Master's Thesis, 2011
140 Pages, Grade: 1.0
List of Figures
List of Tables
Table of Attachments
List of Abbreviations
Background and Need
Objective of the Study
Significance to the Field
Thesis Outline and Approach
Introduction and Approach
Entrepreneurship Theory and Entrepreneurial Process
Lean Startup Approach
Notion of the Lean Startup
Key management practices
Setting, case and participant selection
Case study protocol and case study database
Evaluation of the Research Design
Results and Discussion
Lean Startup Approach Survey
Case Study Analysis
Case and participant description
Disparity in productivity
Lean Startup Approach effect on opportunity development progress
Lean Startup Approach effect on team performance
Derived Model Instance
Summary of Major Findings
Appendix A: Research Context
Appendix B: Literature Review
Appendix C: Research Methodology
Appendix D: Results
List of References
This thesis really has been one of the most demanding challenges for me so far. Not only is the subject difficult to define succinctly but not trivially, but also the lack of directly related considerable academic literature further raised the bar. Conducting a thesis in a three months time frame while working at the same time posed many challenges to me as an ambitious researcher striving for the best possible solution to cover both, theory and practice.
Without the support of the following people, I would probably not have been able to cope with the aforementioned and other academic work related aspects. My thesis advisor Anne Huff shared her wealth of experience in social sciences research, guided me through the whole process and inspired me many times with her passion for qualitative data research. Nathan Furr (BYU) was of tremendous help in providing me with research context, acting as a highly knowledgeable mentor and sparring partner at the intersection of early stage entrepreneurship theory and practice - special thanks to John Lohavichan for establishing the connection. Tom Eisenmann (HBS), Henrik Berglund (Chalmers), Dave Charron (UCB) and Tor Gronsund were all part of the very supportive research community I could set up throughout the course of the study and inspired me several times with research ideas, design feedback and research related documents. Cordial thanks also go to Hans-Georg Gemünden (TU Berlin) and Holger Patzelt (TU Munich) for sharing very helpful research papers with me. All survey participants in general and the case study interviewees in particular deserve highest credit for contributing their time and experience to this endeavor. In this context, my heartfelt thanks go to Elizabeth Yin who helped me significantly in getting access to nearly half of the final interviewees and additionally allowed me to interview her in her role as a mentor at the LSM event. Further credits go to the LSM organization team and the judges Brant Cooper and Patrick Vlaskovits for sharing their insights and supporting me in conducting the event related research.
Furthermore, I would like to express my gratitude to Bernhard Doll, Christine Behre and Timo Azadi for general thesis and motivational support and Falko Wendt for his continuous feedback not only with regard to this thesis, but also in the context of many other papers I had to write throughout the course of the Executive MBA.
Last but not least, my heartfelt thanks go to my whole family who had to put up with all my ups and downs and compromise many times. They supported me exceptionally well and helped me in many situations to not give up.
Figure 1. Agility: the short OODA loop (Observe, Orient, Decide, Act)
Figure 2. The customer development process
Figure 3. The Build-Measure-Learn cycle
Figure 4. Data collection overview
Figure 5. Frequency distribution for “most important / difficult” LS principles
Figure 6. Cross-classified table for “most difficult to implement” x “experience”.
Figure 7. Degree of documentation of business model related assumptions
Figure 8. Inductively derived research model instance
Table 1. Different types of purpose for research
Table 2. Overview on the mandatory LS Approach survey questions
Table 3. Overview on the types of triangulation used in this study
Table 4. Tests and case study tactics for research quality assurance
Table 5. Summary of key choices related to the research methodology
Table 6. Revealed categories for “startup development progress measures”
Table 7. Revealed categories for “strategic startup direction decision metrics”
Table 8. Revealed categories for “customer interest measures”
Table 9. Case and participant description
Table 10. Comparison of the startup teams’ productivity differences
Table 11. Illustration of differences in opportunity development success
Table 12. LS Approach effect on opportunity development progress
Appendix A: Research Context
Appendix B: Literature Review
Appendix C: Research Methodology
Appendix D: Results
illustration not visible in this excerpt
Technology entrepreneurship is vital to economic development. Entrepreneurs shape and drive innovation, speeding up structural changes in the economy. They introduce new competition and hence contribute indirectly to productivity (Bosma, Kelley, & Amorós, 2010, p. 12). New technology ventures (NTV) can have positive effects on employment and could rejuvenate many industries with disruptive technologies (Christensen & Bower, 1996). The first two and a half years of a firm’s life have been shown to be the most risky but in case the entrepreneur survives this initial valley of death, the long run chances of failure are rather low (Casson, Yeung, Basu, & Wadeson, 2008). However, NTVs face the highest failure rate among new ventures with a survival rate1 of only 22% in the United States (Song, Podoynitsyna, Bij, & Halman, 2008) compared to an overall survival rate of 45% (Shane, 2008).
More often than not, many startups fail to make money from their original business plan. Yet, in spite of the widespread belief in the benefits that can be derived from business planning, there is little empirical evidence indicating that founders who write a business plan actually create more successful firms (Gruber, 2010). A recent meta-analysis of business planning research revealed that the relationship between business planning and performance tends to be positive, yet contextual factors affect the nature of the relationship (Brinckmann, Grichnik, & Kapsa, 2010). The research community and many successful practitioners (Mullins & Komisar, 2009) call for emphasizing the importance of “business discovering” and deemphasizing “business planning”. Thought leaders in high technology entrepreneurship - with Steve Blank, Eric Ries, John Mullins and Randy Komisar as some of the key protagonists - and leading business school professors even observe an emerging paradigm shift occurring in entrepreneurship (Ries, Eisenmann, & Furr, 2011).
However, many entrepreneurship educators, established companies and even practitioners in early stage startups still refer to a mostly inappropriate product development process where actual customer feedback is received at the very end of the process only with often highly unsatisfactory results. Consequences typically are a rejected product in the case of an established company and - even more dramatic - a product built by a startup that no real customer actually wants and, at the same time, all money burned and a high risk of costly failure or even company closure (Ries et al., 2011).
Unfortunately, many consumer Internet high-tech startups in the early stage still face the challenge of learning about the market and customer needs in a fast, systematic and cost-efficient way, which affects their chances to survive in a highly uncertain, resource-constrained environment. For software driven technology startups, the market risk is often higher than the technology risk (Vlaskovits, Foundora Interview, 2011) - especially if commodity technology is leveraged (e.g. FOSS2 ). As young startups typically have limited resources - people and money among the most crucial ones - making effective progress often depends on the ability to first find “Problem/Solution fit”3 and subsequently “Product/Market fit” in the shortest time possible at the least cost. Although intuition plays a role in entrepreneurship as well, a lean and agile process not only complements an entrepreneur’s intuition, but also helps to fight other typical challenges observed by business angels or venture capitalists like the team’s “lack of focus”, “culture of slowness” or the “fear of flying” (Beckord, 2009).
Opportunity development is at the heart of the entrepreneurial process and the transformation of ideas into opportunities is vital for entrepreneurial success. Be it in an official ideation session as part of a brainstorming workshop with the goal to create new ideas or just spontaneously while walking through a park, plenty of ideas come to people’s mind every day. Sometimes these ideas are really new, sometimes they solve problems, but most of the time they simply make sense. The creation of ideas is hence typically not the key challenge in entrepreneurship. However, how to approach the evaluation of a business idea in the very early stage with high complexity and uncertainty, and develop it towards a viable opportunity systematically is still considered a significant challenge for many practitioners and in need of further research in the academic world (Tang, Kacmar, & Busenitz, 2010).
