Jur 12

Ghezzi, Antonio
Digital Startups and the adoption and implementation of Lean Startup Approaches:
Effectuation, Bricolage and Opportunity Creation in practice
Article in press
Forthcoming in Technological Forecasting and Social Change
Please cite as:
Ghezzi, A. (2018). Digital Startups and the adoption and implementation of Lean Startup
Approaches: Effectuation, Bricolage and Opportunity Creation in practice. Technological
Forecasting and Social Change, https://doi.org/10.1016/j.techfore.2018.09.017
Antonio Ghezzi*
Politecnico di Milano, Department of Management, Economics and Industrial Engineering, Via
Lambruschini, 4/B, 20156 Milano, Italy.
E-mail : antonio1.ghezzi@polimi.it
* Corresponding author
Short Vitae
Antonio Ghezzi, Ph.D. is Professor of Strategic Management and Head of the Hi-tech Startups applied
research center at the Department of Management, Economics and Industrial Engineering – Politecnico di
Milano. His main research field is Strategy and Digital Entrepreneurship, with a focus on startups' business
model design, innovation and validation, and the role of Lean Startup Approaches. He is author of more than
seventy refereed journal articles (appearing in outlets such as Technological Forecasting and Social Change,
International Journal of Management Reviews, Journal of Business Research, International Journal of
Production Economics, Management Decision and R&D Management), books, book chapters and
conference proceedings.
1
Ghezzi, A. (2018). Digital Startups and the adoption and implementation of Lean Startup Approaches: Effectuation, Bricolage and
Opportunity Creation in practice. Technological Forecasting and Social Change, https://doi.org/10.1016/j.techfore.2018.09.017
Digital Startups and the adoption and implementation of Lean Startup Approaches: Effectuation,
Bricolage and Opportunity Creation in practice
Abstract
Digital startups launching original value propositions can test out and validate their business model using a
recent and emerging set of practices known as Lean Startup Approaches (LSAs), which consist of Customer
Development and the Lean Startup. While LSAs are gaining momentum in the ecosystem surrounding digital
startups, they still suffer from poor theoretical foundations and operational issues that hinder their adoption
and implementation. The aim of this study is to go beyond recalling anecdotal and single-case examples and,
through mixed-methods research involving 227 digital startups, provide a first large-scale analysis of: (i) if
and how digital startups apply Lean Startups Approaches; (ii) the ensuing results; (iii) the main advantages
and disadvantages that stem from adopting and implementing LSAs; and (iv) how digital startups connect
and combine the LSAs with other entrepreneurial approaches and tools for launching startups. The findings
reveal that most of the sample has adopted LSAs and obtained several benefits from their use. A list of
practical guidelines on how to solve the existing drawbacks and enhance the effectiveness of adopting and
implementing LSAs is hence proposed. To conclude, a framework for organizing the empirical findings is
put forward, where LSAs are inserted into the entrepreneurship theory debate on Effectuation,
Entrepreneurial Bricolage and Opportunity Creation. Suggestions are then provided on how to sequence and
bridge effectuation and causation logics and decision-making tools in an “entrepreneurial opportunities
space”.
Keywords: Digital Startups, Digital Entrepreneurship, Business Model, Lean Startup Approaches, Minimum
Viable Product, Pivoting, Entrepreneurial Opportunity, Effectuation, Bricolage
1. Introduction
“There are three principal means of acquiring knowledge... observation of nature, reflection, and experimentation.
Observation collects facts; reflection combines them; experimentation verifies the result of that combination”.
(Denis Diderot, 1753)
1.1. Reasoning behind the study
Entrepreneurial endeavours carry with them an abundance of activities that entrepreneurs must perform in
order to acquire tangible and intangible resources (Barney, 1991; 2001) and transform them into new
ventures or startups. As the French philosopher Denis Diderot (1966) wrote, there are three principle means
of acquiring knowledge available to us: observation of nature, reflection and experimentation. These mirror
what is also required from entrepreneurs, as they need to observe their surroundings closely when trying to
2
identify potential opportunities (Shane and Venkataraman, 2000). Once an opportunity has been detected,
these entrepreneurs must think hard about how to combine their resources creatively and originally (Ireland
et al., 2003) so that they become heterogeneous – i.e. they create value for customers, yet are rare and costly
to imitate (Barney, 2001) – and on how to mould a value proposition and a startup around this unique pattern
of resources. After this, entrepreneurs must test their design (Kerr et al., 2014) to see whether their startup is
likely to create, deliver and capture value is a way that will be accepted by their target customers and,
therefore, be viable. This final action is needed to validate the overall startup business model, in other words,
its architecture of value (Rappa, 2001; Teece, 2010, Ghezzi et al., 2015).
This part is particularly complex within the highly volatile context of digital entrepreneurship (Nambisan,
2017; Autio et al., 2018), where opportunities are quick to arise and vanish, combined resources
unexpectedly acquire or lose their original value, and testing becomes a vital feature of the startup’s learning
process (Contigiani and Levinthal, 2018).
Digital startups launching their products and services in this setting can use a recent and emerging set of
practices to test out and validate their business model, known as Lean Startup Approaches (LSAs), which
consist of Customer Development (Blank, 2007; Blank and Dorf, 2012) and the Lean Startup (Ries, 2011).
These pragmatic methods for transferring the notions of Lean Manufacturing (Womack and Jones, 1996) to
the startups’ environment are built upon a set of key principles: (i) formulating falsifiable hypotheses about
the business idea; (ii) embedding these hypotheses into a designed business model (BM); (iii) developing a
Minimum Viable Product (MVP) to replicate the product’s functions and test the business model; (iv)
identifying the target audience or “earlyvangelists” from whom to receive feedback; and (v) running tests
with multiple iterations to make decisions about persevering with the business idea, letting it perish or
pivoting, that is, undertaking a structured course of correction (according to the “build-measure-learn”
feedback loop).
Although these approaches are gaining momentum within the startup community (Breuer, 2013; Kerr et al.,
2014; Kiura et al., 2014; Frederiksen and Brem, 2017), and sometimes are even penetrating the corporate
world as they are adopted by incumbents (Ries and Euchner, 2013; Powers, 2014), they still lack the
necessary theoretical foundations (Ghezzi and Cavallo, 2018; Yang et al., 2018), with specific reference to
3
the extant theories on Entrepreneurship. Moreover, as the largely practitioner-oriented literature on LSAs
builds on recalling past cases and anecdotal evidence, little is known concerning: (i) the actual adoption and
effectiveness of these approaches in terms of assisting in the digital startups’ launch; (ii) their connection to
other traditional entrepreneurial tools such as business plans; and (iii) their inclusion within more established
entrepreneurial theories and logics.
1.2 Purpose and research questions
This study hence sets forth as its objective to go beyond recalling anecdotal and single-case examples and
offers a first large-scale analysis of four aspects. These are: if and how digital startups apply Lean Startups
Approaches (Research Question 1); the ensuing results (Research Question 2); the main advantages and
disadvantages they feel come from the adoption and implementation of LSAs (Research Question 3); and
how digital startups connect and combine the LSAs with other entrepreneurial approaches and tools for
launching startups (Research Question 4).
Using a mixed methods approach, the study builds on the results of a survey on 227 digital startups launched
between 2012 and 2017 contained in an original database; it further informs these results through a
qualitative research method based on 32 semi-structured interviews, to disclose additional insights.
1.3. Contribution
According to the study’s findings, Lean Startup Approaches are largely adopted in the sample of digital
startups investigated, and digital entrepreneurs gain significant benefits from their implementation, mostly
referring to the following aspects: (i) reducing time and costs to test the startup; (iii) aligning business idea to
customer needs; (iii) verifying and pivoting all business model parameters; (iv) receiving rounds of
financing; and (v) offering alternatives to traditional intellectual property protection. When setting up their
digital startups, however, the entrepreneurs found that the work to implement LSAs was not as
straightforward as they had expected, having often underestimated the true complexity of putting in place the
relative processes and tools. More specifically, digital entrepreneurs claim that they encounter issues
concerning: (i) defining and designing MVPs; (ii) identifying and engaging earlyvangelists and trial users;
4
(iii) defining testing priorities and designing experiments and tests; (iv) missing other market opportunities
and threats; and (v) obtaining information about the startup’s sources of competitive advantage.
As a conclusion, this study suggests that LSAs could be positioned in the debate on entrepreneurship as an
operational and scientific approach to the behavioural theories of Effectuation (Sarasvathy, 2001) and
Entrepreneurial Bricolage (Baker and Nelson, 2005), within an Opportunity Creation setting (Alvarez and
Barney, 2007).
These findings could help to properly address and amend the downsides of LSAs, encouraging digital
startups to adopt this promising - but still scientifically under-investigated - approach.
1.4 Structure
This study is structured as follows. After this introduction (Section 1), Section 2 contains the theoretical
background in which the study is positioned (i.e. the discovery and creation theories of entrepreneurial
opportunity and action, and Lean Startup Approaches); the study method is described in Section 3, leading to
the results given in Section 4. These results are discussed in Section 5, and the conclusions concerning the
study’s contribution are drawn in Section 6, together with its limitations and the avenues for future research.
2. Theoretical Background
2.1 Discovery and creation theories of entrepreneurial opportunity and action
Entrepreneurship is commonly defined as the act of discovering and pursuing profitable opportunities (Shane
and Venkataraman, 2000); once such opportunities are spotted, the entrepreneur should move towards them,
to build a new venture or startup around them. In this phase, entrepreneurs engage in “opportunity-seeking”
behaviour, subsequently followed by strategic or “advantage-seeking” behaviour, in order to integrate the
opportunity with competitive advantage (Hitt et al., 2001; Ireland et al., 2003).
