheses before settling on a specific 

path. In particular the founder’s involvement and ability to embrace feedback, and react to it 

is emphasized since only the founder holds the power to make decisions to change direction 

or key business model components (Blank & Dorf, 2012, ch. 2). However, many startup 

efforts are carried out in teams. This means that the startup learning processes may be a 

product of different individuals’ interpretations of studies that are carried out. 

 

In order to determine how startup knowledge is created, it may be useful to study differences 

in individuals’ information gathering processes. If it is possible to thoroughly understand the 

learning processes of different individuals within the team, it may be possible to create more 

efficient ways to utilize that knowledge. 

 

1.1 Empirical context 

A startup assessing the opportunities for an online e-tendering solution was selected to 

investigate the matter. The startup ProposalsFactory, founded in 2012, was in the process of 

finding a scalable and profitable business model. In the beginning of 2013, ProposalsFactory 

was selected as one of ten startups to join the development program Born Global at 

Chalmers Business Innovation. The program supports Swedish startups with estimated big 

growth potential to create scalable business models following the Customer Development 

process by Steve Blank.  

 

The founder’s vision was expressed as:  

 

“To improve the way in which firms go about procurement, in order to ensure that the results 

of procurement processes are improved, and that costs for procuring are reduced.” 

  

The idea behind the solution is the outcome of practical experience from tendering processes 

where the founder identified a need for procurement process support within procuring 

organizations to lower the total cost of ownership and increase return on investment. In 

addition, the CEO experienced as a sales manager that organizations purchase inferior 

products as a result of subtle bribes rather than superior match against vendor requirements, 

which would be eliminated by the use of the product that the founder is developing. 



 

3 
 

 

The solution aims to improve the tender process lifecycle by assigning roles and 

responsibilities and inviting suppliers to take part in the bid by responding to requirements 

straight into the tool. During the summer of 2013, the startup began implementing the 

Customer Development methodology in their search for a scalable business model (Blank & 

Dorf 2012). Thus, the startup was in the process of finding a scalable and profitable business 

model. 

 

The researchers had limited experience in the field while the founder had extensive prior 

knowledge gathered from several years as a salesperson in a large multinational firm. The 

novices, on the other hand, gained their understanding from discussions with the founder as 

well as from initial scoping interviews. The figure below gives an overview of the different 

roles represented in the research team: 

 

 

 

 Figure 1 – The roles represented in the research team 

 

The researchers are experienced in market analysis and business research. These two are 

considered outsiders to the startup team but during the initial phase they developed and 

shared their learnings with the founder resulting in a shared understanding of the 

possibilities of the founder’s product vision. In addition, one of the outsiders had practical 

experience of purchasing from working within a purchasing department at a large 

multinational company. 

 

1.2 Purpose 

The purpose of the study is to examine how domain insiders and domain outsiders differ in 

their learning styles. The ways in which the individuals collect, interpret, and make use of 

new information are examined in order to understand how learnings of a startup team 

consisting of domain insider and domain outsider are generated.  

 

1.3 Report structure 

Literature related to entrepreneurship and learning processes is presented in order to develop 

a theoretical explanation to startup learning. Thereafter, the methodology chapter describes 

the initial exploratory study aiming to narrowing the research scope. In addition, the method 

used during the demarcated study is explained. Following the methodology chapter, the key 

findings from the empirical study are shown. The differences between the individuals that 

collected knowledge are then presented and analyzed with support from the theoretical 

findings. Finally, the findings are discussed and conclusions regarding learning styles are 

drawn to show how the objective of the research is fulfilled. 

Founder Researcher 

Insider Outsider 

Expert Novice 



 

4 
 

2 Theory 

This chapter identifies relevant theoretical frameworks which are to be used to understand 

the differences in logic behind the results presented by the domain novice and the domain 

expert. First, previous research and literature on startups and entrepreneurial processes are 

presented, followed by dynamics of startup groups and research in teams. Thereafter, the 

processes behind fruitful entrepreneurial development processes and different types of 

learning are discussed in order to understand how and why entrepreneurs must establish a 

learning culture. It is however difficult to learn in an objective manner, and it is important to 

understand the inhibitors to objective learning. Therefore, potential barriers to learning and 

information gathering are reviewed. 

 

2.1 Entrepreneurship and startup  

A startup is a temporary organization designed to search for a repeatable, and scalable 

business model (Blank & Dorf, 2012, intro XVII). The initiator behind the startup is an 

entrepreneur with a vision to create something new that people and/or organizations will 

value. It is important to understand the differences between how to approach organizational 

planning for a new venture and approaches that are appropriate within existing line of 

business. Customer interaction is suggested when developing new products that differ 

greatly from existing products since it is uncertain what product features will be valued the 

highest and what customers will be willing to pay the most (McGrath & MacMillan, 2000, 

preface; Blank & Dorf, 2012, ch. 1; Furr & Ahlstrom, 2011, ch. 1). 

 

The founder’s passion for the vision is what supports entrepreneurs in times of resource 

constraint and uncertainty. The opportunity-focused, innovative, and growth-oriented 

mindset is what characterizes entrepreneurship and the person driving the processes, the 

entrepreneur (Allen, 2010, ch. 1). Traits such as opportunity-focus, and perseverance are 

therefore important during the discovery and development of opportunities (Sardana & 

Scott-Kemmis, 2010). However, the same passion and determination that characterizes 

entrepreneurs may be what causes the initiative to fail (Furr & Ahlstrom, 2011, ch. 1).  

 

The individual entrepreneur’s motivation to solve customers’ problems is an important 

contributor to the development of the initial idea as it results in adoption of certain 

mechanisms to acquire new knowledge and stimulate learning (Young & Sexton, 2003). 

Examples of such important mechanisms are formal education and training, experience, and 

vicarious experience. The entrepreneur often acquires new knowledge from direct 

experience and from observing actions and behavior of others (Young & Sexton, 2003; 

Sardana & Scott-Kemmis, 2010).  

 

The entrepreneur and the startup team must design the venture for systematic failure and 

learn from people and the environment in terms of what will work for the particular solution 

to exploit the opportunities initially perceived by the entrepreneur and to ensure success 

(Furr & Ahlstrom, 2011, ch. 1; Blank & Dorf, 2012, ch. 2; Berglund et al., 2007). 

 



 

5 
 

A study shows that around 32 % of all Swedish startups founded in 2008 had failed before 

the end of 2011 (Ekonomifakta, 2013) and one reason may be that entrepreneurs do not 

adapt and change their business model as new information is received. Entrepreneurs that 

embrace deviations from the original plan and are able to recognize what initially appears 

discomforting would thus decrease the market risk and thereby increase the likelihood for 

success (Young & Sexton, 2003; Furr & Ahlstrom, 2011, ch. 1; Blank & Dorf, 2012, ch. 2; 

Berglund et al., 2007). 

 

The entrepreneurial process consists of evaluation of environment, opportunities, acquisition 

of resources and designing the business model (Allen, 2010, ch. 1), in the search for a 

repeatable, scalable, and profitable business model (Blank & Dorf, 2012, intro XVI). Due to 

the uncertainty involved in starting a new venture, planning and processes differ from 

approaches applied in existing organizations. Within startups the discovery of new data must 

constantly be assessed in terms of what new insights were gained and what to incorporate 

into the evolving plan McGrath & MacMillan, 1995).  

 

However, successful entrepreneurs are action oriented and move fast, developing a strategy 

may present great challenges, particularly due to the conditions of uncertainty. Thus, there is 

a risk that startups move too quickly and as a result, settle down on a specific path (Furr & 

Ahlstrom, 2011, ch. 1; McGrath & MacMillan, 2000, ch. 1) which is why there is a need for 

structured framework and ensuring that the startup does not proceed and ignores important 

patterns presented by the environment.  

 

2.2 The composition of the startup team 

Researchers have suggested that entrepreneurs should look for partners with complementary 

and diverse skills in order to increase performance (Allen, 2010, ch. 8; Furr et al., 2012). 