Most entrepreneurs launching businesses run out of cash faster than they are able to bring in customers and achieve sales profitability - especially those who start up for the first time. Whereas unsuccessful entrepreneurs usually equate an idea with an opportunity, successful entrepreneurs can differentiate between the two (Timmons & Spinelli, 2009). In this sense many entrepreneurs and especially less experienced ones are confronted with the following enduring questions: Are we making any progress? Is what I am working on actually creating value? Is the idea worth to be pursued any further? What should I work on next? (Ries, 2011a). Only few realize that in the very early stage of entrepreneurship, not only the problem is unknown (who is actually the customer?), but also the solution (which features should the product have?) and confuse following a plan with progress.
Other startups fell in love with their initial business model based on untested hypotheses and finally failed due to a lack of customer demand for their products or services. Potentially they burned through their capital, wasted precious time and money on engineering and marketing before realizing that actually no one wants their product (Ries et al., 2011). Hence, quickly evaluating whether the idea has serious potential to become a viable opportunity and deciding how much time and effort to invest is vital for an entrepreneur (Timmons & Spinelli, 2009). Yet traditional approaches like business planning and the waterfall model to solution development are not very appropriate in a highly uncertain context where both the problem and the solution are unknown often leading to a venture created based on assumptions about what customers want (Ries et al., 2011).
In order to reduce risk and increase the odds of success, startups are in need of an adequate process to find the right opportunity in a fast, systematic and learning-oriented fashion.
Treating entrepreneurship as a context only in which existing organizational theories apply ignores critical differences between startups and mature organizations and ultimately led to an inappropriate transposal of existing theories into entrepreneurial settings - for example, strategic planning reemerged as business planning (Furr & Cavarretta, 2011). Following a planning-driven approach with a business plan full of untested assumptions and “selling” strict following the plan as development progress is a common mistake that can cause harm and become counterproductive to the endeavor. Some emerging research and discussion began to highlight those contradictory effects and suggested that existing approaches may not be accurate to tackle entrepreneurial challenges (Bhide, 2000; Kirsch, Goldfarb, & Gera, 2009; Ries et al., 2011; Wasserman, 2003).
A more radical evolutionary approach that deemphasizes planning in favor of an extreme set of learning-focused, iterative processes more akin to design and engineering sciences (Blank, 2005; Eisenhardt & Tabrizi, 1995; Gruber, 2010; Ries et al., 2011) seems especially appropriate in high-tech early stage entrepreneurship. The Lean Startup Approach (LSA) fits that profile and combines fast, iterative development methodologies with customer development principles (Cooper & Vlaskovits, 2010) leveraging commodity technology in order to unleash creativity, eliminate waste and accelerate the time-to-market of a product or service (Ries, 2011b).
Yet there is very little research on the relatively young notion of the Lean Startup (LS) and the few published papers focus primarily on the lean product development process aspects of the LS methodology. Therefore a research initiative to further increase understanding of the LSA with a special emphasis on the early stage of opportunity development appeared very attractive and valuable for practitioners as well as research and education.
The research objective of this study was to explore how the startup idea development progress was experienced by entrepreneurs and how the Lean Startup Approach affected that progress in early stage high-tech entrepreneurship.
The LS movement has gained a lot of traction not only in the Silicon Valley area, but also to some extent in Europe. However, with increasing popularity, criticism starts to get formulated as well. Due to the limited coverage of the LSA, its principles and related management practices in academic literature, it remains difficult to critically assess the validity and effectiveness of the proposed concept. With this study, light is shed on the LS phenomenon by interviewing representative high-tech entrepreneurs and leading proponents of the LS principles and practices.
In order to better understand LS principles at work with an emphasis on experienced startup idea development progress, semi-structure interviews with seven Lean Startup Machine (LSM) event participants as well as one mentoring subject matter expert were conducted. The interviewed practitioners were purposely selected from the most and least performing teams according to the appraisal of the LSM event judges. The main reason for choosing the LSM event is the fact that the LSM projects, while somewhat artificial, do provide condensed versions of real life startups, which provides a high information to effort ratio to me as a researcher - especially in the light of a fixed three-months part-time Executive MBA master thesis study.
Furthermore, a LSA survey tailored to the opportunity development process and a post-LSM-event survey were conducted. The first survey had three primary intentions. First, to gather the LSM event participants and highly interested parties’ view on the LSA, issues in early stage high-tech entrepreneurship and especially their view on development progress and metrics. Second, to source for potential interview candidates. Third, to ensure an up-front availability of socio- demographic data to allow for interview sampling and controls for the analysis.
The post-LSM-event survey was mainly conducted to gather the participants’ immediate feedback after the event, source further potential interview partners who did not participate in the LSA survey and finally, gather data with regard to perceived development progress in order to have an internal basis for later validity and quality assurance (triangulation).
The purpose of the multiple mini-cases study (which treated each team working on an entrepreneurial idea as a mini-case) was to explore the opportunity development progress and the role of LS principles and practices from the perspective of consumer Internet high-tech entrepreneurs and LS proponents (mentors) in a highly time-constrained real world LS practitioner event in the USA. Learning about how participants of the event perceived and experienced development progress and the LSA provided insights into the relevance of the LS methodology as a means to accelerate the development progress of an early stage startup idea towards a viable opportunity.
Typical for flexible design studies (Runeson & Höst, 2008; Robson, 2002) and certain types of case study research, the main research question was developed iteratively over the course of the study. By the end of the study the clear question was:
How does the Lean Startup Approach affect the opportunity development progress of an early stage high-tech entrepreneurship endeavor? The research sub-questions of this study included the following: (I) What explains the disparity in productivity between the most and least productive teams at a recent Lean Startup practitioner event? (II) To what extent does the Lean Startup Approach affect the progress of the startup idea development towards a viable opportunity? (III) To what extent does the application of the Lean Startup Approach affect the team performance? (IV) Which of the Lean Startup principles are the most difficult to implement in the eyes of the practitioners? Why is that the case?
Propagated by leading practitioners in the field of technology entrepreneurship, the Lean Startup has evolved a momentum that could have a significant impact on how companies are built, funded and scaled. Initially centered around the Silicon Valley, today the concept reaches far beyond the USA with so called Lean Startup meetups around the world (see Appendix A1: Practitioner movement: Lean Startup groups and meetups for a recent overview on the current worldwide recognition) and many LinkedIn4, Google or Xing Lean Startup groups for major US cities, Spain, Denmark and even Germany.5
However, there has only been very limited research on early stage entrepreneurship in general (Zott & Huy, 2007) and Lean Startups specifically according to the few researchers in this field (see Appendix C1: List of all interviews: DI01 - DI04). The question of whether ventures that follow the LS methodology are on the whole better off remains to be proven. Hence, this study makes a little contribution to the gap between theory and practice. In a recent practitioner discussion related to the question whether the Lean Startup idea is currently degenerating into meaningless buzzwords, Tristan Kromer (see Appendix A3: Lean Startup research and practitioner community overview) cut right to the chase of the matter:
This is the beginning: Lean Startup as proposed by Eric Ries is incredibly young. Customer development as proposed by Steve Blank is a bit older, but still in its infancy. There simply hasn't been enough real research on the subject to say if companies which follow this methodology are on the whole better off. Ultimately, the success of the method will have to be proven by real data if it wants to avoid being just another "Who Moved My Cheese?" diatribe. It has to produce real results.6
In this section, terms that need a more detailed explanation are defined in order to ensure a common understanding when used throughout the course of this study. The selection of the defined terms followed three rules recommended by Bui (2009): (1) terms unknown to persons outside of the field, (2) terms “coined” by their users, (3) terms that are ambiguous to the reader because the definition of the term depends on the context or the interpretation of a participant. As a result of the selection process, the following terms were defined:
Entrepreneur - “A person who habitually creates and innovates to build something of recognized value around perceived opportunities.” (Bolton & Thompson, 2004, p. 16).
Startup - A startup is a temporary organization designed to search for a repeatable and scalable business model according to Blank’s definition (as cited in Lohr, 2011).