While the literature shows consensus on the centrality of the concept of opportunity in entrepreneurship,
different approaches emerge concerning the nature of these opportunities and the process through which they
5
become apparent. As Alvarez and Barney (2007) argue, two alternative theories of entrepreneurial action
indicate that opportunities can be discovered or created.
According to the discovery theory of entrepreneurial opportunity (e.g. Shane and Venkataraman, 2000),
opportunities exist independently of the entrepreneurs, who should nurture alertness (Kirzner, 1973), that is,
the ability to become aware of opportunities and discover them before others. A discovery view of
opportunities hence advocates the use of data collection and analysis techniques and methods framed in a
business plan (Delmar and Shane, 2003), to make decisions in a risky context.
The creation theory of entrepreneurial opportunity takes a different stance (Weick, 1979; Gartner, 1985).
Starting from the assumption that opportunities are not objective phenomena formed by exogenous shocks
and that they do not exist independently of entrepreneurs, the creation theory posits that opportunities are
social constructs endogenously created by the actions, reactions and enactment of the entrepreneurs’
exploring ways in order to create value in an uncertain context (Alvarez and Barney, pp 13-15). With regards
to enactment, the proceedings entail an iterative, inductive and incremental decision-making process,
supported by heuristics and based on the notion of affordable loss rather than expected return.
Different entrepreneurial opportunity theories and decision-making contexts led to the advancement of
parallel approaches and logics behind the entrepreneurial action. Sarasvathy (2001) restates this apparent
dichotomy in her study comparing traditional causation logics with what she calls “effectuation”. While
causation sees the future as a continuation of the past, and entrepreneurial actions as goal-oriented
undertakings that strive to avoid potential contingencies through accurate predictions and careful planning,
Effectuation considers the future as shaped by entrepreneurs, who act on the basis of the means at hand –
rather than the goals – and plan for the unexpected, leveraging contingencies as they arise (Yang et al.,
2018).
As in the case of Effectuation logic, Entrepreneurial Bricolage (Baker and Nelson, 2005) argues that
entrepreneurs operate in a condition of extreme resource scarcity, and have, therefore, to “make do” by
combining the resources at hand and applying them to new problems and opportunities that arise. Making do
6
hence implies making the most of limited resources, by engaging in experimentation, the early and frequent
interaction with customers and subsequent iteration (Fisher, 2002).
Although the logics of Effectuation and Entrepreneurial Bricolage can describe the actual behaviour of
entrepreneurs involved in the process of creating and developing startups (Fisher, 2012), little is still known
concerning about how they can be placed in a systematic operational frame to enhance opportunity formation
through the entrepreneurs’ enactment process. Although recent studies (Frederiksen and Brem, 2017; Yang
et al., 2018) have come up with the suggestion that Effectuation could be put into practice through the Lean
Startup (Ries, 2011), these contributions are still scattered and unconnected, and show either a largely
conceptual stance, or some misconceptions about the different steps included in the Lean Startup and its
precursor, Customer Development (Blank, 2007).
2.2 Lean Startup Approaches
The Lean Startup Approaches (LSAs), named originally by Ghezzi and Cavallo (2018), include two main
methods to develop startups, with enough similarities to couple them within the same framework: Customer
Development (Blank, 2007) and the Lean Startup (Ries, 2011).
Steve Blank (2007) was the pioneer of LSAs when he introduced the concept of “customer development”
and described the process that entrepreneurs should follow to test and refine their business hypotheses
through customer interaction. In his seminal book “The Four Steps to the Epiphany: Successful Strategies for
Products that Win”, he argued that a startup is a temporary organization looking for a scalable and replicable
business model. Therefore, the primary aim of any startup should be to test whether the overarching
hypotheses of its business model are indeed verified, thus making the BM scalable and replicable.
Based on Karl Popper’s concepts of conjectures and refutations (Popper, 2014), which form the core of
epistemology, the customer development approach states that, rather than developing products or services,
startups should develop customers and, to do so, they must go through a process consisting of two steps and
four sub-steps (see Figure 1).
7
The first is the “search” step, when startups need to find “earlyvangelist”, that is, early customers or trial
users with specific characteristics – i.e. they have a problem or need, they are aware of having a problem,
they are actively looking for a solution, they have even tried to put together their own solution, and they have
or can acquire the funds to buy a good solution once they find it (Blank, 2007). After the earlyvangelists
have been discovered, the startups should work closely with them to create iterative feedback cycles that
allow new solutions to be tested and the business idea to be improved incrementally, so that it gains customer
validation.
Testing and the subsequent validation are based upon a set of assumptions and/or hypotheses set by the
startup to postulate how its business idea will work and create value for customers, how value will be
delivered to these customers and how a share of such value will be captured to ensure the startup’s
sustainability. This means that the startup will need to design an overall business model, that is, an
architecture of value (Rappa, 2001; Teece, 2010, Ghezzi et al., 2015), around its idea.
When testing out their business models, entrepreneurs should introduce what is known as the “business
model canvas” (Osterwalder and Pigneur, 2010), according to which three pillars and nine parameters are
assigned to a business model, as shown in Table 1.
Pillar Parameter Description

Key Activities Describes the most important things that a company
must do to make its business model work
Value Key Resources Describes the most important assets required to make a
Infrastructure business model work
Key Partners Describes the network of suppliers and partners that
make the business model work
Value Propositions Describes the bundle of products and services that
create value for a specific customer segment
Customer Segments Defines the different groups of people or organizations
that an enterprise aims to reach and serve
Value Proposition Customer Describes the types of relationship a company
Relationships
establishes with specific customer segments
Channels Describes how a company reaches and communicates
with its customer segments to deliver a value
proposition
Value Formula Cost Structure Describes all costs incurred to operate a business model
8
Revenue Stream Represents the cash that a company generates from
each customer segment (costs must be subtracted from
revenue to create earnings)
Table 1. The Business Model Canvas (Osterwalder and Pigneur, 2010)
During its search phase, Customer Development, therefore, helps the startup to obtain early insights about its
prospective customers. At this point, the startup must prove that there is a potential market for its
products/services and verify that there are customers willing to acquire them.
By introducing this approach, the startup can reduce the likelihood of it failing. This is because the initial
business model and the assumptions on which it was based, which in the main stem from the entrepreneur’s
personal perception of reality, have been tested and possibly verified in the field through repeated interaction
with potential customers, and the learning outcome derived from these tests comes from the real world – in
line with the directive to “get out of the building” advocated by Blank (2007).
The “execute” phase follows the search phase. Once the business model and the customers are validated, the
startup can create its market demand through investments in marketing and sales spending, and it can scale
up its business by structuring its functional units and introducing customer development teams (Blank and
Dorf, 2012).
Figure 1. Customer Development (Blank and Dorf, 2012).
9
The Lean Startup, a term coined by Eric Ries (2011), revisits Customer Development and consists of a
business approach where the aim is to change the way companies are built and new products are launched.
As with Customer Development, the formulation of falsifiable hypotheses plays a critical role in triggering
the Lean Startup process. The two key assumptions made by entrepreneurs are: (i) the value hypothesis –
testing whether a product or service really delivers value to customers once they are using it; and (ii) the
growth hypothesis – testing how new customers will discover a product or service (Ries, 2011; Hartman,
2013 – p 6).
The Lean Startup introduces two new concepts: Minimum Viable Products (MVPs) – artefacts used for the
fast and quantitative market testing of a product or product feature – and pivots, where certain BM elements
are changed in response to failed hypotheses tests. According to Ries, a MVP is “that version of a new
product which allows a team to collect the maximum amount of validated learning about customers with the
least effort” or, in a simpler way, “the smallest thing you can build that lets you quickly make it around the
build/measure/learn loop” (Ries, 2011).
Unlike other methods for managing early stage ventures, the lean startup approach also balances the strong
direction imposed by a founder’s vision with the need for redirection following market feedback (Eisenmann
et al., 2012), thereby introducing a third element.
In addition, when Ries describes LSAs, a central part is given over to the Build-Measure-Learn feedback-
loop, which in turn is influenced by the Observe, Orient, Decide and Act (OODA) loop developed by the
military strategist John Boyd as a tool for winning battles. Behind the feedback loop is the idea that
entrepreneurs should get their products into the hands of customers – in the form of a MVP – as fast as
possible in order to test it out and receive feedback that can be used to reject or validate their assumptions.
The goal of Lean Startup is to minimize the time spent going through the feedback loop, the implication
being that the startups need to build faster, measure themselves faster and learn faster (Ries, 2011).
Experiments and tests within the loop can also be iterated, as startups should formulate and validate their
problem hypotheses, create and validate their solutions to this problem, and conclude by validating the
10
remaining elements of the business model built around the solution, as well as seeing how to scale it up and
acquire customers (Gustafsson and Qvillberg, 2012; Heiramo, 2013).
Figure 2 depicts the process faced by an entrepreneur when following the Lean Startup approach.
Figure 2: Lean Startup Process Steps (Eisenmann et al., 2012)
Customer Development – with specific reference to its “search” phase – and the Lean Startup together make
up the Lean Startup Approaches (LSAs), used by entrepreneurs when validating their digital startup business
model. Ries (2011) and Blank (2013) both concur that, while traditional methods for New Product
Development largely focused on testing products, services and value propositions (Brown and Eisenhardt,
1995; Krishnan and Ulrich 2001), LSAs should help to experiment on all elements of the startups’ BM.