Further, an effective startup team is characterized by individuals that share the vision for the 

new venture, are passionate about the business concept, are experienced within the industry, 

have contacts within the industry to be able to raise capital, have good credit ratings, are free 

to spend time, are able to handle financial constraints, and have expertise to cover 

marketing, finance, and operation (Allen, 2010, ch. 8).  

 

However, Blank & Dorf (2012), mean that it is important to treat the startup differently in 

terms of what skills are required. Therefore, it may be that Allen’s (2010) perspective is 

more applicable at later stages when the business model has been validated, and there is a 

need for marketing, finance, and operative skills. Although, should such skills and 

experiences be considered to lie outside of the product area, i.e. be an extra-domain skill, the 

ability of the startup team to make radical changes may be significantly enhanced (Furr et 

al., 2012). 

 

During the early stages the initiator or founder, traditionally, is portrayed to be a lone agent, 

a visionary who sets out to commercialize the discoveries and develop their ideas into a 

concept that is offered through a real company (Allen, 2010, ch. 1; Cooney, 2005). In 



 

6 
 

addition, most research is focused toward individual founders or entrepreneurship within 

existing organizations. However, most startups consist of two or more individuals who 

embarque on the journey of finding the scalable business model (Cooney, 2005; Wang & 

Chugh, 2014). The creation of a founding team may be accidental, where individuals meet 

and find that their visions correspond, or the formation of a team occurs through a structured 

process, where the inventor selects individuals that fulfill certain criteria (Allen, 2010, ch. 

8). 

 

Learning and decision-making are integrated processes in startups which demand joint 

development of the venture and the entrepreneur leading the efforts (Sardana & Scott-

Kemmis, 2010). Since the founder will be the driver of the entrepreneurial process, he/she 

will influence the connection between the startup’s resources and actions on the market or 

with customers. Although the startup team conducts research in collaboration, the founder 

may be the one to have the greatest impact on all components of the venture formation 

(Cope, 2003; Allen, 2010, ch. 1). It is therefore necessary to understand the different 

interests' subsequent impact on how research during the early stages is conducted. The 

following chapter discusses potential differences in how members of a startup team collect 

information during research efforts. 

 

2.3 Collaborative research 

Two contrasting scientific approaches dominate the guidance in organizational research, 

called “inquiry from the inside” and “inquiry from the outside” (Adler et al., 2004, ch. 5). 

Similarly, Louis & Bartunek (1992) refers to these scientific approaches by classifying 

researchers as either insider or outsider. During this study the founder is referred to as the 

insider, given that the vision of the product and organization mainly exist within the mind of 

the entrepreneur. Thus, all other members of the startup team are outsiders with less ability 

to directly impact decision-making. 

 

An inside researcher is defined as a member in the organization who is concerned about the 

immediate situation and to uncover situationally relevant knowledge (Sardana & Scott-

Kemmis, 2010). Insiders have a more interpretative approach and results are contextually 

embedded. The benefit of employing only domain insiders or experts could be improved 

ability to quickly make minor changes (Furr et al., 2012). In addition, deeper understanding 

of the context makes individuals better at exploitative research where new information 

results in linear learning (Furr et al., 2012). 

 

Outside researchers have a tendency in gathering factual data and to form context-free 

meanings based on conclusions from logical reasoning (Adler et al., 2004, ch. 5; Louis & 

Bartunek, 1992). Further, outsiders are more detached from the setting and have interest in 

discovering findings that may be generalized. The ability of the outsider in bringing novel 

framing to the search for opportunities, are thus more likely than domain insiders to generate 

information that supports moderate changes during uncertain times (Furr et al., 2012). 

Considering the uncertainty of the early stages of venture formation, it may thereby be 



 

7 
 

concluded that a team that is prone to change initial assumptions and prepared to make 

major changes as new information emerges, is more likely to be successful. 

 

In addition to research members’ connection to the setting, the experience of each member 

further impacts the outcome of the study. The experiences regarding education, career path 

and previous work shape perspectives and interpretations and, in short, the more diverse 

experiences each member has in relation to the others, the more diverse are their views on 

the particular situation (Furr et al., 2012; Louis & Bartunek, 1992).  

 

Looking for complementary skills in the early stages may be part of fruitful entrepreneurial 

processes. Startup teams where members have complementary skills would mean that there 

are a larger number of mental models and thus the likelihood of discovering novel 

opportunities may be higher. In addition, Furr et al. (2012) found that teams composed of 

domain insiders developed higher degree of inertia. Thus, teams consisting of both insiders 

and outsiders are more likely to generate greater heterogeneity due to the existence of 

interpretative and observational approaches in comparison to non-diverse teams (Louis & 

Bartunek, 1992).  

 

The startup team approached potential customers in order to acquire knowledge during the 

entrepreneurial process. Next, approaches to structure the search and review opportunities 

are presented. 

 

2.4 Exploring opportunities in uncertain environments 

As previously mentioned, the entrepreneurial process benefits from implementation of a 

framework that ensures the process of evaluating opportunities is not rushed through. In 

addition, it is important that the founder and startup team are able to apply objectivity when 

searching for opportunities (Furr & Alstrom, 2011, ch. 1). Therefore, iterative processes for 

investigating market uncertainties and conducting user studies are suggested (McGrath & 

MacMillan, 2000, ch. 1; Blank & Dorf, 2012, ch.2).  

 

An example of such a process is the Customer Development method (figure 1), which 

includes continuous interaction with customers and assessment of the startup’s existing 

assumptions regarding their future business model (Blank & Dorf, 2012, ch. 1). 

 

 

 
Figure 2 – Customer Development process (Blank & Dorf, 2012) 

 



 

8 
 

The customer development process is illustrated using circular tracks to make the person 

following the method aware that the process is iterative and will include failures (Blank & 

Dorf, 2012, ch. 2).  

 

The customer discovery phase is similar to the discovery phase presented by the British 

Design Council (2005). The initial phase involves exposing oneself to high uncertainty as 

the discovery or inspiration to develop a new product is yet to be tested. Therefore, it is 

important to conduct market, and user research as well as determine how to manage 

information to and within the design research group (British Design Council, 2005; Blank & 

Dorf, 2012, ch. 3). Divergent thinking is suggested in order to go beyond traditional 

approaches and generate many ideas in the early stage. Applying convergent thinking in the 

early stages may limit the number of opportunities explored when approaching potential 

customers, although convergent thinking is required to assess the ideas generated in the 

initial phase (Sloan, 2012). 

 

By approaching customers, the entrepreneur is first to test the customer’s perception of the 

problem and if a solution to the problem is sought. When evaluating customer input during 

the first stage, the founders are to ask themselves “Is the problem important enough that the 

right product will drive significant numbers of customers to buy or engage with the 

product?” (Blank & Dorf, 2012, p. 25). Blank & Dorf (2012) present a structured approach 

to the search namely the customer development process. According to the authors, all that 

the founder has in the early stages of creating a venture is hypotheses about what the product 

should be and what customers to serve. Therefore, the customer development process should 

be designed for the entrepreneur to explore the hypotheses and discard of assumptions that 

prove to be wrong. 

 

Hence, during the process of acquiring new knowledge and learning about customers’ 

implicit and explicit needs, the entrepreneur will encounter information that goes against the 

existing plan regarding what product to develop and for whom. Such encounters are often 

difficult and result in discomfort, and will be experienced differently depending on the 

entrepreneur’s underlying attitude. Reacting to the information that proves the startup’s 

assumptions to be wrong are referred to as pivots within the customer development 

methodology (Blank & Dorf, 2012, ch. 2). Further, such an event means that the 

entrepreneur must eliminate the mental models which had lead them to pursue that 

opportunity (Young & Sexton, 2003). If the entrepreneur is able to discard existing mental 

models during moments of discomfort resulting from learning that your assumptions were 

wrong they may be able to respond in a way that will benefit the future of the new venture 

(Blank & Dorf, 2012, ch. 2; Young & Sexton, 2003). 