Lean Startup Approach - An approach that combines fast, iterative development methodologies with customer development principles (Cooper & Vlaskovits, 2010) leveraging commodity technology in order to unleash creativity, eliminate waste and accelerate the time-to-market of a product or service (Ries, 2011b).
Lean Startup - A startup that adheres to the principles and pursues practices defined in the Lean Startup Approach (own definition). Throughout the whole document the proper name is used whenever a text refers to the coined term. Problem/Solution fit - A state where a problem is recognized by prospective customers (Maurya, 2010) and a strong, tested hypothesis for the right market segment is given (Cooper & Vlaskovits, 2010). In other words, the opportunity is desirable, feasible and viable (see the following definition of opportunity below). Product/Market fit - A state where a product can demonstrate strong demand by passionate users representing a sizable market (Cooper & Vlaskovits, 2010). Minimum viable product - A product with the smallest set of features necessary to achieve a specific objective, and users are willing to pay in some form of a scarce resource - whereas paying with a scarce resource can mean time, money or attention (Cooper & Vlaskovits, 2010).
Opportunity - Guided by IDEO’s Human-Centered Design process7 terminology and building on work of Timmons & Spinelli (2009), superior entrepreneurial opportunities are characterized by four fundamental aspects:
- Desirability: Significant and perceivable value is created or added to an end user or customer. Value is provided by a solved problem, a removed pain point or a met want or need for which someone is willing to pay.
- Feasibility: Legal, technical and organizational feasibility is given.
- Viability: Sustainable value to potential stakeholders can be estimated and communicated based on robust market ($50 million, 20% growth rate), margin (40% gross margin) and money-making characteristics (early and strong free cash flow, 10-15% profit after tax, etc.).
- Suitability: Good and authentic fit with the founder or team at the time and market-place going hand-in-hand with a sufficiently attractive risk-reward balance.
Opportunity development progress - The degree of validated learning about business model uncertainties (e.g. customer segment) in the context of a startup idea development towards a viable opportunity (own definition). “Validate learning” is a key principle of the Lean Startup Approach and is further described in the literature review section.
The conceptual framework used in this thesis is based on a typical approach of scientific research (Bui, 2009) and tailored to the time-frame and focus of a three months Executive MBA master thesis. Although the thesis deals with a highly practical topic, a link to the academic world and relevant theories was established whenever applicable and appropriate. See Appendix A4: Visual overview on thesis structure and approach for a graphical representation of major components of the thesis and the general research design.
With regard to the outline of the thesis, several reporting structure proposals for a master thesis in general (Bui, 2009; Robson, 2002) and case study reports in specific (Runeson & Höst, 2008) were taken into account and, where necessary, adapted to the needs of this research topic.
The first introductory chapter is very important as it provides the rationale for this thesis and narrows a quite broad problem in the field of entrepreneurship - high failure rates and risks in high-tech entrepreneurship - down to a research problem analyzed in this study - systematically reaching problem/solution fit in situations of high uncertainty. By highlighting relevant background literature and prevailing gaps, defining the actual objective of the study and providing a well fitting main research question, a decent structure is provided for the following chapters to build upon.
The second chapter deals with the most relevant literature in the area of entrepreneurship theory and the entrepreneurial process as well as the Lean Startup Approach with its principles and practices.
The third chapter discusses the research methodology with a motivation for the research strategy chosen, followed by the actual research design in the context of a multiple mini-case study and two surveys. After an evaluation of the design, the chapter concludes with a summary of the main choices made.
The fourth chapter first highlights and discusses the results of the LS Approach survey and commences with a detailed presentation and discussion of the key findings of the case study analysis organized around emerging themes and tailored to the main question and subquestions of the research study.
Last but not least, the fifth chapter concludes with a summary of the key findings, an interpretation towards the respective impact and implications, a critical assessment of the limitations and finally, a suggestion for future research.
This chapter provides the reader with the most relevant and significant literature and research findings in two areas related to the facilitation of opportunity development in the early stage of high-tech entrepreneurship. The first part will address research related to theory of entrepreneurship and the entrepreneurial process of opportunity development. The second part focuses on the Lean Startup Approach and its related principles and practices with links to relevant theory where applicable. Before the actual literature review results are discussed, a brief elaboration on how the review was conducted is presented in the following three paragraphs. Readers not interested in the process might consider starting with the first topic section on the next page right away.
The literature review process typically provides several major benefits. First, knowledge is gained about the research that has already been conducted in the field that is proposed in the study. This does not only include historical and seminal theories, but also the most recent cutting-edge studies. Second, the review can reveal new ideas for the research study and hence serve as inspirational source. Third, conducting the literature review allows seeing how the research endeavor fits into the literature that already exists (Bui, 2009).
A systematic process and several tools and strategies were used to conduct an effective and thorough review. Both, primary and secondary sources were considered and a key terms based search was conducted using Internet search engines (primarily Google Scholar8 ) and credible, electronic databases (primarily Springerlink, Emerald Management, Business Source Premier, EconLit, SSRN). Furthermore, recent issues of A-rated electronic journals according to a recent VHB9 ranking of technology and innovation management related journals (Entrepreneurship Theory and Practice, Journal of Business Venturing and Journal of Product Innovation Management) were scanned to identify related quality research (see Appendix A2: Top 10 VHB ranking of technology and innovation management journals).
Last but not least, informal research community sessions (see Appendix A3: Lean Startup research and practitioner community overview) were used to exchange relevant literature findings, confirm gaps and also discuss potential research purposes and questions (see Appendix C1: List of all interviews where interviews DI01-DI06 were among others also used for the purpose described above).
Although many entrepreneurship scholars have attempted to develop theory in the field of entrepreneurship, there is still a lack of consensus about the “ingredients” of entrepreneurship and a general theory of entrepreneurship has not become accepted (Ricketts, 2008; Alvarez, 2005). A major reason for this lack of consensus is the absence of clarity that entrepreneurship scholars have about the unstated assumptions of entrepreneurship (Alvarez, 2005). From a different angle, Furr and Cavarretta (2011) consider the historical development of organizational theory and the field of entrepreneurship itself responsible for lacking theory and propose an organizational theory of entrepreneurship where entrepreneurship is not just a “context”.
In their line of thought, entrepreneurship viewed as a “context” leads to an inappropriate application of existing organizational theory and traditional, Sloan- style management tools by ignoring the significant differences between entrepreneurial and established organizations. Whereas in established and typically large organizations the primary focus is on the exploitation of known opportunities with efficiency and control, in entrepreneurial settings the focus is on exploration of potential opportunities with agility and satisificing (Furr & Cavarretta, 2011). The act of business planning as such and the business plan as a document are representative examples of large company theory and practice reapplication to approach and describe entrepreneurial problems. What were strategy planning and the strategy blueprint document in the large firm environment became business planning and the business plan in the entrepreneurial context. Other examples are the transformation of corporate marketing into entrepreneurial marketing, and often the traditional product management approach became the process for developing a startup towards a business (Blank & Furr, 2011; Furr & Cavarretta, 2011).
As a result of this simple transposition of existing theory, harm may be caused in entrepreneurial settings and some emerging research and discussion has started to reveal contradictory effects suggesting that existing approaches may not be accurate in such an environment (Bhide, 2000; Kirsch, Goldfarb, & Gera, 2009; Ries et al., 2011; Wasserman, 2003).
Yet the business plan, for example, still plays a dominant role in startups, entrepreneurship education and external financing environments (Lange, Mollov, Pearlmutter, Singh, & Bygrave, 2007). An estimate of 10 million written business plans per year10 exemplifies this statement. The main reasons are partly related to the founders’ belief in their value and partly related to the aspect that founders are made to write a business plan - in example to attract external financing from a bank or venture capitalist (Gruber, 2005; Karlsson & Honig, 2009). Timmons and Spinelli (2009) recognized the dynamic nature of a startup in the early stage and argued that the business plan provides the language and code for communicating the quality of the three driving forces of their model of the entrepreneurial process (team, opportunity, resources) and of their fit and balance.