Recent attempts were made to frame the Lean Startup and Customer Development in the academic literature,
with somewhat unconnected or partly inconsistent results. Frederiksen and Brem (2017) were involved
conceptually in looking for theories that possibly back the Lean Startup, while not addressing the advantages
and disadvantages of applying the method. Yang and colleagues (2018) investigated the entrepreneurial
reasoning behind the Lean Startup – which they divide into the phases of search and execute – and find that
11
search actions are associated to effectual logics, while execution actions are instead associated to causation
logics. Nevertheless, in their hypotheses, they equate the Lean Startup with Customer Development, as
search and execution are phases of the latter rather than the former. Contigiani and Levinthal (2018) carried
out research into the Lean Startup’s theoretical roots and antecedents, but largely without mentioning
entrepreneurial theories. Ghezzi and Cavallo (2018) discussed the Lean Startup and Customer Development
jointly because of their similarities and partial overlapping, introducing the notion of LSAs. They relate
LSAs to Agile Development and Business Model Innovation, although any detailed discussion of LSAs
having roots in Entrepreneurship theory lay outside their scope and aim. Despite the growing interest in
LSAs, extant studies still show only limited consensus about the origins of these approaches and their related
tools; moreover, scholars often adopt a conceptual stance (e.g. Frederiksen and Brem, 2017; Contigiani and
Levinthal, 2018) that needs to be backed by empirical investigation.
3. Methods
This study is based on a mixed-method approach where quantitative and qualitative studies follow one from
the other. (Johnson et al., 2007; Castro et al., 2010). In mixed-methods research, qualitative methods are
often employed to help explain and expand quantitative findings (Molina-Azorin, 2012; Creswell and Clark,
2011).
More specifically, this study uses a sequential approach by combining a quantitative questionnaire-based
survey on a sample of digital startups (step 1) with qualitative semi-structured interviews on a sub-sample of
digital startups representing polar or highly relevant cases – identified by their answers to the questionnaire
(step 2).
3.1 Survey on Digital Startups
In the quantitative step of this research, our aim was to collect large-scale information to address the
following broad research questions: if and how do digital startups apply Lean Startups Approaches (RQ1);
what are the results obtained (RQ2); what do they perceive to be the main advantages and disadvantages
12
stemming from the adoption and implementation of LSAs (RQ3); and how do digital startups use the LSA in
conjunction with other entrepreneurial approaches and tools for launching startups (RQ4).
The sample was obtained from an original database of digital startups based and operating in Italy that
received funding from formal investors – i.e. independent venture capital funds, corporate venture capital
funds, government venture capital funds – and/or informal investors – i.e. business angels, angel networks,
crowdfunding platforms, family offices and venture incubators (e.g. see Bruton et al., 2010, 2015; Kerr et al.,
2011) between 2012 and 2017. This database was set up within a broader research project and, since 2012,
has been mapping Italian-funded digital startups by triangulating data from all actors in the entrepreneurial
ecosystem – e.g. formal and informal investors, incubators and accelerators, governmental institutions,
startups associations and the startups themselves.
Formal and informal investors mostly differed in terms of whether the organization backing them was
structured or largely unstructured, as well as whether they used a systematic or unsystematic approach to
carry out the startup’s due diligence before making an investment (Drover et al., 2017). We decided to focus
on funded startups and include both sources of equity capital financing because this meant that our study
looked at high-quality startups that had passed either a formal or an informal due diligence process, allowing
us also to potentially contrast and compare the different roles that LSAs play in enabling or hindering the
funding process.
The original database, as of January 10th 2017 (when the survey was launched), held 484 digital startups that
were still active and were or had been funded by formal and/or informal investors.
To collect data, an e-mail based survey was sent to the personal or business e-mail addresses of a
representative digital startup founder, which had previously been collected and archived in the original
database, together with information about the contact person in the founding team to be interviewed for
research purposes. It was important to interview the startups’ founders since they were the people behind the
business ideas to which the LSAs were being applied, and they had control over the development processes
enabled by LSAs.
13
A pilot survey was submitted to three digital startup founders who were willing to provide feedback on its
structure and the clarity of the questions. This kind of feedback allowed us to revise the order of our
questions and the wording of some of them, as well as tweaking the list of items to be included as alternative
options, to come up with the final version of the survey.
This final survey was broken into four sections of questions covering the adoption of LSAs, the results, the
advantages and disadvantages, and their integration with other entrepreneurial theories, approaches and tools.
The full list of survey questions is given in Appendix 1.
Where the answers to the questions were not of a Boolean yes/no type or chosen from a set of items, they
were given on a four-point Likert scale, where the mid-points choices were eliminated in order to polarize
results and help us to select the startups best suited to a qualitative interview (Garland, 1991).
The initial personalized contact and subsequent telephone follow-ups ran up to March 15th 2017 (Porter and
Whitcomb, 2003; Baruch and Holtom, 2008), at which point we had received 235 answers, giving a survey
response rate of 48.55%. After cleaning this list by removing incomplete answers, the remaining 227
complete records was used as the basis for our quantitative data analysis.
The respondents belonged to various fields of application in the broad digital industry, which also spans a
number of traditional markets (Bharadwaj et al., 2013), and the sample was well-representative of the
database’s whole population in terms of industry and average startup age. Table 2 shows the distribution of
the sample across different fields of application.
Field of application Number of % of sample Average age (in years)

startups
Mobile Applications and Services 49 21.59% 4.1
Internet of Things 38 16.74% 3.6
eCommerce and Retail 29 12.78% 3.8
Industry 4.0 26 11.45% 2.4
Big Data 23 10.13% 3.1
Social Media 19 8.37% 3.5
Mobile Payment 13 5.73% 3.3
Cloud Computing 10 4.41% 3
Machine Learning 8 3.52% 1.5
Security 7 3.08% 1.8
Digital Wearables 5 2.20% 2.2
Total 227 100% 3.3 years (weighted average)
14
Table 2. Sample characteristics
Given the rather explorative nature of this first analysis on a wide sample of digital startups engaged in LSA
practices (Eisenhardt, 1989), the data analysis mostly relied on descriptive statistics to show the
aggregations, means and trends that could be further explored through qualitative analysis. On this point,
quantitative analysis provides the means to inform theoretical sampling in qualitative analysis (Molina-
Azorin, 2012; Creswell and Clark, 2011): as such, by leveraging on the contribution by Johnson and
colleagues (2007) on the qualitative-quantitative continuum, this mixed methods research can be defined as
“qualitative dominant”.
3.2 Interviews with significant digital startups
The qualitative step of the research consisted in gathering insights that could help to explain and expand the
findings gained from RQs 1, 2 and 3, while further exploring how to address RQ4, which covered the LSAs
and how they connected with other entrepreneurial approaches and tools.
This step was designed as a set of semi-structured qualitative interviews (Yin, 1984; Eisenhardt, 1989;
Eisenhardt and Graebner, 2007), starting from a protocol of questions similar to those of the survey (see
Appendix 2), although these questions were more explorative and open to the informants going off topic or
including any new aspects, thus allowing innovative issues to emerge from the open discussion (Walsham,
1955; Yin, 1984).
Case sampling was performed theoretically (Meredith, 1998; Eisenhardt, 1989) and was informed by the
outcome of the quantitative research step. The cases were selected according to their heterogeneity in terms
of: (i) within-case consistency or inconsistency of answers to the survey; and (ii) cross-case polarities and
significant differences.
By carrying out multiple qualitative interviews, it was possible to generalize the results more firmly
(Meredith, 1998) and undertake a comparative analysis of the findings, because the theoretical sample
potentially contained extreme cases, polar types or niche situations (Meredith, 1998). Hence, in order to
15
include the digital startups and associated informants in the qualitative research, they needed to show a high
level of consistency or inconsistency in terms of their internal positioning towards the adoption of LSAs, or
strong similarities or differences when compared to other startups.
The face-to-face, semi-structured interviews were, therefore, the primary source of information when
collecting data in this research step. Between May and October, 2017, 51 digital startups which had
answered the questionnaire were contacted to ask for a follow-up interview. Of them, 37 startups agreed, and
32 of them were eventually interviewed (14% of the sample of respondents, and 62.7% of the subsample of
startups contacted in the qualitative research step).
The interviews were held in a semi-structured format with the digital startups’ founders and lasted on
average 1 hour and 8 minutes.
For the data analysis, the interviewees’ responses were first recorded and fully transcribed. If any
information was still unclear and/or more data was needed, the informants were later contacted by telephone
to ask for clarification.
Then, following the recommendations set by Eisenhardt (1989), a within-case data analysis was carried out,
so as to generate the necessary insight into the issues under scrutiny. A subsequent cross-case analysis made
it possible to compare the various responses from informants belonging to the different digital startups.
Interview content analysis was performed for both the within-case and the cross-case analyses, borrowing the
use of open coding from Grounded Theory methodology (Glaser and Strauss, 1967; Strauss and Corbin,
1998), as this method is suitable for studying complex phenomena as it applies a clearly defined procedure
based on coding – i.e. using labels, concepts and words to produce theory from interviews, rather than the
mere finding of facts (Glaser and Strauss, 1967). The empirical material was codified through textual
analysis, and stored in a software archive.
In accordance with Eisenhardt (1989), as virtually no research can be based on a completely “clean
theoretical slate” (p. 536), coding was guided by the few extant studies where there was the explicit attempt
to frame LSAs in theory (Frederiksen and Brem, 2017; Contigiani and Levinthal, 2018; Ghezzi and Cavallo,
16
2018; Yang et al., 2018). The literature streams seen as plausible antecedents or theoretical roots for LSAs
were used as the lenses through which to interpret the empirical results of the interviews.