 

To sum up, by implementing iterative learning strategies the entrepreneur will be better 

equipped to learn and develop an understanding of potential customers and the opportunities 

that exist in terms of commercializing their product or service. Thus, the entrepreneur will 

acquire new knowledge and understanding of potential customers and the existing 

opportunities when attempting to commercializing their product or service. To better 



 

9 
 

understand how such learning processes within a startup can be explained, the different 

types of learning are explained in the following chapter. 

 

2.5 Knowledge acquisition in uncertain environments  

Argyris (1976) defines learning as acquisition of knowledge or knowing that enables the 

individual to detect and correct features that make actions ineffective. The objective of a 

startup must be to become a learning organization, not a knowing one, in order to thoroughly 

understand the customers’ needs (Furr & Ahlstrom, 2011, ch. 1; Blank & Dorf, 2012, ch. 1).  

 

Since the entrepreneurial process is about finding novel entry into new or established 

markets, and exploiting new or existing products and services (Allen, 2010, ch.1) there is a 

need for path breaking discoveries resulting from thinking outside of the proverbial box. 

Learning during the entrepreneurial process can therefore not only be a linear process of 

acquiring new knowledge, making it important to understand how different types of learning 

may be developed. Berglund et al. (2007) explored the possibilities of including venture 

capitalists that expanded the scope of tested hypotheses, resulting in increased value of the 

venture to which investments were contributed.  

 

Learning through testing of assumptions, such as the hypothesis testing approach suggested 

within the customer development framework, is constituted by discrete tests of hypotheses 

in order to determine what data fulfills predetermined criteria (Berglund et al., 2007). The 

hypothesis testing approach thereby corresponds to validating and rejecting hypotheses as 

described by Blank & Dorf (2012) as a part of the customer development process. When 

developing learning based on a previously known context the discoveries exploit current 

mental models and deepen the understanding of what is already known. Such learning may 

become linear and could be compared to lower-level or single-loop learning (Argyris, 1979; 

Cope, 2003). Exploitative learning processes and hypotheses testing may thus improve the 

understanding of a particular customer problem. In relation to developing a product 

exploitative learning will generate information that is beneficial in developing a specific 

product but is limited in exploring the possibilities of radical changes to the set of product 

features. 

 

On the other side of the spectrum is explorative learning processes, where discoveries are 

generated through enactment and interpretation of results. Included in theories related to 

such processes are the hermeneutic perspective (Berglund et al., 2007), higher-level learning 

(Cope, 2003), and double-loop learning (Argyris, 1979). Data collected during exploratory 

processes are wide spanning and the value yielded may not be instantly visible (Wang & 

Chugh, 2014). Path breaking discoveries necessary during the innovation process often 

result from so called double-loop or higher-level learning (Argyris, 1976; Cope, 2003; 

Harrison & Leitch, 2005), which may be stimulated by e.g. critical examination of the 

individual’s identity and life story (Argyris, 1976). The following section explores 

interviews as learning tools to apply during the entrepreneurial process. 

 



 

10 
 

2.6 Interviews as learning tools 

Qualitative interviewing is a useful tool when seeking to acquire market and customer 

related information. However, there are several types of interviewing and categories of 

questions, some more appropriate than others given the research setting. 

 

In interviewing, knowledge is constructed through dialogue between the interviewer and the 

respondent. The interviewer must make sure that the respondent feels comfortable despite 

the clear asymmetry of power in order for the respondent to talk freely about his or her 

experiences. The interviewer must be attentive and show interest, understanding, and respect 

both to what the respondent is saying and even not saying (Bryman & Bell, 2011, ch. 18; 

Kvale, 1996, ch.8). Thus, it lies within the ability of the interviewer to grasp the meaning of 

the answer and determine how the interview is to proceed. The ability is a function of the 

interviewer’s knowledge and interest in the subject matter and the human interaction (Kvale, 

1996, ch. 8). 

2.6.1 Deciding on the type of interview 

There are several types of interviewing, ranging from the structured interview to the 

unstructured interview, where the latter tends to be utilized in qualitative research. In 

qualitative interviewing, emphasis is on the respondent’s own perspective and the purpose is 

to generate thick and detailed answers. A less structured interview is preferable when the 

researcher wants to obtain thorough understanding of the respondent, however, to enable 

comparison between interviewing styles some structure is required (Bryman & Bell, 2011, 

ch. 18; Kvale, 1996, ch. 8). Thus, it is key to let the respondent talk about problems from the 

respondent’s point of view in order to gain a thorough understanding of the customer, which 

makes it appropriate to use the semi-structured interview.   

 

The semi-structured interview allows the respondent to take off on interesting sidetracks. 

The interviewer is permitted to depart from the interview guide, change the order of 

questions, and add questions as the interview process unfolds. It is important that the 

researcher is familiar with the setting under study in order to understand the respondent’s 

point of view before conducting the interviews (Bryman & Bell, 2011, ch. 18; Kvale, 1996, 

ch. 8). It is essential that the interviewer listens carefully without interrupting or prejudging 

respondent’s answers, from a phenomenological perspective on qualitative interviewing 

(Kvale, 1996, ch. 8). 

 

Further, the type and design of questions affects the quality and usability of the information 

that is gathered. Next, the characteristics of questions and their impact on the gathered 

information are explained. 

2.6.2 Designing questions to explore the environment 

Questions must be brief and simple for the respondent to comprehend the question, 

according to Kvale (1996, ch. 8). The interviewer will ask personal factual questions when 

demographic or personal information concerning e.g. age and education are of interest. This 

category of questions also includes questions about behaviour and e.g. length and frequency 



 

11 
 

of certain behaviours. The interviewer may also want to ask questions regarding the 

respondent’s work position and certain attributes about the company, referred to as informant 

factual questions. Finally, the interviewer may be interested in asking questions considering 

the respondent’s knowledge or attitudes in certain areas (Bryman & Bell, 2011, ch. 10). 

  

There are nine main types of questions to be asked during semi-structured interviews. The 

first introducing questions let the respondent describe the main characteristics or dimensions 

of the phenomena from the respondent’s point of view. Then, the interviewer is able to 

follow up on dimensions brought up by the respondent (Kvale, 1996, ch. 8).  

 

Following up may include either probing questions or interpreting questions. Probing 

questions do not suggest upon what dimensions of the answer to elaborate on but it is up to 

the respondent to decide (Kvale, 1996, ch. 8). Further, reasons for asking probing questions 

are to find root causes to customer problems. By for example using the “five whys” 

technique, interviewers are able to uncover underlying meanings of responses (Ries, 2010). 

In addition, if an answer includes general or ambiguous terms the interviewer may ask 

specifying question to obtain specific data. Interpreting questions on the other hand may be 

used to clarify responses, the interviewer may e.g. rephrase the respondent’s answer in order 

to have it confirmed (Kvale, 1996, ch. 8). 

Later on in an interview, after the respondent has talked about the phenomena in her own 

words, the interviewer is allowed to ask more direct questions by introducing dimensions of 

the phenomena to the respondent. Indirect questions are used when the interviewer is 

interested in investigating how the respondent refers to e.g. attitudes or behaviours of others. 

As previously mentioned, it is the interviewer that steers the course of the interview, thus, if 

the respondent elaborates on an answer that is of no interest for the study, the interviewer 

may ask structuring questions to keep the interview on track. At the same time, the 

interviewer may chose to remain silent, which gives the respondent time to reflect. Finally, 

interpreting questions may be addressed to clarify responses, the interviewer may e.g. 

rephrase the respondent’s answer in order to have it confirmed (Kvale, 1996, ch. 8). 

 

Besides the type of question to ask, the interviewer must consider how questions are 

constructed. The following rules should be considered and borne in mind when conducting 

interviews (Bryman & Bell, 2011, ch. 10).  

 

First, the interviewer should avoid using ambiguous terms, such as “often”, which is 

interpreted differently dependent on respondent’s frame of reference. In addition, common 

words, such as for exampe “quality” have different meanings for different people. To avoid 

such fallacies the interviewer may first ask the respondent to define certain words and 

concepts before they are used. 