However, although there is a common belief in the benefits that can be gained from business planning and the business plan as such, there is little empirical evidence indicating that those entrepreneurs who write a business plan actually perform better and create more successful businesses (Gruber, 2010). While some researchers see business planning as a vital activity for successful company creation as a decision making support (Shane & Delmar, 2004), other researchers believe quite strongly that planning can actually lead to cognitive rigidities and consider the time invested for the planning as a kind of waste where time could be better spent on actual company and customer relationship building (Bhide, 2000). In a recent meta-analysis regarding the relationship between business planning and performance relationship a beneficial effect was indicated, yet contextual factors such as the newness of the company and the cultural environment significantly affect the relationship. In a context of high uncertainty and a limited information base - typical for the early stage of high-tech entrepreneurship - basic, rather than sophisticated business planning is advised. In the general conclusion, a more dynamic approach that combines planning, learning and doing was suggested (Brinckmann et al., 2010).
From a different angle, Eisenhardt and Tabrizi (1995) found a similar result when investigating strategies for accelerating the development of products under uncertainty. They concluded that under high uncertainty, the use of prototypes, rapid iteration, and sense making through direct contact were more valuable than a "compression" strategy in which well-understood links in a system were squeezed together. In other words, variance-creating strategies were more valuable than mean-enhancing strategies (McGrath, 2001).
With regard to the understanding of key success factors for startup development from initial exploration stage to later exploitation phase, two elements have been identified that lead to a better identification of opportunities by entrepreneurs: prior knowledge and “alertness” (Kirzner, 1973). In a recent study, Tang et al. (2010) further specified the meaning of alertness and provided a validated 13-item alertness scale as instrument to empirically measure this construct in a rigorous manner. Their resulting definition of alertness consisted of three distinct elements: scanning and searching for information, connecting previously disparate information, and making evaluations on the existence of profitable business opportunities. In a summarizing view on recent development in the area of alert evaluation and judgment, the combination of beliefs, insights and desires for a judgment on the prospects of the new venture were mentioned. Yet how entrepreneurs should deal with beliefs that might be leaps of faith was not covered.
Furthermore, a significant relationship between the entrepreneurs’ prior knowledge11 and alertness was noticed. However, how this knowledge is acquired is not explained. Likewise, why some entrepreneurs are more alert than others and whether a certain opportunity development process could compensate insufficient prior knowledge and alertness to some extent is not known. Ardichvili, Cardozo and Ray (2003) showed that there is a weak relationship between opportunity identification and personality traits. Some researcher speculated that it might be a condition resulting from what an entrepreneur has done that others have not, rather than what she or he is, in example, that there is a path-dependence between alertness and knowledge (Midler & Silberzahn, 2008). An interactive relationship between action and cognition was identified by Crossan, Lane and White (1999) who remarked that understanding guides action, but action also informs understanding.
Two other factors that entrepreneurs mention now and then when asked about entrepreneurial success factors in their opportunity development efforts are - besides the well researched aspect of network - intuition and luck. In a recent large survey, Liechti, Loderer and Peyer (2010) concluded that luck was responsible for less than 17% of the performance whereas toil, experience, talent, education and network were among the more important contributors according to the received responses. Although consistency checks were made, some skepticism with regard to the results is certainly appropriate. Not only related to the data collection method used as such, but also related to the complexity of ensuring a certain objectivity and verifiability when asking about the role of luck in this context. With regard to intuition, Blume and Covin (2011) analyzed whether entrepreneurs actually use intuition or just claim to do so. They suggested that attributions to intuition as noted by the entrepreneurs may have little relationship to the actual use of intuition. Furthermore, they stressed the importance of domain-relevant knowledge and experience due to their direct applicability to opportunities and foundational character regarding the formation of complex knowledge structures from which all true intuitive capability is derived.
Specific to new technology ventures (NTV), Song et al. (2008) researched success factors in a meta-analysis by analyzing the findings of 31 studies and revealed the 24 most broadly researched success factors. Among those, eight were homogenous significant success factors (e.g. homogeneous positive significant metafactors correlating with startup performance) that were suggested as the only universal success factors for the performance of the NTVs: (1) supply chain integration, (2) market scope, (3) firm age, (4) size of founding team, (5) financial resources, (6) founders’ marketing experience, (7) founders’ industry experience, (8) existence of patent protection. Surprisingly, the founder’s experience with startups or research and development was declared as not significant. At the heart of their suggested integrated framework of new entrepreneurial firm performance, strategic and organizational fit with the subitems competitive strategy, structure, processes and systems were mentioned. However, processes were not part of the conducted meta-analysis and leave a gap for further research.
Another important stream in the field of entrepreneurship intensively analyzed how entrepreneurs deal with situations of pure, Knightian uncertainty. Sarasvathy, (2001, 2008) increased our understanding of the entrepreneurial process with a description of two different approaches to the creation of new ventures: causation and effectuation. By inverting the principles of causal reasoning, entrepreneurs constitute a new comprehensive logic based on those inversions that Sarasvathy called “effecutation”. In contrast to causation which is consistent with planned strategy approaches, effectuation is consistent with emergent strategy and primarily means-driven where goals emerge as a consequence of stakeholder commitments rather than vice versa. Furthermore, effectuation includes a selection of different alternatives based on flexibility, loss affordability and experimentation (Chandler, DeTienne, McKelvie, & Mumford, 2009; Sarasvathy, 2001, 2008). Taking a progressive approach to product and market definition by following the effectuation logic suggests success in the context of Knightian uncertainty (Wiltbank, Dew, Read, & Sarasvathy, 2006). Therefore, the question of how those products get created by the venture in absence of a current market for future products is much more important than what kind of products and markets they had chosen ex-ante (Shane & Venkataraman, 2000). In the same context, how this process supports the venture team in resolving the uncertainty appears to be crucial.
Concerning development process models related to early stage high-tech entrepreneurship, three recent case study based research projects revealed interesting findings, but also directly or indirectly highlighted some important gaps in research. Klocke and Gemünden (2010) analyzed the balance of exploration and exploitation processes in different stages of organizational development as well as the speed at which firms reach a certain stage. They proposed a learning based development model with five stages and a separation in technology- and market-related activities. Related to progress, increased speed, and particularly very high speed, was interrelated with high performance and high output (e.g. products, turnover, positive income, etc.). Not only knowledge itself, but also the speed of its creation was mentioned together with challenges caused by speed: e.g. shortened process and decision times or increased failure rates. Related to the model and the scope of this thesis, purely technology-related activities were listed as part of the exploration in the first stage of the model (Klocke & Gemünden, 2010).
In contrast, the customer development model advocated by serial- entrepreneur-turned-educator Steven Blank defines market and product development as parallel processes already in the early stage (Blank, 2005). Further support for this viewpoint was provided by the second case study research where simultaneous product development and market exploration based on lineage management with cumulative learning were recommended for dealing with uncertainty while preserving flexibility to adapt to changing environments (Silberzahn & Midler, 2007). In that context, the importance of focus in uncertain environments got highlighted by pointing out two major benefits: (1) an unnecessary variation is limited and a random walk on products and markets is avoided, (2) an effectual approach of development is supported by providing a product basis upon which the next effectual iteration can build (Silberzahn & Midler, 2007).
Finally, the third case study research was conducted by the same authors and dealt with the question of managing the startup-development process through a series of consecutive exploration projects in high-tech entrepreneurship. By comparing the learning from project to project of two cases where one case followed a portfolio approach without cumulative learning and the other a lineage- based management strategy (cumulative learning, simultaneous exploration and exploitation), Midler and Silberzahn (2008) concluded with the following finding: while the startup in the second case did not arrive at a promising opportunity immediately, the firm converged towards it - thereby reducing the cost for each iteration and increasing the chances of success.