“In vivo” and constructed codes were collected for each interview (Glaser & Strauss, 1967), recording the
exact wording used by informants to describe the processes for adopting and implementing LSAs, as well as
the constructed wording introduced by the researcher. By contrasting and comparing the content of different
interviews iteratively, we were able to identify any idiosyncratic positions as well as the aggregated clusters
of concepts (Clark et al., 2010), and these helped us to deepen our understanding and gain further insight into
the quantitative research results. The results of the coding process, which consists of identifying a selection
of codes and concepts and the broader themes derived from them, are given in Table 3. Beyond the themes
determined straightforwardly through the various sections of the questionnaire – i.e. LSA adoption, results,
advantages and disadvantages, complementary entrepreneurial methods and tools (not shown in Table 3) –
other themes connected with Entrepreneurship, Strategic Management and Innovation Management research
emerged from the interviews. These themes are further elaborated in Section 5 – Discussion.
Selection of codes and concepts Theme

E.g. proactivity, shaping reality, iteration, Effectuation
experimenting, testing, resource scarcity,
leveraging contingencies, accepting change,
options, limiting investment
E.g. resource scarcity, bootstrapping, Bricolage
challenging the status quo, resource
recombination, creativity, new opportunities,
customer interaction, feedback
E.g. proactivity, shaping reality, creating data Opportunity Creation
and information, uncertainty, iteration,
experimenting, testing, limiting loss
E.g. value creation (products, services, Business Model Innovation
solutions, bundles), value delivery (channels,
marketing, sales, customer relationship), value
capture (revenue model, pricing, cost)
E.g. waste reduction, customer value, customer Lean
perspective
E.g. iteration, feedback, loops, customer Agile
interaction, cross-functional teams, incremental
delivery of features
E.g. product/service innovation, new solutions, New Product Development
launching, experimenting, testing, prototyping,
feedback
E.g. learning, knowledge, build-measure-learn, Organizational Learning
learning from failure, learning to fail,
evolutionary path
E.g. options, alternatives, limiting upfront Real Options
investments, keeping opportunities open,
valuation
17
Table 3 – Codes, concepts and themes
To conclude, the interviews’ results were reviewed and confirmed by the interviewees, to correct any
potential errors or bias and ultimately enhance the correctness of our interpretations.
4. Results
The results from our mixed-methods research are presented as a combination of quantitative findings derived
from the survey on digital startups and the qualitative findings derived from the semi-structured interviews
conducted on a subsample of the survey’s respondents.
The results are aggregated around the four broad research questions addressed in this study.
4.1 Lean Startup Approaches - adoption
The results concerning RQ1 about LSA adoption are given in Table 4 below.
Research item Results

Adoption of LSAs by digital startups 93% Yes
7% No
Main reasons for adopting LSAs 1. Need to achieve product-market fit fast (33%)
2. Need to avoid waste of scarce resources (21%)
3. Need to organize startup’s development process (17%)
4. Need to find alternatives to traditional Business Planning
(8%)
5. Need to please investors (3%)
Main reasons for not adopting LSAs 1. Product-market fit already achieved (45%)
2. Perceived inability to apply to the startup’s business idea
(25%)
3. Perceived complexity of the process (20%)
Sources of information and training on LSAs 1. Incubators/accelerators (32%)
2. Workshops (22%)
3. Universities (19%)
4. Own reading (16%)
5. Investors (5%)
Main concepts, tools and models* 1. Business Model Canvas (91%)
2. MVP (64%)
3. Build-measure-learn loop and pivoting (38%)
4. Earlyvangelists and trial users (14%)
5. Falsifiable hypotheses (11%)
18
6. All LSA tools (11%)
*tot. % > 100% as up to 5 alternative selectable
Stages of LSA adoption* 1. Early stage (58%)
2. Late stage (32%)
3. Early and late stage (25%)
4. Seed stage (6%)
5. Seed, early and late stage (3%)
*tot. % > 100% as up to 3 alternative selectable
Table 4. Quantitative results on RQ1 – LSA adoption
With regards to whether digital startups had adopted Lean Startup Approaches, a striking 93% of the sample
of founders declared that they had explicitly adopted and implemented LSAs with the purpose of driving the
launch and development of their startup.
Startups where these approaches have been adopted do so mostly to achieve a fast product-market fit: they
are well-aware of the need to run experiments on their original version of the business model and to verify
whether their business model is aligned to the market’s needs; alternatively, it could even drive or cater to
new latent needs. The founders also know that timing and speed are instrumental in achieving and sustaining
competitive advantage in such a crowded and dynamic environment as that of digital industries.
As the founder and CEO of a mobile applications startup put it: “All of our apps look great to us and to our
developers – because we came up with them and sweated over them for months. So it’s pretty obvious, isn’t
it? What is way less obvious is knowing if users are ultimately going to like them as much as we do and we
need to test that out”.
Another founder and marketing president of an Internet of Things startup said that “[…] there’s no such
thing as a good idea if there’s no market for it: we need to find out whether there is a market out there or not
ASAP and lean startup is just what I need”.
The founders of digital startups also rely on LSA to address another traditional critical issued faced by any
startup: resource scarcity. The founder in a machine learning startup claimed that “LSAs give our team clear
guidelines and a process to minimize the resources we need in the product-testing phase, which is what we
need to burn less cash”.
19
LSAs are also seen as a way to put the complex and often unstructured launch and development stages of a
startup into a common and formal framework. As a cloud computing startup’s founder claimed: “our startup
is a living organism in its early years… just like a baby. How can you sum up your early years with one
word? I’d say: messy. Many things are going on, you want to be part of them all, and you still cannot figure
out what to do or how to do it, and sometimes why you have to do it. Well, lean approaches tell you what
should be done first and what’s next, while allowing you to keep your flexibility. It’s a kind of step-by-step
flow that allows iterations but helps you to put order into a messy environment”.
A limited number of informants also declared that they had adopted LSAs to find more recent alternatives to
traditional business planning. In the words of the CEO of a mobile payment startup, LSAs were implemented
“[…] in order to avoid mere desk research, as Blank said”. On this point, others also found that adopting
LSAs was something investors pushed and evaluated positively. The marketing manager in a mobile
applications startup said that “when I and [co-founders] first met a business angel, she asked: ‘Did you use
lean startup approaches to see if your ideas make sense?’. We said we hadn’t, and she told us to come back
after we had got some test results – which is what we did, and she became our first investor”.
When it comes to the decision not to adopt LSAs, the main reason seems to be that the startup had apparently
already achieved product-market fit, so no other processes were required to verify this point. Interestingly,
when these startups were analyzed in detail through the qualitative interviews, what emerged was that almost
all of them had been financed through Corporate Venture Capital (CVC) funds, with the mother company
already providing a sort of “captive market” for the startup to tap into. Their answer was hence probably
unconsciously biased, as there was this pre-existing corporate-startup relationship, although the literature
claims that even CVC-backed startups could probably benefit from the adoption of LSAs. The second reason
was that startups perceived LSAs either as inapplicable to the startup’s business idea and domain or too
complex compared to traditional business planning. The founder of a big data startup said that “Lean Startup
is cool but it’s not for everyone: we sell business-to-business solutions, so who should our target evangelists
be? Incumbent companies? How do you test on companies without jeopardizing your potential partnership
with them?”. As another entrepreneur in artificial intelligence put it: “Lean is only apparently simple, but if
20
you go beneath the surface, there is a lot of complexity, mostly related to what an MVP is and who your trial
users should be - and how do you even get in touch with them?”.
With regards to the LSA-related sources of information and training that the digital startups could draw on,
most informants claimed that they practiced lean approaches when the startup was being incubated or
accelerated. Different founders agreed with the CEO of a mobile application startup, who stated that: “A
good 90% of our training sessions with tutors, coaches and mentors were actually about lean principles
applied to our startups”; and “here in [incubator’s name], lean startup is everywhere… you’re literally
flooded with lean principles, there’s no escaping from it! [Laughing]”. Dedicated workshops were another
major source for training, although their results often did not meet the startuppers’ expectations: the Chief of
Operations in a social media startup claimed that “lean startup workshops, tutorial and training sessions
spring up like mushrooms, but most are just [obscenity] and a total waste of time, so you need to be very
careful about who’s running and holding them”. In this adverse selection setting, bad choices are set to play
an increasing role in the adoption of LSAs, as more and more actors and institutions start holding courses
and training sessions – sometimes apparently picking their stuff out of thin air rather than having any real
expertise. To solve this matter, many startups are turning to universities and university-based schools of
management and/or entrepreneurship outfits. “Lean Startup courses in a college are rather expensive, but
they’re becoming more and more hands-on and practice-oriented rather than purely theoretical; so they’re a
good way to gain professional and certified knowledge”, as an Industry 4.0 founder and marketing manager
said.
Several entrepreneurs also declared that they learned about LSAs on their own by reading the books written
by Blank (2007), Blank and Dorf (2012) and Ries (2011). A very small number of founders also declared
that it was their investors who made the suggestion to adopt LSAs – and this largely overlaps with the
informants who claimed that they had adopted these tools to “please investors”.
Several insightful considerations emerge on the use of specific LSA concepts, tools and models. The most
widely used tool is the Business Model Canvas, whereby “[…] we easily make our business idea come down
to earth” – as the CEO of one eCommerce startup said – in a “unified, simple and straightforward way” –
according to the Chief Financial Officer in a security startup. In turn, the startups seemly adopt MVPs to
21
outline the most important functions of their business idea and possibly enable the early interaction with
customers. The whole build-measure-learn loop based on the MVP was also applied by startups striving to
hear the “customers’ own voices and opinions based on something they touch and feel” – as a founder of a
startup in digital wearables said, and so prepare for pivoting. The least used concepts, according to the
quantitative analysis, were earlyvangelists and falsifiable hypotheses, and this finding relates to the
disadvantages and criticalities that startups encounter when applying lean processes (see Section 4.3).