 

Second, the use of leading questions to steer the respondent’s answer into particular areas 

should be avoided as it may result in the respondent feeling stressed to answer in certain 

ways and refrain from objecting even though opportunity is given by the interviewer 

(Bryman & Bell, 2011, ch. 10; Rowley, 2012). 

 



 

12 
 

Third, the interviewer should avoid negatives and double-barrelled questions, such as “how 

satisfied are you with pay and conditions in your job?” (Bryman & Bell, 2011, p. 256), 

which may lead to misconceptions (Bryman & Bell, 2011, ch. 10; Rowley, 2012).  

 

Finally, the interviewer should avoid asking questions in the form “when did you last discuss 

your training needs with your supervisor?” (Bryman & Bell, 2011, p. 258) since it takes for 

granted that the respondent discusses training needs with the supervisor, which may not be 

the case. The interviewer should avoid questions that are too general since it may lead to 

diverse interpretations among respondents, which ultimately impacts on results (Bryman & 

Bell, 2011, ch. 10).  

 

The quality of the interview conducted will impact the quality of the subsequent analysis, 

therefore, the interviewer should be aware of quality criteria to guide the interviews (Kvale, 

1996, ch. 8). According to Kvale (1996, p. 145) the quality of an interview is dependent on:  

1. The extent of spontaneous, rich, specific, and relevant answers from the respondent. 

2. The shorter the interviewer’s questions and the longer the respondent’s answers, the 

better. 

3. The degree to which the interviewer follows up and clarifies the meanings of the 

relevant aspects of the answers. 

4. The ideal interview is to a large extent interpreted throughout the interview. 

5. The interviewer attempts to verify his or her interpretations of the respondent’s 

answers in the course of the interview. 

6. The interview is “self-communicating”, it is a story contained in itself that hardly 

requires much extra descriptions and explanations. 

 

In sum, the interviewing technique enables researchers to learn about the respondent and its 

setting. By objectively searching for such opportunities and threats to the startup, the amount 

of resources dedicated may be decreased, building lean processes and learning into the 

organization from the very beginning (Blank & Dorf, 2012, ch. 2; Furr & Ahlstrom, 2011, 

ch. 1). However, the search presents challenges when integrating the individuals’ learnings 

as such learning can be limited in different ways. In the following chapter, the barriers to 

double-loop or higher-level learning, objective search and thus barriers to path breaking 

learning are presented. 

 

2.7 Barriers to learning 

The human intellective capacity is limited in comparison with the complexity of problems 

facing an individual or organization, as a consequence, the human mind creates simplified 

models only including the main features of the problem (March & Simon, 1967, ch. 6). 

Thus, the founder, who wishes to validate assumptions rather than being wrong, may 

unintentionally present subjective stimuli during the data collection, thereby limiting the 

ability of objective search (Hellevik, 1996, ch. 5). 



 

13 
 

2.7.1 Individual level barriers to learning 

Since individuals are not able to search and interpret information accurately (Simon et al., 

2000), individuals reduce complexity of a situation or problem by simplifying, using so 

called heuristics (Kahneman et al., 1982, preface). Thus, by using heuristics the individual is 

able to speed up the process of getting acquainted with a new situation. 

 

People tend to base beliefs of uncertain events on probabilities (Kahneman et al., 1982, ch. 

1). Economical concepts of rationality only holds true for cases of certainty, however, in 

cases of uncertainty rationality is limited since definite probabilities cannot be assigned 

(Kahneman et al., 1982, ch. 1; March & Simon, 1967, ch. 6), as for example the uncertainty 

involved in start-up processes.  

 

Existing models of rationality suggest that all alternatives of choice are given, all 

consequences attached to alternatives and, further, rules and principles for ordering those 

consequences are known. From a phenomenological viewpoint, humans behave rationally 

only relative to a frame of reference or some set of ”given” characteristics, which define the 

situation as it appears to the actor (March & Simon, 1967, ch. 6). In times of uncertain 

events, people tend to rely on heuristics, or subjective assessment of probability when 

interpreting the information that is being received. The use of heuristics can lead to biases 

that hinder the ability to make accurate decisions (Kahneman et al., 1982, ch. 1). 

 

Cognitive biases are common types of mental shortcuts to help individuals cope with their 

limited cognitive capacity when making judgements of uncertain events (Simon et al., 

2000). Cognitive biases stem from three categories of heuristics, namely; representativeness, 

availability, and adjustment and anchoring, according to Kahneman et al. (1982, ch. 1).  

 

Representativeness heuristic is explained by people making judgements based on how well a 

sample represents a stereotype, e.g. if A is similar to B, then the probability that B originates 

from A is considered to be high. Biases stemming from availability heuristic concern the 

degree to which the occurrences of an event can be brought to mind. In other words, people 

estimate probability based on the ease to recall a situation or an example. Finally, adjustment 

and anchoring heuristics occur due to people's tendency to rely on initial values or 

information, and as a result, estimates are biased toward the reference point. Then, by 

altering the starting point different estimates are obtained (Kahneman et al., 1982, ch. 1).  

2.7.2 Different types of bias 

Several biases can be connected to the heuristics presented previously in accordance with 

Kahneman et al. (1982). The confirmation bias occurs due to people’s tendency in filtering 

information that confirms their own beliefs in favor of contradictory information (Furr & 

Ahlstrom, 2011, ch. 3; Lau, 2011, ch. 20). As a consequence, people fail to use information 

and evidence properly. People are in general more prone to stick to their beliefs, even though 

contrary evidence is presented, and to look for evidence that support their own beliefs in 

favor of searching for opposing information (Lau, 2011, ch. 20). As a result, it becomes 



 

14 
 

difficult for entrepreneurs to thoroughly understand their market and customers since 

contrary feedback is filtered out (Furr & Ahlstrom, 2011, ch. 3).  

 

The overconfidence bias suggests experienced individuals are overconfident about their 

abilities (Furr & Ahlstrom, 2011, ch. 3; James & Barnes, 1984) or unaware about their own 

limits in knowledge (Simon et al., 2000). According to Lau (2011, ch. 20), people think they 

perform above average, e.g. research shows that more than 50 % of drivers think they drive 

better and safer than average. The overconfidence bias occurs due to individuals tendency to 

avoid adjusting initial data accurately after new information has emerged or because 

individuals are prone to base certainty on the ease to recall (Simon et al., 2000). Related to 

the overconfidence bias is the optimism bias suggesting individuals are overly optimistic 

about the outcome of future events (Lau, 2011, ch. 20) and further, also have the ability to 

control or predict future events (Simon et al., 2000). Further, the overconfidence bias is 

amplified in complex situations resulting in entrepreneurs that are less eager to learn and 

change their beliefs (Furr & Ahlstrom, 2011, ch. 3) and may ignore obstacles (Simon et al., 

2000).  

 

In addition, familiarity or capability bias suggests that organizations tend to rely on their 

competencies, which might result in a future lock-in. The implications for entrepreneurs 

might be the reuse of ideas from familiar settings into new, unfamiliar and inappropriate 

settings (Furr & Ahlstrom, 2011, ch. 3).  

 

An individual’s judgments are affected by how the problem is presented, thus estimates are 

manipulated by differing the point of reference (Kahneman et al., 1982, ch. 1; Lau, 2011, ch. 

20). Research shows that when subjects were asked to estimate quantities and were given 

numbers in forehand by spinning a wheel of fortune, estimated quantities depended on the 

number given by the wheel of fortune. In short, a relatively high number on the wheel of 

fortune resulted in a higher estimation of quantity than an estimation followed by a low 

number in the wheel of fortune (Kahneman et al., 1982, ch. 1). The framing bias suggests 

that the formulation of a problem affects decision-making, i.e. people think differently 

regarding the same choice depending on how alternatives are described. In general, people 

tend to avoid losses over acquiring gains (Lau, 2011, ch. 20), which is explained by sunk 

costs theory suggesting people feel obligated to keep investing, otherwise previous 

investments will have been wasted (Furr & Ahlstrom, 2011, ch. 3). This is why people tend 

to assign more resources to a failing project rather than terminating it. Further, entrepreneurs 

often put their hope and reputation on a startup, which turns decision-making into a difficult 

matter when exposed to failure (Furr & Ahlstrom, 2011, ch. 3). 