Therefore, a process model that increases speed at sustained low failure rates and supports appropriate decision-making appears promising as a means to mitigate high risk and increase the odds of success for early stage high-tech startups. The Lean Startup Approach represents a promising concept to combine many of the latest findings in research and practice that support the entrepreneur in the process of effectively dealing with uncertainty and limited resources (time, people, and money).
There has only been very limited research on early stage entrepreneurship (Zott & Huy, 2007) and Lean Startups in specific according to my own literature review. A Google Scholar search using “Lean Startup” as key term did not reveal a single search result12 after a manual adjustment with regard to false positives, practitioner books and articles. Leveraging other search techniques and databases, only one published academic case study paper could be identified that deals with Lean Startup principles. Yet it focused solely on the product development related Lean thinking principles and did not cover market or customer related aspects. Discussions with the few researchers in this field (see Appendix C1: List of all interviews: DI01 - DI04) confirmed the identified situation. Besides Nathan Furr’s longitudinal Lean Startup Research project which just started in the second half of 2010, no further completed research initiatives or published papers, theses or dissertations are known. Therefore, the decision was made to strive for academic definitions to the best degree possible based on one unpublished, forthcoming HBR article and the originator’s as well as leading proponents’ view or books on the subject at hand. Nevertheless, links to theory were provided wherever applicable.
The term “Lean Startup” was originally coined by Eric Ries, a Stanford graduate and former student of Steve Blank, serial entrepreneur and popular blogger in September 2008. Despite the fast rise in popularity with more than 14,000 members in Lean Startup Meetup Groups in 57 cities and 13 countries13, the “Lean Startup” has not reached notable attention in the academic world yet as described in the beginning.
According to Ries (2008), a Lean Startup is best described by three characteristics that are also part of a major trend respectively: (1) Ferocious customer-centric rapid iteration, as exemplified by the Customer Development process, (2) The application of agile development methodologies, and (3) The use of platforms enabled by open source and free software. Cooper and Vlaskovits (2010) argue that a fourth element, “the use of powerful, low-cost and easy-to-use analytics” (p. 28), might be appropriate to be added supplementary. Ries, Eisenmann and Furr (2011) define the LS Approach as follows: “rapidly build and test products and then, based on customer feedback, quickly refine promising concepts and ruthlessly cull the flops”.
As the name “Lean Startup” suggests, the concept is an application of Lean Thinking. Originally based upon the Toyota Production System (TPS), lean thinking strives for systematic identification and elimination of waste in manufacturing and administrative processes through a process of continuous improvement by flowing the product and service at the pull rate of the customer (Widman, Hua, & Ross, 2010). Lean is defined as the process for doing more with less and less - less human effort, less equipment, less space, and less time - while getting better and better at providing the customers exactly what they want (Womack & Jones, 2003).
Whereas short production cycles are used to reduce inventory and eliminate waste in manufacturing, short product development cycles are used to gain rapid market feedback and eliminate waste in development process and capital expenditure. Besides lean thinking, other concepts are borrowed from areas such as design thinking, customer development, agile development (Ries et al., 2011) and the OODA (Observe-Orient-Decide-Act) loop. However the difference between Lean Startups and the narrow application of these ideas is that the LS methodology synthesizes these principles with novel practices employed by successful entrepreneurs to develop the first beginning-to-end model for building a new business (Ries et al., 2011). While the customer development and agile development processes are rather considered as management practices to implement the LS principles (Ries et al., 2011) and are hence discussed in a following section, the ideas behind OODA deserve particular attention due to their formative influence on the Lean Startup thinking model.
The OODA loop, an innovative time-based theory, was originally pioneered by military strategist and U.S. Air Force Colonel John Boyd who later became a professor at the U.S. Military academy (Marchisio, Shepherd, & Woods, 2010). Boyd’s basic idea is that every decision-maker repeatedly executes a cyclic process of observation, orientation, decision and action which is commonly referred to as the OODA loop:
illustration not visible in this excerpt
Figure 1. Agility: the short OODA loop (Observe, Orient, Decide, Act)14
Often reduced to an over-simplified conclusion of “the shorter OODA loop always wins”, Osinga summarized Boyd’s main ideas in his recent book about science, strategy and war as follows: “Life is full of complex adaptive systems. Such systems require understanding at the system level and are perpetually changing. Therefore, our methods of gaining understanding must be perpetually adapting, as well.” (as cited in Samuelson, 2010, p. 36). Applied in a competitive business environment, a simplified operation of an OODA cycle could be described as follows: “decision makers gather information in the Observe phase, they filter this information in the Orient phase and then make Decisions (hypotheses) and take Action. The cycle is repeated continuously as the organization works to execute rapid OODA loops.” (Marchisio et al., 2010, p. 10).
This kind of agility requires a continuous cycle of interactions with the environment, constant assessment of change as well as means to mitigate risks and finally, iterating faster than the competition to yield a substantial advantage - in the case of a startup potentially the survival in the initial years where scarce resources and high uncertainty can be best compensated with a mindset of agility. The validity of the OODA loop as a useful strategic tool in a business context has been advocated by Osinga (as cited in Marchisio et al., 2010, p. 9), and just recently an explicit fit with the nature and characteristics of an entrepreneurial context was declared (Marchisio et al., 2010). In concert with the findings of Furr and Cavarretta (2011), Marchisio et al. (2010) advise against assuming strategy and strategic thinking as it has been understood in large enterprises applies to an entrepreneurial context, but rather suggest considering different tools, frameworks and heuristics with OODA as a particular framework of choice.
The interactive, non sequential process of the OODA loop provides the necessary flexibility in making critical decisions required in unpredictable, constantly changing environments therewith assuring long term survival (Richards, 2004). The ambidextrous character of the framework that offers not only a useful heuristic tool for entrepreneurs, but also a useful way of thinking about strategy, entrepreneurial action and the decision making progress (Marchisio et al., 2010) appears to be well suited as an appropriate guiding theme of the Lean Startup Approach. Baron (2004) describes the OODA loop as a heuristic that can act as a “mental aid for making fast but accurate decisions” and Alvarez and Barney (2002) argue that the extensive application of heuristics by entrepreneurs allows them to more readily navigate through a wide array of problems and irregularities inherent in the development of new opportunities. The attainment of knowledge in this way is an intangible asset that, given its rareness among business leaders, may be a source of competitive advantage for entrepreneurs (as cited in Marchisio et al., 2010, pp. 14-15).
According to Ries, Eisenmann and Furr (2011), there are four principles that characterize the Lean Startup Approach: (1) Validate learning, (2) Pivot as necessary, (3) Iterate rapidly, and (4) Avoid premature scaling. Although Ries refers to five, partly different principles of the Lean Startup on the landing page of his upcoming book (Ries, 2011b), the decision was made to refer to the ones presented above as they were elaborated for a research-based magazine (Harvard Business Review) and are characterized by a more neutral language. Furthermore, there is substantial overlap when it comes to the description and the bottom line appears to be the same. Hence, the four principles mentioned above are now briefly described to ensure a common understanding:
Validate learning. As startups face significant uncertainty, a mindset of relentless experimentation, learning, and adaptation is required in order to raise the odds of success (Gruber, 2010). Lean startups experiment relentlessly and strive for frequent market feedback to alleviate the degree of uncertainty. By structuring experiments and rigorously confirming or disproving hypotheses about uncertain parts of the business model, this process finally yields validated learning - the effective unit of progress in a Lean Startup. Many entrepreneurs struggle to know when they are actually making progress in resolving those areas in their business model where they assume the highest uncertainties. Conventional wisdom defines a new product development effort as successfully making progress as long as monetary budgets and deadlines are met. However, potentially launching a product that no real customer actually wants is not considered good management in the definition of entrepreneurship management as presented in the beginning of this chapter. Rather than following a plan and processing output to meet the requirements of traditional project management milestones, a Lean Startup processes input in the sense of validated learning that can be leveraged in the next Build-Measure-Learn cycle (Ries et al., 2011).