Surprisingly, a few startups declared that they adopted all the models at the same time, as if they somewhat
personalized the LSAs during their implementation processes in line with the startups’ specific needs: “I
found the [business model] canvas and MVPs quite useful, while frankly I never used those falsifiable
hypotheses and evangelists… they sounded a bit tricky”. Such cherry-picking and piecemeal approach may
have significant implications when assessing the LSA results and performance.
The last point concerning the adoption of LSAs referred to the stage of startup development when the
practices were leveraged. Here, many startups claimed they employed LSAs mainly in their early stage
development, when getting their business model ready for the informal investors’ approval; however, the
founders also observed that adopting LSAs could prove useful in later stage development, as “VCs look for
indications about market traction, and lean startup approaches can show them your traction” – as a founder
in a mobile services startup put it. It was apparently unusual to adopt LSAs at the seed stage, where the
business idea is at its embryonic stage of development, because, at that time, “the idea is too fuzzy even in
your own head, you could probably sketch a canvas around it, but rather than embarking on a full lean
startup or customer development process, you ask around for informal feedback, quizzing your network of
peers, while you are asking them for pre-seed and seed money”, as noted by an entrepreneur in a IoT-related
startup. A number of startups also declared that they continuously used LSAs in cycles and iterations, since,
as a CEO in a mobile payment startup said, “you really never stop pivoting and iterating a digital service”.
4.2 Lean Startup Approaches - results
The findings concerning RQ2 on LSA results and outcomes are given in Table 5 below.
22
LSA implementation process length Average: 8.2 months
Upper bound: 13.5 months
Lower bound: 4.1 months
LSA implementation process cost (total cost) Average: $ 34,000
Upper bound: $ 180,000
Lower bound: $ 19,000
LSA implementation process cost (% of capital Average: 24%
raised) Upper bound: 43%
Lower bound: 18%
LSA adoption overall satisfaction 2.8 on a 4-point Likert scale
(1. Dissatisfied – 4. Fully satisfied)
Table 5. Quantitative results on RQ2 – LSA results and outcomes
With regards to the duration of LSAs implementation, the founders declared that it took them an average of
8.2 months to go from the first set of falsifiable hypotheses embedded in a business model and a MVP, to
achieving a reasonable product-market fit that gave them the go-ahead to proceed with scaling – with a lower
bound of 4.1 months and an upper bound of 13.5 months. During the interviews, it became clear that changes
to the LSAs’ “lead time” depended mainly on the following reasons: (i) ability or inability to find team
consensus about the right way to set hypotheses and design an initial business model; (ii) ability or inability
to craft the right MVP and to properly prioritize the MVP tests; and (iii) ability or inability to spot the right
cluster of earlyvangelists or trial users, engage them and get informed feedback from them. The LSA length
was clearly correlated to its cost – in terms of capital expenditure for equipment and, above all, operational
expenditure in human resources, materials and advertising – such as AdWord campaigns – both in absolute
terms and as a percentage of the capital raised.
The best performers in the sample, those keeping the LSA implementation length and cost to a minimum, felt
that “the key here was to make all founders agree on one single statement that summarized our view on the
startup and its best future” – as stated by the CEO of a machine learning startup – while “cutting all features
that were making our MVP too ‘bulky’ and too similar to a real prototype, rather than a pretotype as it
should be” – in the words of an Industry 4.0 CFO. This was coupled with “a few weeks spent in finding
where our target audience actually hung out, to target the right early users. It may have looked like time
wasted, but as it turned out it clearly wasn’t. It made us save a lot of time repeating tests in different contexts
23
and with different user groups”. The low performers argued that their time and costs went out of control
because of several factors, the most relevant being: “the inability to make plans about when to start and,
most of all, end testing” – as stated by the COO of an eCommerce platform startup; “our probably
inaccurate definition of the right metrics to use and the right prospective clients to include in the experiments
caught us up in a never-ending iteration process” – in the words of the CEO of an Internet of Things startup;
and “our unwillingness to pivot was based on the sunk costs we had already incurred in device development”
– as the CEO of a digital wearables startup admitted.
The last survey question on LSA results covered overall satisfaction. Surprisingly, although most digital
startups by far said they used the approaches and that they saw something in return their implementation
efforts, the overall satisfaction hit a relatively poor 2.8 in a four-point Likert scale, leading to slightly less
than moderate satisfaction. The qualitative interviews revealed that behind this mild satisfaction was the fact
that they recognized the complexities and possible drawbacks of the approach (as discussed in Section 4.3),
something that even the best performers experienced. “All in all, I enjoyed applying LSAs and I think they
were beneficial for our startup in its early stages of development. Having said this, these apparently
pragmatic tools sometimes work great in theory but not so well in practice, and did they make us sweat!”,
the founder and general manager of a big data startup said. This was reinforced in an interview with another
founder of an Internet of Things startup, who stated that “lean startup tools, are, let’s say, a great start… but
they need a lot of work to define, design and refine them before they’re ready to tell you something you need
to know about your company and its market”.
4.3 Lean Startup Approaches - advantages and disadvantages
Table 6 summarizes the main findings gained from RQ3 on the principle advantages and disadvantages of
LSAs.

Main advantages of LSAs 1. Reducing time and cost for startup testing (74%)
2. Aligning business idea to customer needs (68%)
3. Verifying and pivoting all business model parameters (52%)
4. Receiving rounds of financing (39%)
24
5. Offering alternatives to traditional intellectual property protection (28%)
*tot. % > 100% as startups could select up to all 5 options
Main disadvantages of LSAs 1. Defining and designing MVPs (82%)
2 Identifying and engaging earlyvangelists and trial users (69%)
3. Defining testing priorities and designing tests (52%)
4. Missing other market opportunities and threats (39%)
5. Obtaining information about the startup’s sources of advantage (36%)
Table 6. Quantitative results on RQ3 – LSA advantages and disadvantages
With regards to the advantages of implementing LSAs, most digital startup informants claimed that these had
helped “significantly in cutting the time and costs we had expected for developing the startup, especially
during testing” – as stated by the CEO of a mobile applications startup. She continued by saying that “I have
experience with previous startups where we didn’t use startup and customer development, and I can say that
our savings here in time and costs are remarkable. Of course, this happens if you learn to use the right tools
properly”. Discovering and validating the alignment between the startup’s business model and the needs of
its target market was also a key outcome stemming from the adoption and implementation of LSAs: “before
we used LSAs, we really had no clue as to whether customers would actually like and be willing to buy our
products. This changed radically after they sent us their feedback and we learnt from it”, as said by the
founder of an Industry 4.0 startup.
Another advantage felt by founders implementing LSAs was that of verifying and/or knowing when to pivot,
all the business model canvas parameters, which is something that goes well beyond the bundle of products
and services embedded in the value proposition. As the CEO of a cloud computing startup explained, “we
thought that LSAs were meant to help us verify our value proposition, but we found that it's a good set of
tools to test other parts of our business model, like its channels and its revenue model”. The COO of a
mobile applications startup added that, “by applying Lean Startup and Customer Development, not only did
we pivot six of the nine parameters in our canvas, we also changed two of them so radically that our
launched startup looked very little like our original idea”.
Interestingly, the founders also noted that adopting LSAs made it easier for them to receive rounds of
funding from formal and informal investors alike. The CEO in an eCommerce startup explained that “the
25
evidence, metrics and figures first shown to business angels, and later to venture capitalists, were so real
and backed by testing, that they certainly helped us to demonstrate our business potential and its growth
traction”. Another entrepreneur in a mobile applications startup said that “the tools and metrics championed
in lean approaches, like cohort analysis, are something investors really like. If you use them, you are
showing them you really know your stuff. That’s when they start trusting you and probably think: ‘I could
give this guy my money’”.
A last, yet insightful, hidden advantage of adopting and implementing LSAs concerns their ability to force
entrepreneurs to rethink traditional Intellectual Property (IP) protection logics. In the words of the CEO of a
machine learning startup, “when I read the books, the first thing I thought was ‘ok, nice, but once I launch
my MVP and run experiments, aren’t I putting my ideas out there for competitors to steal?’ Yet, as I deep-
dived into the lean startup philosophy, I understood that being lean doesn’t mean being dumb, and rather
than IP protection, it’s all about speed”. Another COO in a mobile applications startup added the following:
“If you go for build-measure-learn loops and experiments, it’s key to being fast. You can’t just run the
experiments and sit there and wait: once you get the green light, you must be ready to pivot, develop and
launch at full throttle. By then competitors know you’re there, but you will still have the advantage by being
first in learning what counts for your customers, and can act fast accordingly”. Coupled with IP protection,
the informants also addressed the matter of reputational risk associated to poor MVPs. Our interview with
the founder of a mobile application startup shed light on a view common among digital startups: “Of course
launching a sloppy MVP on your market could harm your brand: that’s why you should definitely use ‘fake
brands’ when doing your experiments. Prospective customers may get upset by the fake brands’ products,
but, at the end of the day, they will turn to your brand’s final product once it’s been thoroughly LSA-tested
and is ready”.
As anticipated in the previous sections, many insights were derived from the discussion of the main
disadvantages of LSAs. By combining quantitative and qualitative data and information, a clear and true
picture emerges of the weaknesses in LSAs, and this could lead to improving these processes.
First, many founders encountered severe problems when defining and designing their MVPs, and also when
designing MVP-based tests. The founder of a mobile applications startup said the following: “In line with the
26
lean rationale, we interviewed customers as early as possible, using an online survey to ask them about the
features of the would-be product they liked the most. We also asked how much they were willing to pay for
each feature. The results were rather disappointing: as they said they liked all the functions, we learnt
nothing about customer preferences… and to compound it, they were so unwilling to pay, there was no
justification for our investment in any of these functions… practically speaking, that test was useless”. This
statement is revealing, as it shows how entrepreneurs still confuse MVP and MVP-based testing with
marketing research: according to its original proponents, MVP-based testing is not about market research,
but is should enable the experimenting on customers’ natural behaviour; tests should not refer to willingness
to pay, but rather simulate an environment where customers believe they are actually making a purchase
(Blank, 2007; Blank and Dorf, 2012; Ries, 2011; Ries and Euchner, 2013).