 

Individuals are prone to draw conclusions even from limited sample sizes suggesting that the 

sample is representative for the population from which it is drawn, referred as the “law of 

small numbers” (Kahneman et al., 1982, ch. 1; Simon et al., 2000). Individuals tend to base 

forecasts on a limited amount of positive information, consequently, if feedback from 

potential customers is limited due to the number of respondents it may lead the entrepreneur 

into fallacy. In addition, a small sample size increases the risk to only retrieve positive 

feedback, however, if feedback is gathered from a larger amount of potential customers it 



 

15 
 

would probably comprise negative feedback as well serving a better representation of the 

population (Simon et al., 2000).  

 

The way that organizations evolve in terms of culture and potential action paths develops 

through historical events. Historical events are created in the early days of the new venture 

and therefore could include any action taken by the founder or the startup team. In the case 

the founder is a domain expert, with knowledge on prevailing practices within the industry, 

the likelihood for the venture taking a deviant approach may be limited due to the 

individual’s existing mental models (Staber, 2005). 

 

The cognitive biases may inhibit individuals from developing learning in an objective and 

explorative way. Further, Sydow et al. (2009) explain how self-reinforcing patterns result in 

certain emotional reactions, cognitive bias, and political processes. In organizational 

research the term path dependence is used to explain how past behavior may affect or force 

certain subsequent actions (Sydow et al., 2009). Thus, the biases of the individuals shaping 

the learning and decision-making processes within the startup may subsequently affect the 

design of the business model (Staber, 2005).  

2.7.4 Overcoming barriers due to biases  

In order to overcome cognitive bias, entrepreneurs must establish a learning culture that is 

able to seek and receive feedback, which is essential to discover real opportunities (Furr & 

Ahlstrom, 2011, ch. 3). Entrepreneurs must balance between confidence and distrust in what 

they know, be receptive to real-time feedback in order to stick to the reality as well as avoid 

overconfidence in assumptions. Since the individual’s cognitive frameworks, routines and 

heuristics shape the processing of the input (Sardana & Scott-Kemmis, 2010), this could 

mean that entrepreneurial learning tends to be path dependent and thus that the path of the 

new venture may be largely affected by the characteristics of the entrepreneur.  

 

An organization with strong dynamic capabilities is developed by ensuring that previously 

mentioned self-reinforcing patterns do not result in lock-in situations prohibiting 

organizational change. Such action is difficult to take, since the judgement of the individual 

is subjective. The behavior and actions of the founder and its team determine the firm’s 

future ability to adapt to changing environment and its ability to survive (Sydow et al., 

2009). 

 

Further, decisions must be based on accurate data from potential customers and not on 

beliefs or gut feelings (Furr & Ahlstrom, 2011, ch.3). Such a balance is difficult to achieve, 

not only because the entrepreneurs are constrained by cognitive bias, but also that they do 

not embarque on the journey alone, receiving input from various sources along the way 

(Staber, 2005). Sydow et al., (2009) suggest that path dependencies may be broken by taking 

a critical stance and reflecting on the drivers that made it happen, and by understanding the 

social mechanisms driving the path process.  

 

Thus, understanding the ways in which individuals gather information and what that 

information includes, may unveil the drivers of decision and explain how different types of 



 

16 
 

learning are created during startup processes. By analysing the ways in which individuals 

within the startup team gathers information during interviews the understanding of how 

learning can be explained increases. In the following chapter, the methods used when 

collecting empirical data to develop the theoretical explanation are described. 



 

17 
 

3 Methodology 

In this chapter the inductive research approach employed to develop theoretical 

explanations to the entrepreneurial learning phenomenon is explained. In addition, the 

research design, sampling, data collection methods, and quality criteria are described. 

 

3.1 The initial study 

An initial study including exploratory interviews was carried out to get for the researchers 

familiar with the setting. In addition, it is essential to define a strict scope for what to include 

in the description of the phenomenon to enable comparison and to ensure that variations in 

collected information is a result of interviewer characteristics rather than differences in 

customer traits. Further, the initial study played an important part for the outsider 

researchers’ understanding of the founder’s product vision. The results from the initial study 

were then used to define the scope of the demarcated study. 

 

During the initial phase, the founder first drew the business model consisting of hypotheses 

about the nine building blocks. In addition, the outsider team held several discussions with 

the founder to understand the underlying idea behind the solution and the product vision. 

The outsider team developed an interest in the idea as they saw great potential in the 

founder’s vision.  

 

Jointly the team set up an interview guide (Appendix B, Interview guide B1), consisting of 

20 open questions seeking answer to how procurement activities are organized within firms 

of different size. The founder and the outsiders each conducted 7 semi-structured interviews 

(Appendix A, Interview list A1), reasons for selecting a semi-structured approach during 

interviewing is further explained in the theory section as learning tools are explained.  

 

3.1.1 Results from the initial study 

The results from the initial study indicated that there were differences in how the founder 

and the outsider research team gather new information during customer discovery. More 

specifically, patterns of varying depth and specificity of the information collected were 

indicated.   

 

Information gathered by the insider interviewer included more detail and enabled validation 

of hypotheses to another extent than the outsiders due to the use of more specific questions. 

The outsiders on the other hand found a new area that appeared to present opportunities for a 

niche product that some customers were searching for. In order to investigate the differences 

further, a demarcated study was carried out, where the researchers were separated to ensure 

that there were no parts of the collaboration that affected the way in which information was 

gathered by the two individuals.  

 

The results from the initial interviews were analyzed jointly, resulting in segmentation 

parameters to investigate the potential of developing a product that tends to particular needs 



 

18 
 

for improvement within procurement processes. Within procurement there are several 

processes and activities, potentially involving multiple stakeholders, internal and/or external. 

The hypotheses concerning what type of individuals to approach during the demarcated 

study are described below:  

 

 Individuals managing procurement are potential customers as they will want to 

be able to capture all information and activities involved in the procurement 

process. 

 Individuals in charge of cross-functional teams are potential customers. 

 Purchasing and procurement managers at companies acting in industries where 

output requires product innovativeness are potential customers. 

 Individuals in charge of many and large procurements are potential customers. 

 

Certain parameters were found to affect the firms’ need and willingness to develop their 

procurement activities, these are listed below: 

 

 Complex organizations where stakeholders for a certain procurement may be 

found in several departments such as quality, R&D, and production. 

 The firm’s yearly turnover has been found a potential ground for segmentation as 

it affects the distance between individuals and departments. The strata therefore 

includes firms of more than 500 million SEK in revenue and a top limit of 

approximately 10 000 million SEK. 

 Purchasing organizations faced with complex requirement definitions due to new 

task or modified rebuy situations which result in involvement of multiple internal 

stakeholders/cross-functional expertise.  

 

Situations where problems may be solved using the tool were found to occur mostly in large, 

complex procurement projects, consulting the illustration below the demarcated study aimed 

to include firms that may be placed toward the second upper quadrant, i.e. “many and large”, 

in the figure below. 

 

 

 
Figure 3 – Classification of types of procurement projects 



 

19 
 

 

In short, the results from the initial study indicated differences in information gathering 

between insider and outsider interviewer. Therefore, it was decided to conduct a demarcated 

study to further investigate the potential differences. The results from the initial interviews 

were analyzed jointly and then used to set the scope of the demarcated study. 

 

3.2 The demarcated study 

It is necessary to separate the interviewers over some time and let the founder and the 

outsider conduct interviews individually to identify differences in interviewing techniques 

and how the interviewer perceives the situation of the respondent. This is in accordance with 

the phenomenological approach, where the purpose is to describe humans’ experiences 

related to a phenomenon (Berglund, 2007).  