Iterate rapidly. There are two kinds of waste that need to be taken into account: (1) activities that do not generate direct value, but are currently necessary to create value, (2) activities that do not generate value and could be eliminated immediately (Blank & Ries, 2010). To avoid waste, fast and iterative development cycles are practiced along with customer development methodologies in order to validate core hypotheses around the customer, the problem and the solution (Cooper & Vlaskovits, 2010). By developing a series of minimum viable products (MVPs), learning cycles are additionally accelerated. While Ries, Eisenmann and Furr (2011) define the MVP as “the smallest set of features necessary to secure the next round of validated learning” (p. 3), Cooper and Vlaskovits (2010) chose an even more customer-centric definition: “A product with the fewest number of features needed to achieve a specific objective, and users are willing to pay in some form of a scarce resource” (p. 26) - whereas paying with a scarce resource can mean time, money or attention. Striving for “just enough” features avoids not only the risk of wasting time on features that none of the customer wants, but also the impairment of experimental designs. Regarding the latter aspect, an evaluation of a potential product rejection by customers is hampered in such a way that it may be difficult to determine whether the product got rejected due to concerns about the MVP feature set, or just about the unnecessarily added “extra feature” (Ries et al., 2011).
Pivot as necessary. By iterating rapidly and testing hypotheses, Lean Startups seek for validated learning. A refuted business model hypothesis leads to a new pivot. As a result, affected elements of the business model are changed - typically one element only - while other elements are retained to avoid waste (Cooper & Vlaskovits, 2010; Ries, Eisenmann, & Furr, 2011). Depending on the magnitude of the pivot, one can differentiate the frequent form of small pivots from large pivots that are likely to appear occasionally only (Ries et al., 2011). Furthermore, Maurya (2010) argued that there is a difference between pivots and optimizations which depends on the stage the startup is currently in. Whereas the focus before Product/Market fit is on validated learning by means of pivoting, the focus changes to growth via optimizations after the fit. Product/Market fit, in its simplest form, can be described as a state where a product can demonstrate strong demand by passionate users representing a sizable market (Cooper & Vlaskovits, 2010).
Avoid premature scaling. A low-burn mentality where investments in marketing, product development and infrastructure (predominantly software and hardware) are limited is another means to eliminate waste. Before Product/Market fit is reached, investments in the aforementioned areas are restricted to the minimum extent necessary to allow for fast and continuous cycles of validated learning.
However, there are two exceptions to this rule: (1) long lead times in the deployment of capacity and significant preemption risk, (2) hyper-growth models where typically viral network-effects lead to hyper-growth of user base. Twitter, for instance, took four years to unveil the first stage of its business plan with so called Promoted Tweets to finally generate more than US$150-million in advertising revenue in 2011 according to an industry estimate (Hartley, 2011). Until the end of 2009, Twitter had yet to figure out what an attractive monetization model around the significant user base could look like. These kinds of startups had to scale before the business model was validated in order to reach a user base size that is appealing to ecosystem partners who then could participate in experimentation to identify effective ways to monetize the platform. For those hyper-growth models, the recommended approach is as follows: (1) seek for a path to reach viral network-effects, (2) aggressively follow the discovered path to continuously increase the user base and thereby form partnerships, (3) acting in concert with the partners, strive for a viable monetization model, and (4) aggressively pursue the identified model (Ries et al., 2011).
Although the described path might appear different from the Lean Startup Approach, those startups that successfully pursued it still adhered to the principle of rigorous hypothesis testing to reach validated learning through rapid iteration cycles and pivoting (Ries et al., 2011).
The Lean Startup Approach combines two main management practices to pursue the principles described in the preceding section (Ries et al., 2011): (1) Customer development, (2) Lean and agile product development. Both practices are briefly described in the following paragraphs to ensure a sufficient understanding to comprehend details of the research design, analysis and discussion. For both topics, the section first starts with an overview and key components (tenets, process/approach) and ends with empirical case study research support where existent and directly relevant.
Customer development. According to Silvernagel and Clement (2010), the customer development model is a conceptual model that frames innovation and new product development in terms of two focus areas: technical and market feasibility. Developed by Steven G. Blank, a serial-entrepreneur-turned-educator, the model is based on a four-step approach and is characterized by the following three central themes: (1) product development and market development are parallel processes - in contrast to the traditional thinking of serial processes where market development follows product development, (2) “bipolar” conceptual design with a non-linear thinking and an interplay of technical and market feasibility as a line with technical feasibility on one end and market feasibility on the other, (3) elimination of waste by repeated validation of the entrepreneur’s assumptions about the problems of customers and the potential solutions by communicating with the customers (Silvernagel & Clement, 2010; Ries et al., 2011).
On a very abstract level, two quite simple but powerful tenets characterize the customer development model: “Question your assumptions” and “Talk to your customers (stakeholders)” (Cooper & Vlaskovits, 2010). An important aspect to consider is that adhering to those two tenets is the centerpiece of the model and the process is a means to an end as well as a guiding framework to support the systematic practical execution.
The four steps of the customer development process are depicted in the following graphic, whereas the first two are most relevant for lean startups (Ries et al., 2011):
illustration not visible in this excerpt
Figure 2. The customer development process15
As visualized in the figure above, the first two phases and especially Customer Discovery were in the scope of this study.
Despite the huge popularity of the customer development concept, no directly related empirical research could be identified which further underlines the importance of a research endeavor in this field.
Lean and Agile Development. In a nutshell, Agile Development is a light- weight method that works with short, iterative development and feedback cycles and involves the customer and other important stakeholders tightly in the
software development process. The most popular Agile (Software) Development
methods are eXtreme Programming16 and Scrum17. Lean Development, in essence, is about applying Lean principles to software development from an end- to-end perspective and putting a strong focus on eliminating waste from the development process in order to optimize customer value contribution (Petersen, 2010).
Although Lean Development and Agile Development are two different paradigms, Petersen (2010) concluded in a thorough analysis of the respective goals, principles, practices and processes: (1) Agile and Lean share the same goals, (2) Lean is Agile as principles of Lean reflect principles of Agile, (3) Lean has adopted many practices known in the Agile context, while putting emphasis on using practices that are related to the end-to-end flow. Agile uses practices that do not exist in Lean (See Appendix B2: Comparison of Lean and Agile principles for the complete comparison).
Another perspective is provided by Robert Charette - the originator of Lean Development - who sees the key difference between Lean and Agile in Lean representing a top-down approach and Agile a bottom-up approach (Highsmith, 2002). In the context of this thesis, Agile Development is considered as a supportive practice in the context of a Lean Development philosophy and treated as one resulting unit while bearing the distinction in mind.
In contrast to the plan-driven and sequential waterfall model of new product development, Lean and Agile development is learning-focused by employing many short iterations of the development cycle and getting customer feedback as early as possible in the process (Ries et al., 2011). This is in line with earlier findings of Eisenhardt and Tabrizi (1995) who investigated strategies for accelerating product development under uncertainty and concluded that under high uncertainty, the use of prototypes, rapid iteration cycles, and sense making through direct contact were more valuable than a "compression" strategy where well-understood links in a system were squeezed together.
The mantra can be summarized as: turn ideas into a product (MVP), measure the product against reality (customers) and learn for the next iteration (validated learning) in the fastest possible way by balancing the two extremes “Release early and often” and “Maximize the chances of success”. The challenge is to find out about the minimum set of features necessary to engage with those early evangelists to start the learning feedback loop - the so called Build-Measure- Learn cycle:
illustration not visible in this excerpt
Figure 3. The Build-Measure-Learn cycle18
The key practices can be summarized as follows (Ries et al., 2011):
Short Cycles via MVPs. To keep cycles short and get feedback quickly, only few features are added to an iteration. Right from the beginning, high-fidelity prototypes are tested rather than working versions of the envisioned final product. Low priority features are held for the product’s hypothetical version 2.0.