Another entrepreneur in a social media startup complained that “In the first version of our MVP, we created
a video showing how our application should work. That’s something we learned from other popular lean
startup cases, like Dropbox. However, beyond counting the number of times the video was watched, we
learned little from the process”. This statement shows that, when applying MVPs, it is crucial to create an
MVP that is actionable, in the sense that it should trigger the customer’s natural behaviour and actions that
carry useful information and evidence for startups – like registering to a beta waiting list, as in the case of
Dropbox.
A third point, stated by a founder in an IoT startup, showed how apparently difficult it can be to place MVPs
within a business-to-business environment: “We serve companies with complex systems, so it’s kind of hard
for us to isolate key functions and test them on their own; what happened when we applied the lean approach
was that we ended up by creating a full prototype – and its budget was far from being cheap or lean…”.
A concluding remark on MVPs from a COO in another Internet of Things startup is also worth mentioning:
“Lean tells you to build a MVP, but gives you no clear guidelines or indications whatsoever on how to do
so!”.
Spotting earlyvangelists and trial users is another excruciating difficult exercise for many startups. As a
founder in a mobile applications startup noted: “We kept on asking ourselves ‘who, which customers, should
27
we test our business model on?’, but we never came to a definite conclusion. We eventually launched our
MVP after testing indiscriminately across our target market, with mixed results". Another marketing
manager in a machine learning startup highlighted a further issue relating to identifying and engaging with
trial users: “we identified our target segment and actively engaged with several influential users, receiving
interesting feedback from them. We then found another potential target segment, engaged with those users
too, and received totally different feedback compared to the first lot. What were we meant to do with all this
assorted feedback? And how were we to know when to stop?”.
Many of these drawbacks, which on the surface were connected to LSAs, stemmed from the fact that startups
seldom see the need for setting assumptions and hypotheses about themselves, their market and their
customers (this was highlighted in Section 4.1, where we saw that, among the lean concepts, falsifiable
hypotheses were the least used). Hypotheses were often taken for granted, or defined too broadly. Some
founders did not address the most penetrating questions about their customers and what they want, their
pains and gains (Osterwalder and Pigneur, 2010), which ultimately made it difficult for them to find out who
their earlyvangelist actually were. This emerged from the words of a CEO of a mobile payment startup, who
admitted honestly that: “The problem was never the process. It was us. We never really accepted that we had
to fully adopt it in all its steps, with an open mind and true commitment: that’s why it did not pay off as we
had expected”.
In line with this sometimes limited commitment to adopting and implementing LSAs as a whole, the findings
of both the survey and the interviews revealed a common misconception about lean approaches, their reach
and their role in supporting startup development. The informants pointed out that the LSAs gave very little
support in spotting industry-related opportunities and threats hidden in external trends or in the market’s
structural determinants; or they received poor indications about the startup’s sources of competitive
advantage, and about whether such sources could actually help to beat the competition. Both of these
arguments show that entrepreneurs may tend to have unrealistic expectations about lean approaches and
overestimate their own ability to support startup development. In actual fact, analyzing the features of the
external industry – alongside those of potential customers – as well as the startup’s internal sources of
differentiation goes well beyond the scope of LSAs. This interesting finding also indicates that digital
28
entrepreneurs must complement LSAs with supplementary strategy analysis models and an overall business
strategy.
4.4 Lean Startup Approaches and other entrepreneurship methods and tools
The findings that relate to RQ4 - how digital startups connect and combine the LSAs with other
entrepreneurial tools for launching startups, as well as their overall business strategy – are given in Table 7.

Other models/tools/approaches used to 1. Business Plan (91%)
complement LSAs 2. Agile Development (82%)
3. SCRUM and Sprints (53%)
4. Feature-Driven Development (FDD) (44%)
5. SWOT Analysis (41%)
Table 7. Quantitative results on RQ4 – LSA and other entrepreneurship and strategy tools
A first, significant finding emerged concerning the approach most widely used in conjunction with LSAs,
with a striking 91% of the founders saying that they use Business Planning. This result may seem surprising,
as Customer Development was proposed as an alternative to business plans, as the latter kept the startup’s
development too far removed from reality for too long (Blank, 2007). Nearly all the startups declared,
however, that they had solved this false dichotomy by fruitfully integrating LSA-derived data into the first
versions of their business plan, thus combining a “get out of the building” stance with a well-structured
document describing their strategic, marketing, operational, economic-financial, organizational and human
resource-related plans (Delmar and Shane, 2003). “Lean startup testing data and metrics proved a useful
input to our business plan, where data came from our experiments with customers rather than from
secondary sources”, claimed the marketing manager in a mobile applications startup. This arrangement also
helped founders to raise capital, since “[…] angels first, and VCs later, seemed to really appreciate the
combination of real-world data from the LSAs and the traditional business plan structure”.
The informants also emphasized the fact that they could combine LSAs with Agile methods (Beck et al,
2001; Nerur et al, 2005; Holmström et al, 2006; De Cesare et al, 2010; Dingsøyr et al, 2012; Cram and
Newell, 2016), to fill some of the gaps in Lean Startup and Customer Development. More specifically, the
29
founders were keen to introduce methods such as SCRUM and sprints (Schwaber and Sutherland, 2011), as
well as Feature-Driven Development (Abrahamsson, 2003). The COO in an Internet of Things startup said
that “sprints gave us timelines for iteration cycles in the build-measure-learn loops”, while another founder
in a security startup observed that “Feature-Driven Development helped us to define the features to be
embedded in the MVPs”.
Moreover, digital entrepreneurs appeared not to underestimate the role of the traditional Strengths-
Weaknesses-Opportunities-Threats (SWOT) analysis (Lorange, 1980) in feeding the LSAs: “Before you even
start any lean development process, you should be well-aware of what is going on in your target market, its
trends, opportunities and threats, as well as of what you can offer in that market, what really makes you
different from your competitors and will make you win in the long run”, as the CEO of an Industry 4.0
startup explained.
5. Discussion
This study examines how digital startups’ adopt and implement Lean Startup Approaches, pointing to a
number of opportunities and issues that could be addressed to enhance the value of LSAs for digital
entrepreneurship theory and practice.
5.1 Framing LSAs within entrepreneurship theory
With its insights, this research indicated a number of contact points between LSAs and other theories that are
worth discussing and clarifying.
With regards to the theoretical foundations of the Lean Startup Approaches, the empirical findings of this
study back the claims made in the few recent works that looked into the antecedents of LSAs (Frederiksen
and Brem, 2017; Contigiani and Levinthal, 2018; Ghezzi and Cavallo, 2018; Yang et al., 2018). As shown in
Table 3, our qualitative research brought up nine themes that form the theoretical roots of LSA, namely: (i)
Lean; (ii) Agile; (iii) New Product Development (NPD); (iv) Real Options; (v) Organizational Learning; (vi)
Business Model Innovation (BMI); (vii) Effectuation; (viii) Bricolage; and (ix) Opportunity Creation.
30
From a comprehensive overview, LSAs can hence be framed in a theoretical model where these approaches
stand at the crossroads of the above research streams and disciplines (Figure 3). These streams are far from
being stand-alone silos and some of their key elements often partly overlap, as our empirical research shows.
Figure 3. Antecedents and theoretical roots of the Lean Startup Approaches
While the LSAs’ relationship with themes (i) to (vi) had already been discussed in the recent works
mentioned previously, the findings of this study mostly contribute to the debate on the role of LSAs with
reference to Effectuation, Bricolage and the whole theory of Entrepreneurial Opportunity.
When discussing Lean Startup Approaches, the interviews with our digital startup founders largely revolved
around their standpoint and reasoning, and their actions to interpret their own role of entrepreneurs in a
resource-constrained and uncertain environment.
With regards to the logic behind entrepreneurial behaviour, the entrepreneurs showed how adopting and
implementing LSAs drove them to take an “effectual” (Sarasvathy, 2001) or “bricolage” (Baker and Nelson,
31
2005) stance on many occasions, underscoring how these logics are, in essence, connected to LSAs’ main
steps and constituting elements.
Formulating hypotheses about the world’s evolution and the startup’s role in influencing it is similar to the
Effectuation principle stating that the future is shaped by entrepreneurs, rather than being a mere
consequence of the past.
Moreover, building a minimum viable product to experiment and test hypotheses on the market and argue for
or against the viability of the startup’s business model reflects an effectual and bricolage-oriented logic, in
the ways in which resources are perceived and used, goals are set and learning is achieved.
Despite operating in a resource-constrained environment, entrepreneurs can take full advantage of their
extremely scarce resources by combining them originally into low-cost - yet highly informative - pretotypes
of their value propositions – the MVPs – and then deploy them to test their BM assumptions.
In addition, experimenting and testing efficiently though iterations and short startup-customer loops that
provide valuable feedback is also consistent with Effectuation and Bricolage, in terms of having the
flexibility to handle and even exploit unexpected contingent events and information brought up by the
experiments. At this stage, the startup’s paramount objectives boil down to minimizing upfront investments
during the initial learning phases, in line with the concept of affordable loss, while making do with limited
bootstrapped resources (Sarasvathy, 2001; Baker and Nelson, 2005; Fisher, 2012).
This study hence empirically supports the conceptual claim that LSAs provide operational support to
Effectuation and Bricolage in a systematic and scientific way, overcoming in this manner the heuristic
approach to problem solving – based on intuition, unclear paths and somewhat casual attempts – that
traditionally went hand-in-hand with these logics and, to some extent, limited their effective implementation.