 

3.2.1 Research design 

A comparative research design was used to examine differences in approaches between 

insider and outsider during interviewing. According to Bryman & Bell (2011, ch. 2), a 

comparative research design is equal to a multiple-case study research design when a 

qualitative research strategy is applied. However, in contrast to the initial study, the 

demarcated study was designed to allow comparison between interviewer approaches by 

examining two groups of interviews, i.e. interviews conducted by the insider against 

interviews conducted by the outsider, rather than examining respondents’ answers, which 

was the purpose of the initial study to set the scope of the following research. 



 

20 
 

A comparative design was selected since social phenomena is easier understood when two or 

more contrasting cases are compared (Bryman & Bell, 2011, ch. 2). By using a replication 

logic, viewing all cases as series of experiments and then comparing relationships across 

cases, it is possible to validate the consistency in relationships and determine when those 

relationships hold. In other words, a relationship that is confirmed with another case adds 

validity to the findings (Eisenhardt, 1989). 

3.2.2 Research methods 

Identical research methods were used to enable comparison, in this case semi-structured 

interviews. Due to time restrictions considering transcribing and coding of audio recordings, 

eight semi-structured interviews was conducted in total (Appendix A, Interview list A2). The 

founder and the outsider team conducted 4 interviews each over a period of 30 working 

days. Reasons behind selecting a semi-structured approach during interviewing are further 

explained in the theory chapter.  

3.2.2.1 Data sources 

Respondents were selected based on the customer segment mapping performed after 

exploratory interviews had been conducted. In addition, discussions with the founder were 

held as the researchers developed an interview guide for the demarcated study (Appendix B, 

Interview guide B2). Since interviews were audio recorded, it was important to explain 

reasons for recording and how the collected data will be utilized before starting the interview 

(Bryman & Bell, 2011, ch. 18). In case the respondents refused to be recorded their 

responses were excluded since transcription of recordings was needed in order to conduct 

the analysis. Further, consideration was taken to overly short interviews without 

comprehensible answers as such responses indicate that there are motives for covering up. 

3.2.2.2 Design of interview guide 

The interview questions aimed at investigating customer problems related to procurement 

processes. It was highlighted that requirements management was the most difficult task to 

deal with during the initial set of interviews with purchasing managers. Therefore, the pains 

and pain relievers connected to requirements capture were further investigated by focusing 

on the founder’s hypotheses regarding early involvement and requirements management, 

while building on the features that solve issues in those areas. 

 

An interview guide (see Appendix A) with open-ended questions was compiled to examine 

customer pains and to validate and/or pivot value proposition hypothesis. The main 

questions were supplemented with probing questions in order to provide as comprehensible 

answers as possible. Extensive pretesting of the interview guide was undertaken by 

interviewing six individuals within the target customer segment to ensure that the questions 

were understood without explanations. 

 

The value proposition hypotheses selected were derived from a thorough evaluation of the 

current business model, which is based upon the founders’ perception of the company 

realisation, and insights from interviews. The value propositions as described by the founder 

and concretization of the problem are illustrated below: 



 

21 
 

 

1. Purchasing managers conducting large procurements (cost or time) need a tool where 

team members collaborate to capture requirements. 

2. The purchasing manager lacks overview of ongoing procurement projects. 

3. Purchasing managers need a tool with which he/she will be able to review all other 

relevant aspects of a procurement earlier in the process. 

4. Purchasing managers lack access to details on each procurement conducted within 

the organization. 

5. Requirements management is easier and more efficient by tying each individual to 

any number of roles and activities. 

6. The purchasing manager is frustrated about members of the procurement team 

spending a lot of time searching through e-mails and documents. 

 

The outsider research team let one person carry out the interviews during the demarcated 

study to ensure that the data from the insider/founder and the outsider could be compared. 

The data was collected over a time period of 30 working days and interviews were 

conducted over telephone. 

 

The interviews were audio recorded and transcribed in order to allow analysis of how 

interviewer input is formulated and not solely on what the respondents say. The audio 

recording enabled the researchers to fully concentrate on what was being said in order to 

detect e.g. inconsistencies and nuances in language (Bryman & Bell, 2011, ch. 18). The 

transcription of interviews enabled both qualitative and quantitative analysis of the data. 

Thereby, it was possible to detect patterns and form categories within the interviewer’s input. 

Since the research team consisted of two people, of which only one conducted interviews 

during the demarcated study, it was possible for the other researcher to transcribe the 

outsider’s interviews directly as they were conducted, which is suggested by Bryman & Bell 

(2011, ch. 18). The researchers received the founder’s recordings after the interviews were 

conducted. Thus, the transcribing of founder’s interviews was carried out after all interviews 

had been conducted, ensuring that the founder did not affect the outsider interviewer. 

3.2.3 Data analysis 

The collected data from the interviews was reviewed in terms of qualitative and quantitative 

dimensions, illustrating differences from multiple perspectives. The ways in which data was 

reviewed and analyzed are presented below. 

All transcripts from the interviews were first reviewed by the researchers individually in 

order to ensure that the perception of the interviewers’ outputs did not differ between the 

person carrying out the interview and the researcher that had observed the interviews. 

Thereafter, the researchers analyzed one of the insider’s interviews and one of the outsider’s 

interviews to detect patterns in interviewer input. This process resulted in a set of categories 

that described the different types of interviewer input.  

 

Descriptions of each input were generated to support the researchers in developing insights 

about the cases, allowing for unique patterns to emerge before developing a generalization, 



 

22 
 

by carrying out a thorough within-case analysis of each interview (Eisenhardt, 1989). The 

large number of descriptions of input was then analyzed within the research team, resulting 

in elimination of some descriptions that were similar and development of a set of micro-

categories (see Table C1) that could be used to explain all input within the data. The micro-

categories were then grouped together, forming main categories of input types and focuses 

to be analyzed. 

 

Each of the interviewers’ input was then assigned to one of the micro-categories in order to 

ensure that the researchers had the same perception of how to assign micro-categories in a 

cohesive manner. The qualitative analysis was conducted by analysing the categories to 

which input is assigned as well as comparing insider and outsider data.  

 

Results from the interviews were developed by categorizing each interview input and 

comparing the extent to which they occur within the insider and the outsider interviews 

respectively. Thus, a quantitative assessment of the data was conducted as well by 

comparing the occurrence of categories between insider and outsider interviews. Further, a 

dynamic dimension, i.e. differences in category occurrence between the beginning toward 

the end of the interviews was added to complement the static dimension, which views 

interviews at large to detect differences throughout interviews. 

 

3.3 Validity and reliability 

The quality criteria presented by Yin (2009, ch. 2) and Bryman & Bell (2011, ch. 16) have 

guided the assessment to evaluate the quality of the research. Some researchers argue that 

same criteria for establishing quality in quantitative research can be applied to qualitative 

research, whereas others suggest the opposite due to the fundamental difference between 

qualitative and quantitative research (Bryman & Bell, 2011, ch. 16). In this research, the 

quality criteria as defined by Yin (2009, ch. 2) were applied since these are developed in 

particular for case studies. 

3.3.1 Construct validity 

The construct validity criterion is addressed to ensure that the observations represent what 

was supposed to be investigated. Researchers must find evidence from multiple sources and 

establish a chain of those evidences to meet these criteria (Yin, 2009, ch. 2). The 

categorization of input formed the basis of the qualitative analysis and was first carried out 

individually. A structure displaying interviewer characteristics, slowly started to appear by 

reviewing transcripts and categories multiple times in combination with reading relevant 

literature. The structure was compiled by discussing the framework of categories thoroughly 

as well as with an external supervisor. Construct validity is considered high due to the 

consistency of categories across interviews when conducting the analysis.  