Intensive Communication. Face-to-face communication is emphasized, often co-locating all development functions in a shared project room (war room concept). Close, ongoing collaboration with internal customers (e.g. product managers) is valued higher than relying on detailed planning and contract negotiations. Feedback is appreciated and change is embraced regarding the specifications.
Metrics as necessary. While measurement is a key component of the BuildMeasure-Learn cycle, metrics are only created where important to increase learning. The rule is to use as few as possible in order to reduce waste (adapted definition). To guide product development decisions, measurements are chosen that are actionable, accessible, and auditable.
Commodity Technologies. Whenever possible, existing and proven technology, in particular open source software, is used. Likewise, partners are leveraged who can provide shortcuts and fixed costs are variablized to save precious time and capital.
Five Whys. With its origin in the Toyota Production System, the Five Whys question-asking method strives for the identification of a root cause based on the assumption that behind every supposedly technical problem, there is actually a human problem. While traditional TPS would focus on fixing the root cause, the Lean Startup Approach advocates to only proportionally invest at each of the five levels. Only when problems reoccur again and again, more and more time is invested in the investigation. Thereby, the adapted Five Whys technique acts as a natural speed regulator and balances “going to fast” with “overinvesting in problem solving” (Ries, 2010).
Continuous deployment (CD): Complimentary to Five Whys, CD simultaneously reduces the batch size and increases the work tempo. Thereby development teams reduce waste in the process and stay in a condition of flow for sustained periods. Attaining and keeping that state facilitates innovation, experimentation and achievement of sustained productivity.
Last but not least, two empirical case studies could be identified that relate to the subject at hand. The first study is a practitioner case of the Finnish Lean Startup Huitale19 that implemented a Lean (Product) Development process with a predictable workflow within acceptable variance. According to Taipale (2010), Huitale implemented nearly all practices mentioned above - only “Intensive Communication” was not mentioned specifically - and could achieve the following results: lead times for new features could be predicted accurately, the business model could be adapted to the changing market needs accordingly, software updates could be released daily and the quality level became extraordinary high (only two production bugs in the past three years). Finally, as part of a lessons learned statement, discipline was stressed as key factor on executing, implementing and improving the Lean and Agile Development process. Workflow visualization, root cause analysis, customer value metrics and measurement of lead and cycle times were claimed beneficial for continuous improvement and transparency.
The second study is an academic research case where IMVU - the Lean Startup of Eric Ries which pioneered the Build-Measure-Lean cycle - was analyzed with regard to the application of Womack and Jones' (2003) five Lean Management principles in software development processes. Widman, Hua and Ross (2010) concluded that IMVU could successfully implement the Lean principles at the technical level and identified as well as eliminated common wastes - specifically overproduction, waiting, defects and process. Furthermore, a chaotic, frequently changing process could be turned into a more predictable, fast moving and streamlined process. As the focus of this case study was primarily on the technical level and the case was conducted in the company of the person that coined the term Lean Startup, further research with unrelated companies and a focus on the customer and business related aspects appears to be highly valuable. The case study research in this thesis tried to start with filling that gap.
An emerging paradigm shift towards a new theory of entrepreneurship sets out to increase the odds of entrepreneurial success and reduce the still relatively high failure rate of high-tech startups. The Lean Startup methodology promises to play an important role in this happening. In the absence of empirical data and academic research, a better understanding of the Lean Startup Approach and its potential impact on startup idea development towards a viable opportunity is needed.
The following main research question was addressed in this study that evolved during the study iterations - typical for flexible design20 studies (Runeson & Höst, 2008; Robson, 2002): How does the Lean Startup Approach affect the opportunity development progress of an early stage high-tech entrepreneurship endeavor? Furthermore, the following sub-questions were defined: (I) What explains the disparity in productivity between the most and least productive teams at a recent LS practitioner event? (II) To what extent does the Lean Startup Approach affect the progress of the startup idea development towards a viable opportunity? (III) To what extent does the application of the Lean Startup Approach affect the team performance? (IV) Which of the Lean Startup principles are the most difficult to implement in the eyes of the practitioners? Why is that the case?
This chapter now discusses the research strategy in general and the research design in specific. Therefore it is divided into four sections and starts with an elaboration of the research strategy chosen and the related motivation. The second section details the research design by covering the most important aspects such as unit of analysis, setting/selection of site, sample/participants, measurement instruments, data collection, data analysis and ethical considerations. The third section highlights key tests and tactics considered to evaluate the design. Finally, section four concludes with a summary of the key choices made related to the research methodology.
In order to make an informed decision on the research strategy, the different types of purposes for research based on Robson’s (2002) classification need to be considered respectively (Runeson & Höst, 2008):
illustration not visible in this excerpt
Table 1. Different types of purpose for research21
The research objective and the formulation of the main research question as a “how” question speak for an exploratory research. However, the pre- and post- study survey and interview elements as well as the formulation of one sub- question (“What explains …”), could lead to a categorization as a hybrid.
According to Woodside (2010), “case study research is an inquiry that focuses on describing, understanding, predicting, and/or controlling the individual (i.e. process, animal, person, household, organization, group, industry, culture, or nationality)” (p. 16). Therefore, inducting theory using case studies constitutes a useful research strategy for the analysis of a phenomenon in its context and is especially appropriate in new topic areas (Eisenhardt, 1989). In concert with the research questions and the scarce knowledge as well as empirical data available in this research field, a multiple mini-case study analysis was conducted, studying themes and patterns in polar type cases of very early stage startup teams applying the LS Approach in a real-life context. The term mini-case is used to account for the resource constraints of this enquiry with regard to time, especially for data collection, and researchers involved.
Similar or comparable multiple case study designs have been used in new topics and new theory areas (Zott & Huy, 2007; Midler & Silberzahn, 2008; Lettl, Hienerth, & Gemuenden, 2008) where the focus is on understanding the “how” and “why” in unexplored research areas and where the objective is to further develop an emerging theory (Eisenhardt & Graebner, 2007). In order to not only ensure greater validity in the development of insights, but also consider context dependency, the more robust multiple case studies with across-case comparisons were favored over a single case study (Eisenhardt, 1989; Yin, 2009).
The general research process followed is characterized as flexible (Robson, 2002) where the design “unfolds” and key parameters of the study are subject to be changed during the course of a study (Runeson & Höst, 2008). Furthermore, an integrated multi-method approach around an educative, real-world oriented Lean Startup practitioner event was used that included both quantitative and qualitative data collection - although the focus was predominantly on the qualitative side. The approach consisted of four key elements which are described in more detail in the following research design related sections: (1) an online survey (started before the LSM Boston event), (2) the Lean Startup Machine Boston (LSM-BO) event, (3) a post-LSM-event online survey (shortly after the event) and most importantly, and (4) the post-LSM-event semi- structured interviews (retrospective: one to two and a half weeks after the event).
The integration of quantitative and qualitative approaches was not only pursued for data and method triangulation in order to reduce threats to validity (Robson, 2002), but also to prepare and complement the third part, semi- structured phone interviews with event participants. The interviews were conducted one to two weeks after the event to allow for a reflected view on the LS methodology as experienced at the event, complemented and where applicable challenged by the survey findings in order to better understand the LS methodology and its related principles in practice. Although not typical, surveys may be conducted within a case study and are well accepted as means for triangulation (Runeson & Höst, 2008). The narrative data were completely transcribed in a way that is appropriate for this type of research and the given time frame, coded, and categorized into themes related to the research questions.