On a bigger picture, another insightful contribution emanating from this empirical research concerns the
theory of entrepreneurial opportunities implicitly embodied by LSAs.
Assuming that there are two alternative theoretical settings where opportunities can be discovered or created,
entrepreneurs using LSAs take up a creational stance towards entrepreneurial opportunities, since the
32
strategic tasks for entrepreneurs in a creational setting are to ask customers the right questions, design new
experiments, remain flexible and learn (Alvarez and Barney, 2007).
LSAs seem to implicitly address one of the research questions for creational theory anticipated by Alvarez
and Barney (2007 – p. 22): “How can entrepreneurs use incremental, iterative and inductive processes to
make decisions?”. This study provides empirical evidence to answer this creation theory question, claiming
that LSAs are the operational, systematic and essentially scientific processes that enable opportunity creation
through enactment.
Positioning the investigation of LSAs within the debate on opportunity discovery vs creation helps to bring
about empirical contributions for advancing entrepreneurial opportunity theory.
Alvarez and Barney (2007) present these two internally consistent theoretical perspectives on opportunities
as dichotomous: while discovery theory assumes that the environment is risky and entrepreneurs recognize
and exploit opportunities by means of risk-based decision-making tools, such as the business plan (Delmar
and Shane, 2003), creation theory considers the environment to be uncertain and states that entrepreneurs
form opportunities through enactment based on alternative iterative and incremental processes – like LSAs,
as this study contends. However, what we found empirically is that, although LSAs lead entrepreneurs to
adopt a creational and effectual approach, other actors in the entrepreneurial ecosystem still expect them to
follow a largely discovery and causation approach, at least in terms of the output they produce and present
(once again, the business plan), since the latter approach is more closely aligned to and compliant with their
traditional “due diligence” process based on risk assessment and management (Drover et al., 2017).
The digital entrepreneurs interviewed solved this contrast pragmatically by, in essence, sequencing the
creational and discovery processes and tools. They first iterate and experiment through LSAs to create
opportunities and generate data, information and knowledge; they then feed such knowledge into a well-
structured business plan, to discover opportunities for the scaling, market penetration and company building
of their startups. The turning point for moving from a creational to a discovery setting is achieving a
“product-market fit”, in other words, the alignment between the startup’s validated business model and what
customers demand (Blank, 2007; Blank and Dorf, 2012; Ries, 2011).
33
The findings of this study suggest that, while LSAs fall into the domain of creation theory, their output – i.e.
data, information and knowledge created with reference to an uncertain environment – could provide the
basis for a discovery approach by being fruitfully used as input to a business plan – i.e. a collection of risk-
based decision-making tools.
As a result, these findings also offer an answer to the question of how uncertainty can be best converted into
risk and managed by entrepreneurs (Frederiksen and Brem, 2017). LSAs’ iterations, interactions and
experiments can help in the transition from uncertainty, where knowledge about opportunities is created, to
risk, where knowledge is further elaborated to discover opportunities through planning. This is also in line
with the assertion by Brinckmann and colleagues (2010) that entrepreneurial planning must be appropriate to
the circumstances.
This contribution concerned with how entrepreneurial approaches, models and tools are connected,
empirically bridges the two theories on entrepreneurial opportunities (creation and discovery), by positioning
them within an “opportunity space” that helps to explain the empirical findings emerging in the study
(Figure 4).
34
Figure 4. The Entrepreneurial Opportunity space
The first wave of opportunities is created by entrepreneurs through an enactment process based on
experimenting and iterating on the startup’s business model and its elements. The underlying cognitive logic
driving opportunity creation in an uncertain environment is the same as that found in Effectuation and
Bricolage; and the operational decision-making tools to create opportunities systematically and scientifically
is embodied in the Lean Startup Approaches.
Iterations proceed in incremental loops until the product-market fit is achieved, and the business model is
proven to be viable, replicable and scalable (Ries, 2011; Blank, 2013). At this time, the decision-making
context is converted from uncertain to risky, setting the scene for entrepreneurial cognition based on
causation. Entrepreneurs can then discover a second wave of opportunities through planning, where the set of
decision-making tools encompassed by the business plan are such that the expected returns and the
probability distributions of such returns can be estimated. These, in turn, lead to different alternative
scenarios and related plans.
As the opportunity space helps to bridge the alternative approaches to entrepreneurial opportunities,
cognitive logics and operational decision-making tools, it is useful to note that the gap between the
opportunities created and those discovered is not clear-cut – as indicated by the vertical dotted line in Figure
4. Iteration and planning can sometimes alternate or coexist in specific stages of startup development, as the
context changes from uncertain to risky and calls for different entrepreneurial logics and tools. As Alvarez
and Barney (2007 – p. 19) suggested, entrepreneurs are required to make decisions using context-appropriate
tools, which will help them to form and exploit opportunities more effectively.
Recognizing that entrepreneurial opportunities are, in essence, a doubleheader also helps to explain why both
discovery and creational theories are internally consistent (Alvarez and Barney, 2007). They are so because
they can adequately depict a different evolutionary stage for the new venture within the opportunity space;
and all the actions that determined the evolution of the startup could be attributed ex post to one or the other
35
theory depending on the stage that the entrepreneur or the external observer wishes to emphasize or can
recollect.
5.2 Enhancing the practical adoption and implementation of LSAs
The combined results of this mixed-methods study into digital startups helped us to make sense of the
practice of adopting and implementing LSAs, and so provide practical guidelines that are backed by theory
and can help digital entrepreneurs to set this LSA process into operation. The unexpected result indicating
that digital entrepreneurs are only modestly satisfied by the outcome of the LSA implementation shows that
these approaches are still burdened by many limitations and need to be amended.
A first guideline to direct the adoption and implementation process examines the kind of digital startups that
are encouraged to adopt these approaches. Startups backed by informal investors or independent venture
capitals benefit from LSA adoption, but not just them, as even CVC-backed startups find Lean Startup and
Customer Development useful for verifying and validating their business ideas – despite maybe already
having a captive market for their products and services. With regards to the stage in the startup’s
development lifecycle when LSAs are used, although most startups implement LSAs in their early stage, the
introduction of a continuous adoption cycle that takes into fair account the different goals and metrics at the
various stages of startup development is beneficial in supporting iterations and validations, as the internal
and external contexts vary and can turn out to be uncertain, this being in line with Agile principles (Beck et
al., 2001). Not starting too late also helps startups to avoid those sunk costs that make it much harder to
pivot, as emerged from the qualitative interviews.
Making bad choices in an adverse selection setting also appeared to play a role in hindering the adoption and
implementation of LSAs (Jullien, 2000), caused at least in part by the digital startups’ difficulty in finding
the right suppliers providing courses and training, indicating that there is a strong case for properly assessing
the quality of their suppliers before the training starts. This result shows that universities, colleges and
research institutions will play a growing role in the digital startups’ ecosystem, as the certified means for
obtaining LSA-related knowledge. To some extent, this finding also goes against some recent claims that
36
entrepreneurship cannot be taught and a college degree does not really help people to become a startupper
(e.g. see Forbes, 2012). As Frederiksen and Brem (2017) sharply pointed out, Eric Ries (author of The Lean
Startup) has a degree himself, and his background more than likely informed his work in generating the Lean
Startup.
Another key suggestion for digital founders embarking on the process of LSA implementation is to place the
right focus on phrasing falsifiable hypotheses. These assumptions provide the basis for testing, since
entrepreneurs should build their business model around them and their MVPs should stem from them.
Assumptions were, however, too often taken for granted or defined too broadly. Digital entrepreneurs have
to realize that defining hypotheses is not a mere exercise, but acts as the operational trigger for the whole
scientific method behind LSAs.
With regards to the work behind crafting the MVP, designing MVP-based experiments and running the
build-measure-learn loop, several takeaways emerge from this study. Startups should spend enough time in
actually understanding who their target customers are, together with their pains, gains and needs
(Osterwalder and Pigneur, 2010), and so properly identify the earlyvangelist and trial users able to provide
informed feedback and stimulate useful learning.
Once the earlyvangelists are identified, the MVPs should be built and placed in their hands. This research
highlights a significant distinction that makes LSAs different from traditional approaches to startup
development. MVPs are not related to marketing research, as instead they should enable experiments on the
customers’ natural behaviour. MVPs are artefacts that should have the following key characteristics: (i)
resemble and embody the business idea; (ii) be actionable, as they should activate customer response and
actions to learn from; (iii) be measurable through the MVP testing outcomes; and (iv) be less wasteful than
prototypes. When possible, entrepreneurs are also encouraged to make the MVP a paid-for product rather
than just a free version, to further qualify the input from early adopters who become paying customers (Ries,
2011). This was seldom the case in the sample analyzed, thus impairing the measuring and learning stages of
the loop.
37
To design MVPs, our findings suggest that entrepreneurs can leverage Agile methods such as Feature
Development Design (Abrahamsson, 2003) to provide guidelines on how to identify and design the
minimum features and run iterations. In addition, when it comes to running iterations, digital products and
services could be run through Scrum sprint cycles (Schwaber and Sutherland, 2011) to control for time and
budget.
When the digital startup operates in a B2B context, our results suggest that these startups should find their
right evangelists among existing business customers by carefully assessing their customer companies’
purchasing processes – as decisions about buying into the startup’s digital solution can come from different
stakeholders. Additionally, MVPs in a B2B environment cannot be overly “minimum”, in the sense that they
should incorporate a sufficient number of features at a satisfactory level to compete with existing offers.