3.3.2 Internal validity 

For this criterion to hold the researcher must establish a good match between relationships 

and explain how these relationships emerged (Yin, 2009, ch. 2). Further, inconsistent 



 

23 
 

explanations should be considered as well. Multiple sources from various authors were used 

to achieve a comprehensive understanding of the field to ensure internal validity. In addition, 

the demarcated study was conducted during a limited period of time to reduce the likelihood 

of rival explanations to the causal findings and to ensure that the individuals had a similar 

perspective concerning the offering. Finally, the research team did not conduct any bias 

literature study until data collection was finished to not affect the result. 

3.3.3 External validity 

The external validity deals with the degree findings can be generalized into other settings 

(Bryman & Bell, 2011, ch. 16). According to Yin (2009, ch. 2), external validity is the major 

concern when conducting case studies. Theory must be tested and findings replicated in 

another setting where the same results should occur to meet this criterion. This strategy of 

testing is referred to replication logic and was used to test findings from one interview to 

another. The research was concentrated around one single insider and outsider interviewer, 

thus, further research is required in order to validate connections between relationships and 

interviewer profiles even though there was consistency in relationships across interviews. 

3.2.4 Reliability 

The last criterion is reliability and refers to the replication of the study. This is also 

considered a difficult criterion to attain since the social setting in which the research takes 

place changes over time (Bryman & Bell, 2011, ch. 16). Therefore, researchers must 

document procedures to allow for repetition of findings (Yin, 2009, ch. 2). The replication of 

findings is considered high since the research focuses on the difference in learning styles, 

which is independent of the specific company setting or situation. Further, both teams let 

one person carry out the interviews during the demarcated study to avoid inter-interviewer 

variability to ensure internal reliability of the study. It was also decided to conduct all 

interviews over telephone to make the interviewing setting as similar possible across 

interviews.  

 

 

3.4 Self-criticism 

The limitations to the study in terms of how it has been carried out are treated as the subjects 

which may be disputed arise. There are however some general limitations which deserve to 

be mentioned here in order to clarify shortcomings of the results presented previously. 

 

First, due to time limitations resulting from the fast moving startup environment, the 

interviewers that were studied only managed to conduct four interviews each within the 

customer segment that was selected for the demarcated study. After reviewing the results 

from the pre-study and finding a different customer segment, a major pivot was made. Since 

the demarcated study required extensive testing of interview guide and hypotheses to be 

tested by the interviewers there was no room for developing new material in order to 

increase the number of interviews from which insider/outsider data could be collected.  



 

24 
 

 

Second, we do not evaluate whether or not the information received from the respondents 

actually generate a larger number of opportunities, neither by generating specific 

information nor general information. Thus, it is not possible to draw conclusions regarding 

what information was more valuable, rather we have shown what type of learning was 

encouraged by the use of a particular category of interviewer input. 

 

Third, due to the sample size and the amount of empirical data consequently means that no 

precise conclusions regarding the type of bias resulting in particular input could be drawn. 

We only touched upon the possibility of certain interviewer characteristics stemming from 

the interviewer’s way of using input but were not able to categorize all input in terms of the 

bias causing them. 

 



 

25 
 

4 Findings 

In this chapter the results of the insider and outsider interviews are presented in a 

quantitative respectively qualitative section. The quantitative section includes a calculation 

of the amount of words in order to determine the proportion of speech between insider and 

outsider interviewer in interviews and between phases. Further, the categories found by 

reviewing of transcripts are presented in a figure in the qualitative section. The 

categorization of the data is then presented in multiple formats in order to visualise 

differences in insider and outsider data. Tables that are not presented in the report are found 

in appendix. 

Patterns regarding amount and characteristics of input were revealed as the transcripts were 

reviewed. Each interview was divided into five phases in order to determine if input was 

present in all phases of the interview, or if certain input was only used in particular phases. It 

thereby became useful to divide the presentation of findings into static and dynamic 

parameters within the quantitative and qualitative approaches respectively. The reason for 

dividing the interviews into five phases was to enable detection of differences across stages 

of the discussions. More phases would have made findings harder to review due to the low 

number of input in each phase, and fewer phases would not have displayed results as clearly.  

A slight decrease in interviewer proportion of speech is possible to detect using five phases. 

A decrease in proportion of speech is important to detect as it may result from extensive 

respondent elaboration, and therefore does not contribute to explaining patterns in 

interviewing characteristics.  

 

4.1 Quantitative 

4.1.1 Results from static assessment 

The collected data shows that the insider contributes a larger number of words, ranging 

between 33.0 % and 54.9 % of the total amount of input during the interview, compared to 

the outsider whose contribution ranges between 11.9 % and 29.1 %, which is seen in the 

table below. 

 

Interestingly, the word count of the respondent does not differ when comparing insider and 

outsider respondents as seen in Table C2, thereby suggesting that the differences in 

interviewing characteristics will generate significantly different types of information and 

thereby result in different kinds of learning. 



 

26 
 

 

4.1.2 Results from dynamic assessment  

The table illustrates that in all interviews except the first one, the interviewer leaves more 

space for the respondent during one of the five phases. In interview two and three more 

space is allowed in phase four, and in interview number four, the respondent contributes 80 

% of the total amount of words in phase 5 (Table C3).  

 

The lower percentages, only occurring in one phase per interview are considered outliers, as 

they do not seem to represent the interviewer’s general characteristics. Thus, the average 

insider interviewer contribution is evenly distributed over the phases. During the first phases 

the interviewer input is between 27.9 % and 51.5 %, and during the last two phases between 

17.2 % and 60.7 % (Table C3). Further, by looking at the proportion of speech throughout 

the five phases, it is shown to vary between 37.4-48.7 % (Table C4). 

 

The outsider interviewer inserts an average of 18.3 % of the total amount of words, ranging 

between 4.5 % in interview two and 37.9 % in interview one. In interview two and three the 

amount of input is kept under 22 % during the entire interview whereas in interview one and 

four the interviewer input exceeds 30 % in some cases. During interview one the interviewer 

is contributing more than 20 % in all phases, on the other hand, interview four reaches a 

percentage of 34.4 % in the final phase (Table C5). The proportion of speech varies between 

15.1-22 % throughout the phases (Table C6). 

 

The insider interviewer exceeds 30 % proportion of speech in the later stages of the 

interview in all but two instances, interview 3 phase 4, and interview 5 phase 5 (Table C3). 

The outsider interviewer also exceeds 30 % in the final phases of the interview, occurring 

two times in phase 5 (Table C5). 

 



 

27 
 

4.2 Qualitative 

The following section includes the qualitative representation of the collected data. As 

presented in the quantitative section, the amount of input differs between insider and 

outsider due to differences in interview length, which consequently affects the amount of 

data to base conclusions. Furthermore, the observed patterns must be weighed against the 

sample size, i.e. conclusions are restricted to the size of the sample. 

4.2.1 Categories 

The collected data is divided into categories in order to observe differences in insider and 

outsider approaches during interviewing. The connections between focus, input type, and 

input nuance is illustrated in the figure below (Figure 4), and further described in the 

following passages. 

Figure 4 – Categorization of inputs 



 

28 
 

4.2.2 Focus 
The interviewer focused the questions toward different areas during the discussions. At some 

points, the interviewer explores contextual factors, either by asking personal factual 

questions, to understand the respondent’s role and responsibilities, or informant factual, 

including characteristics of the organization in which they work. The interviewer also had 

the possibility of using problem exploring focus in the attempt of getting objective 

information about difficulties that the respondent experiences in daily work. In addition, 

since the interviewers had untested hypotheses regarding potential customers and their 

problems, at certain points during the interview it would be valuable to test such hypotheses 

in order to get confirmation or rejection from the respondent. Such focus is referred to as 

hypothesis verification, and includes confirmation of problems not yet stated by the 

respondent being interviewed, and matching against product/solution features.  

4.2.3 Input types and nuances 

The input was also categorized by type: new question, follow-up question, statement, and 

interjection, to further define and understand the dynamics of different questions. The 

different types are described below and by using the input nuances and examples from those, 

further micro categories for each of the input types and nuances are found in appendix 

(Table C1).  