In a nutshell, the research design is the logical sequence that links empirical data and the conclusions to be drawn to the study’s research questions (Yin, 2009). In other words, the design constitutes a blueprint, chain of evidence, or logical model of proof. It strives for maximum construct validity, internal validity, external validity and finally, reliability (Robson, 2002).
The research design followed in this thesis is described in the following paragraphs and deals with five methodological issues: (a) setting, case and participant selection, (b) data collection, (c) data analysis, (d) case study protocol and case study database, and (e) ethical considerations. Finally, the research methodology chapter is wrapped up with a summary of the key choices in this context. All aspects were presented in a way so that other researchers can potentially replicate this research (Bui, 2009; Yin, 2009).
Setting, case and participant selection.
Although the case study approach used is flexible, a certain level of planning is still appropriate. Good planning for a case study is even considered to be crucial for success. Among the most important issues to be planned are the methods to be used for data collection, which site to visit, what documents to study, which people to interview, how often these interviews should be conducted, etc. (Runeson & Höst, 2008). A common approach to formulate these plans is to use case study protocols (see section Case study protocol and case study database. in this chapter).
The overall study was designed around an educational competitive Lean Startup practitioner event in Boston, Massachusetts that took place from Friday, February 25th (6pm) until Sunday, February 27th (7pm) 2011 at the Microsoft New England Research & Development Center. The event - called the Lean Startup Machine (LSM) - is tailored to entrepreneurs who want to learn the Lean Startup methodology and its related principles through real-word problem solving in a competitive environment:
At Lean Startup Machine, you’ll learn how to figure out what customers really want. In 48 hours, you’ll form a team with other participants and work on a new startup while competing for cash, mentorship and glory - Lean Startup Machine will change the way you think about building startups (“Lean Startup Machine / BOSTON,” n.d.).
The initial plan had foreseen to be on site as well and additionally observe the participants at work. However, the research related “observation allowance” was provided only a few days before the event so that administrational22 and research related preparatory tasks were not manageable anymore.
The event began on Friday evening (6 pm) with networking and food, followed by presentations about the Lean Startup methodology (7-8:30 pm) and finally, idea related project pitches (8:30-10 pm) from half of the participants out of 50 event participants in total.23 After the pitches, the top 10 ideas were selected by voting (participants, judges and mentors) and teams were formed around the people who pitched the top 10 ideas (first checkpoint). After the first out of three checkpoints in total, the event proceeded with team work in each of the different teams to progress the ideas to the best degree possible to finally pitch to a panel of experienced and knowledgeable entrepreneurs for a cash prize and additional mentoring (see Appendix C7: LSM Boston event idea, schedule, speakers, judging criteria and resources for a visual depiction of the aforementioned and further information).
The LSM projects, while somewhat artificial, provide condensed versions of real life startups, which allow for a high information to effort ratio to me as a researcher. Furthermore, the event setup requested all participants to start from scratch with an idea selected by the judges and all participants by voting. All the participants got the same input information on the Lean Startup Approach and had only a very restricted time frame of 56 hours overall to achieve a meaningful result - a business idea that is developed towards a viable opportunity with the most customer validation possible.
Theoretical sampling of cases was used to illuminate the phenomenon of interest, assist in ruling out alternative explanations and enable the elaboration of the emergent theory (Eisenhardt & Graebner, 2007). Particular emphasis was put on a theoretical sampling approach that increases the likelihood to reveal unexpected, critical elements and impediments of startup idea development progress as the single unit of analysis, while simultaneously allowing deeper exploration of anticipated influence factors. To achieve these goals, a “polar types” sampling approach was chosen where extreme cases (e.g. very high and very low performance according to judges) were sampled in order to ease a potential observation of contrasting patterns in the data and support a very clear pattern recognition of the central constructs, relationships, and logic of the explored phenomenon (Eisenhardt & Graebner, 2007).
The original plan was to select two to three startup teams from the best and least performing teams according to the judges’ assessment and interview two team representatives per team. Quite surprisingly, it was much harder to recruit interviewees from the top three teams (finalists in the competition) so that finally “only” one case could be selected in group A (high productivity), but five cases in the group B (low productivity). Although this is certainly a limitation and not ideal, the decision was made to stick to the approach as additional data about the other finalists could be gathered via the post-LSM-event survey as well as one mentor interview (see section Data collection.). Furthermore, a first analysis of the cases was expected to reveal more insights to form a solid decision basis.
The mini-cases are briefly described in the section Case and participant description. of the chapter Results and Discussion. Team and idea names have been anonymized and taken from a top 10 list of the greatest inventors (last name only).24
1 Survival defined as still in operation after five years with more than five full-time employees (Song, Podoynitsyna, Bij, & Halman, 2008, p. 7).
2 Free open source software.
3 See the section Definitions in this chapter for a brief explanation of this and the following term. 2
4 LinkedIn operates the world’s largest professional network on the Internet with more than 90 million members in over 200 countries and territories according to their “about” page information accessed on 2011/02/15.
5 See https://www.xing.com/net/pri0aac0fx/lean-startup.
6 Source: Quora question ("At this point is the 'Lean Startup' idea degenerating into meaningless buzzwords?, n.d.).
7 IDEO’s free Human-Centered DesignToolkit can be downloaded at: http://www.ideo.com/ work/ human-centered-design-toolkit/.
8 Google Scholar (http://scholar.google.com/) is a service provided by Google to conduct a general search for academic literature and comprises many areas and sources (theses, papers, etc.).
9 VHB stands for: Verband der Hochschullehrer für Betriebswirtschaft e.V.
10 According to David E. Gumpert’s estimate on written business plans on a global scale in the context of his research for the book „Burn your business plan!: What investors really want from Entrepreneurs (2003).
11 About markets, serving those markets and customer problems or needs. 12
12 The last search was conducted on 2011/03/29. Using „Lean startup“ or „Lean start-up“ as key term and „Since 2008“ as further criterion led to 10 or 8 unadjusted search results.
13 See Appendix A1: Practitioner movement: Lean Startup groups and meetups for an overview of the distribution around the world.
14 Taken from Samuelson (2010).
15 Modified, taken from Blank (2005), Cooper & Vlaskovits (2010) and Ries, Eisenmann, & Furr (2011).
16 See http://www.extremeprogramming.org/ for detailed information.
17 See http://www.scrum.org/ for further information. 22
18 Taken from Ries (2011b).
19 Huitale Ltd. is a provider of Lean and Agile consulting services and a product development company that hosts a national consumer portal (Nextdoor.fi) with 30,000 unique visitors per month and 2,000 active users.
20 The notion of flexible design is briefly explained in the following section Research Strategy. 25
21 Modified, taken from Robson (2002).
22 Flight, hotel, admission/entry permit, etc.
23 All data taken from the event homepage and additionally validated as well as complemented via mentor and participant interview data.
24 Modified, taken from http://www.toptenz.net/top-10-greatest-inventors-in-history.php (Accessed: 2011/03/27).
Master's Thesis, 130 Pages
Master's Thesis, 135 Pages
Doctoral Thesis / Dissertation, 142 Pages
Research Paper (postgraduate), 15 Pages
Master's Thesis, 206 Pages
Term Paper (Advanced seminar), 26 Pages
Scientific Essay, 17 Pages
Thesis (M.A.), 30 Pages
Seminar Paper, 23 Pages
Term Paper (Advanced seminar), 22 Pages
Master's Thesis, 32 Pages
GRIN Publishing, located in Munich, Germany, has specialized since its foundation in 1998 in the publication of academic ebooks and books. The publishing website GRIN.com offer students, graduates and university professors the ideal platform for the presentation of scientific papers, such as research projects, theses, dissertations, and academic essays to a wide audience.
Free Publication of your term paper, essay, interpretation, bachelor's thesis, master's thesis, dissertation or textbook - upload now!