 

Examples from the interviews are presented in the following passage. Insider and outsider 

interviewer are referred to as In(n), and Out(n) respectively, where n indicates in which of 

the interviews that the example was found. Respondents are consistently referred to as R. 

 

New question input can further be divided into three categories: neutral question, ambiguous 

question and limiting question. A neutral question input is either an open question, a 

question from the interview guide or a slight adjustment to an interview guide question 

which keeps the interview guide question open and neutral.  

 

Example of neutral new question input: 

In(4): Yes, I see, I see. If I leave the subject and get us a little closer to the 

end [laughter] for not keeping you too long since you are busy. Eh, what 

would you say based on your, probably quite long experience in purchasing, 

what are the main parameters for a procurement to be successful in the end, if 

you’d say 3-4 things that are the most important parts? 

 

Ambiguous questions contain e.g. use of jargon, undefined terminology, are long or in other 

ways complicate the interpretation of the response given by the respondent, or results in 

respondent’s inability to understand the question to provide a useful contribution. 

Ambiguous questions include, but are not limited to, use of phrases such as ‘early 

involvement’ and ‘procurement’ without any definition being established.  

 

Example of ambiguous new question input: 

  In(2): Yes, exactly  



 

29 
 

I am thinking, for example, I may just take an example of something that has 

come up. Sometimes you set requirements on something and then you 

wonder a year later why that requirement was important, once you are out 

there and about to do something about this with a contract, for example, or 

whatever it may be. And there is nobody really knowing who owned the 

requirement from the beginning and who is really behind it at all, such 

problems can easily arise afterwards if you do not have, somewhere, 

documented it. If there is some rotation of people and so on. Do you 

recognize yourself in this type of problem or you might not have that kind of 

challenge?  

R: Yes, I recognize it in some way. But I can’t say that I think it feels like 

there has been a problem because I don’t recognize it. 

 

The third and final type of new question input is limiting question, which involves leading 

phrases, giving alternatives or asking a closed question. 

 

Example of limiting new question input: 

In(3): Right  

Does that mean that you’re not involved that much in centralized purchasing 

but more in decentralized purchasing decisions and purchasing processes?  

R: Well, we think that sometimes it’s good to be big and sometimes it doesn’t 

matter. [...] 

 

Follow-up questions are used as the interviewer aims to further probe the respondent’s 

previous answer. However, follow-up questions can be either limiting or exploring 

depending on how such an input is constructed. Limiting follow-ups are characterized by the 

interviewer asking to clarify particular areas of the previous response, whereas explorative 

follow-ups are neutral and includes asking how, why questions and the like. Below are 

examples of each type 

 

Example of limiting follow-up question: 

R: It could be both, for example a new product, or an existing product that we 

want to look at whether  

Do we have the right price on this one, the right contract and so on, then one 

can start procuring.  

Out(1): But when there is a new product [thinking] how do you proceed, you 

said it was the concept development where, uh, are you involved and set 

requirements? 

 

Example of explorative follow-up question: 

R: [thinking] The difficult part is collaboration, between stakeholders to 

really set a strategy, that the project is clearly defined and what it consists of 

and what the project should contain and what roles we will have in the 

project. And also to set goals, what is a successful project?  

One really has to discuss the objective and that you have a number of 



 

30 
 

milestones in the project to see that it’s right in both time and cost.  

Out(3): And how do you do it? [laughter] 

 

Statements are input that is either true or false and could therefore be used to test the 

respondent’s willingness and ability to confirm or reject such an input. Therefore, statements 

are categorized based on the space given to confirm or reject. 

 

Example of statement with space to confirm: 

R: No, first there is a seller who says one thing and then a technical manager 

who says something else, and then there is someone who should pay for it at 

last. There are people working on the “floor”, of course, who have a say 

regarding the work environment. As a large company, we have to behave in 

certain ways according to authority and not let out water and emissions, so 

there are a lot of things that must work in order for it to be good. 

Out(2): Yes there are many people with various interests 

R: Yes, in the case of a purchase of a machine, I will not say that it affects, it 

is not the right word but it will affect many people. After all, ultimately it is 

an operator that will run a machine, it should be safe and okay for the 

environment. We have an industry that generates a lot of heat [hesitates] so in 

energy terms, if we’re talking heating and such things, our cheapest days are 

when it’s minus 20 outside. Then we run at our cheapest, which requires 

cooling of our production and we need water for cooling and so on, of course 

we want to use municipal water to cool the production, it’s very expensive 

[...] 

 

Example of statement without space to confirm: 

In(3): Exactly, and then naturally, it is managed from a purchasing 

perspective from start.  

Are there also situations where slight larger procurements are initiated from 

the organization, erm, we have found, as a background to the question, an 

interesting little [hesitant] a challenge about where you, where the 

organization has pushed a specification quite far and then it comes to the 

purchasing department a little too late in the process. Is there anything that 

you are aware of or have addressed, or do you not see that as a problem at 

all?  

R: No, we do not see that [sounds doubtful] I don’t recognize it, no I don’t. 

 

Finally, the fourth category of input is interjections, i.e. short units that may or may not 

interfere with the prevailing information sharing. Thereby, interjections can also be limiting, 

which is in the case the interviewer gets excited or interrupts to steer the discussion toward a 

particular part of the response without actually asking a follow-up question. Then, attention 

must be paid to how the data is interpreted when including the interjection input in order to 

ensure that the analysis of the data does not render misleading conclusions.  

 

Example of limiting interjection input: 



 

31 
 

R: Yes, that’s exactly as you say, that the specification can be too directed 

against one supplier for example.  

Or that the requirement specification does not allow having as many options. 

In(1): Ah, and this is where it will be interesting, when the  

Because that’s just, that’s the view many  

But then, there are differences in what you do when you get into that 

situation. [laughter]  

If you reverse the tape, if you simply try to do the best given the timetable. 
 

Neutral interjections on the other hand, may include interruption but without affecting the 

following response, or the input is simply added to confirm the response and encourage the 

respondent’s continued elaboration. 
 

Example of neutral interjection: 

R: It may not always be, how should I put this, the one who has the need who 

sets the technical requirements. They are perhaps more interested in functions 

as such or that you get a really really good product at any cost.  

Out(2): Exactly  

R: We have a reality to relate to, in economical terms, so yes.  

But sometimes it’s a bit [puff] it requires some decisions that we later on have 

to solve, I agree with that. 

 

In the following chapters the categories are used to assess the data, first a static review is 

made, thereafter a review of when the particular input type occurs in the different phases.  

4.2.4 Presentation of qualitative static assessment  

The static part of the qualitative assessment of the results views insider and outsider data as 

two groups of interviews. Then, is it possible to compare the total of each group of 

interviews in order to detect differences on a general level by examining e.g. the use of 

different input types. 

4.2.4.1 Focus of the input 

Obviously, the focus of the input differs depending on what the interviewer seeks to explore. 

Investigating the amount of different focus input used by insider and outsider interviewer 

will reveal potential differences between interviewer profiles.  

 

 

 

Focus Total Interview 1 Interview 2 Interview 3 Interview 4 

Number  

of input 

Personal factual 6,0% 2,2% 9,8% 6,9% 5,7% 9 

Informant factual 52,7% 42,2% 46,3% 48,3% 77,1% 79 

Hypothesis verification 14,0% 20,0% 9,8% 13,8% 11,4% 21 

Problem exploring 27,3% 35,6% 34,1% 31,0% 5,7% 41 

            150 

Table C7: Insider, focus of the input in total 

 



 

32 
 

The data shows that the insider mainly has an informant factual focus in the interviews (52.7 

% of all information-carrying input), followed by a problem exploring focus (27.3 %). The 

four interviews follows more or less the same structure regarding the focus, except from one 

case. In interview number 4, the informant factual focus amounts to 77.1 % of the input, 

which is high/much in contrast to the mean/average value of the three preceding interviews 

(45.6 %). As a result, the informant factual figure is slightly amplified