MVP-based testing should even lead digital founders to radically rethink their views on IP protection and
their approach to launching the startup. Rather than fighting to control, hide and defend their business idea,
LSAs place significant emphasis on collaboration and openness, where the key to competitive advantage is
not the original know-how, but the startup’s learning speed. This point is inherited from an effectual-type
logic, which favours partnership and cooperation rather than competition (Sarasvathy, 2001; Yang et al.,
2018).
Insights were also derived from the span and width of the process of validating and, when appropriate,
pivoting the business model. This research underscores how LSAs support innovation to the whole value
architecture of the digital startup’s business model, including value delivery and capture as well as value
creation. Changes to such value parameters can be incremental or radical, resulting in slight modifications or
even dramatic changes to the original business idea. Although Ries and Blank themselves argue that the
business model of a startup is up for experimenting on (Gustafsson and Qvillberg, 2012; Hartman, 2013;
Heiramo, 2013; Ghezzi and Cavallo, 2018), a common misunderstanding among entrepreneurs is to place the
focus of LSAs on testing product features. Our findings strengthen the belief that digital entrepreneurs
should, instead, take entire business models as their units of analysis for experiment purposes and as the
entity to which any pivoting should apply. Moreover, and remarkably, informants often declared that testing
38
on the startup’s go-to-market strategy and its execution was even more important than validating the product
itself.
For digital entrepreneurs wishing to embark on LSAs, a major takeaway of this study is that a piecemeal
approach, where founders cherry-pick concepts and models they like and ignore other parts, is fundamentally
flawed. Several founders used LSAs in a personalized, unsystematic and disconnected fashion, maybe due to
poor training or to the heuristic intuitive approach they were accustomed to. They should, instead, have been
fully committed to adopting the whole process: for entrepreneurs, LSAs are the means to go beyond mere
heuristics and embrace a scientific method towards entrepreneurial opportunity and action.
The opposite of a piecemeal approach is what is needed, and rather than adopting bits and pieces,
entrepreneurs should combine all of the LSAs’ process with other tools, one of which is certainly the
business plan. In line with existing studies that claim that the business plan plays an important role in linking
entrepreneurship and strategic management (Kraus and Kauranen, 2009; Delmar and Shane, 2003), and
building on our discussion on the entrepreneurial opportunity space, our argument is that Blank’s (2007)
famous motto “instead of writing an intricate business plan, design a business model [and apply customer
development]” should be altered to “before writing a business plan […]”. LSA experiments and tests,
together with their real-world results and metrics, should be used as input to the planning operations in a
well-structured business plan that can be easily assessed by both formal and informal investors.
As a concluding remark, this research shows that digital entrepreneurs sometimes blame LSAs for pitfalls
that fall outside the domain and scope of these approaches, and instead relate largely to the process of
formulating an overall business strategy, which startups tend to overlook. Reflecting on the startup’s vision
and mission, and strategically analyzing its internal and external environment, can better inform the way
entrepreneurs formulate their falsifiable hypotheses and design their preliminary business model.
Table 8 contains practical guidelines that entrepreneurs will find useful to follow when adopting and
implementing LSAs.
39
Issue Practical Guidelines
Type of digital startup  All startups, including those that are CVC-backed, benefit from adopting and
implementing LSAs.
Stage of startup  Startups are to adopt LSAs in their early stages of development, while
development continuously implementing them following Agile principles whenever the context
turns out to be uncertain.
Bad choices  Startups are to carefully assess and select suppliers for LSAs courses and training
concerning the sessions.
provision of LSA-  Startups are to rely on certified and experienced actors – e.g. universities, colleges
related knowledge and research institutions, top-ranked incubators and accelerators.
Formulation of  Entrepreneurs must think carefully on how they can accurately formulate
falsifiable hypotheses falsifiable hypotheses about their startup’s business model (a step they often
neglect).
 Falsifiable hypotheses constitute the operational trigger for the scientific methods
embedded in LSAs.
Identification of  Entrepreneurs are to properly evaluate who their earlyvangelists and trial users are
earlyvangelists and where they “hang out”, in order to target the right prospects and receive
informed feedback.
 In B2B settings, evangelists are found among existing business customers by
carefully assessing the customers’ purchasing processes and identifying the key
decision-makers.
MVP design  MVPs are artefacts with these key characteristics: they (i) resemble and embody
the business idea; (ii) are actionable; (iii) are measurable through the MVP testing
outcomes; and (iv) are less wasteful than prototypes.
 MVPs as paid-for products increase the amount of information they carry when
tested upon.
 MVP design can leverage Feature Development Design (FDD) to provide
guidelines on how to identify and design minimum features and run iterations.
 In B2B settings, an MVP cannot be too “minimum”, as it should incorporate a
sufficient number of features at a satisfactory level to compete with existing
offers.
Experimenting and  Experiments are to be MVP-based, which in turn means they are BM-based.
testing  Digital products and services can be run through Scrum sprint cycles, to control
for time and budget.
 Entrepreneurs are to radically rethink their views on IP protection, embracing
openness and collaboration through fast experimenting and learning.
Business Model  Startups are to experiment and test on all elements of their business models, not
validation and pivoting just their value proposition (product, service, solution, bundle).
 Executing a go-to-market strategy often requires more testing than the value
proposition itself.
LSA broad adoption  Entrepreneurs are to adopt LSAs comprehensively, rather than cherry-picking the
and implementation steps and elements they perceive as most useful.
process  Entrepreneurs are to go beyond heuristics and apply a scientific method by means
of the LSAs.
 Entrepreneurs are to integrate LSAs with business planning, thus altering Blank’s
motto to: “before writing a business plan, design a business model and apply
LSAs”.
 Entrepreneurs are not to overlook the process of strategy formulation and strategy
analysis which can inform the formulation of falsifiable hypotheses and design of
a preliminary business model.
Table 8 - Practical guidelines for LSA adoption and implementation
40
6. Conclusions
This study presented a first large scale research into if and how digital startups adopt and implement Lean
Startups Approaches. The findings emerge from a mixed-method combining quantitative and qualitative
analyses, and the contributions are relevant for both the theory and practice of entrepreneurship.
In terms of value for theory, the empirical evidence shows that the entrepreneurs’ behaviour when applying
LSAs mirrors that described in the theories of Effectuation and Entrepreneurial Bricolage. Looking at the
bigger picture, the findings support the claim that LSAs are a set of operational, systematic and scientific
decision-making tools for supporting entrepreneurial opportunity creation: entrepreneurs use LSAs to make
sense of the opportunities they create, rather than to discover and plan around them.
Empirical evidence, however, shows that entrepreneurs solve the creation vs discovery dichotomy
pragmatically by sequencing their use of LSAs and business planning. This suggests that there is an
“opportunity space” that can bridge the various alternative theoretical views on entrepreneurial actions,
cognitive logics, decision-making contexts and decision-making processes and tools.
In terms of value for practice, this study can provide a set of guidelines and real-world insights into the
adoption and implementation of LSAs, and this can extend the anecdotal evidence currently available about
common practice among digital entrepreneurs, while suggesting a move towards combining these approaches
with other agile, entrepreneurial and strategic models.
This study’s limitations mostly refer to the following: sampling bias – i.e. focus on a specific country and on
funded digital startups – which could have influenced the statistical and theoretical samples selected; and
observer bias (Yin, 1984), which could have distorted the interviewees’ perception of the interviewer’s
questions, and/or the interviewer's interpretation of the interviewees’ answers. While the sound research
design may have helped to amend these limitations, future studies could explore different samples of startups
- e.g. non-digital startups, or non-funded ones – and other countries, while investigating how to combine
Lean Startup Approaches systematically with the other theoretical and practical streams indicated in this
study.
41
7. Acknowledgements
I would like to thank the Editor in Chief, the Guest Editors and the two anonymous Reviewers, who helped
significantly to improve the contribution of this study as a result of the revision process.
8. References
P. Abrahamsson, Extreme programming: First results from a controlled case study, IEEE, In null (p. 259)
(2003).
Alvarez, S.A., Barney, J.B., 2007. Discovery and creation: Alternative theories of entrepreneurial action.
Strateg. Entrep. J. 1 (1-2), 11–26.
E. Autio, S. Nambisan, L. D. Thomas, M. Wright, Digital affordances, spatial affordances, and the genesis of
entrepreneurial ecosystems, Strategic Entrepreneurship Journal 12(1), (2018) 72-95.
T. Baker, R. E. Nelson, Creating something from nothing: resource construction through entrepreneurial
bricolage, Administrative Science Quarterly 50(3) (2005) 329–366.
J.B. Barney, Firm resources and sustained competitive advantage, Journal of Management 17(1) (1991) 99-
129.
J. B. Barney, Is the resource-based “view” a useful perspective for strategic management research? Yes,
Academy of management review 26(1), (2001) 41-56.
Y. Baruch, B. C. Holtom, Survey response rate levels and trends in organizational research, Human relations
61(8) (2008) 1139-1160.
K. Beck, M. Beedle, , Van Bennekum, A., Cockburn, A., Cunningham, W., Fowler, M., J. Kern, The agile
manifesto (2001), last accessed on July 30th, 2018 at: http://agilemanifesto.org/.
A.Bharadwaj, O. A. El Sawy, P. A. Pavlou, N. V. Venkatraman, Digital business strategy: toward a next

generation of insights, MIS Quarterly 37(2) (2013) 1-18.
S. Blank, The Four Steps to the Epiphany: Successful Strategies for Products that Win, K&S Ranch (2007).
S. Blank, Why the lean start-up changes everything Harvard Business Review 91(5) (2013) 63-72.
S. Blank, B. Dorf, The Startup Owner's Manual: The Step-By-Step Guide for Building a Great Company,
K&S Ranch Consulting (2012).
J. Brinckmann, D. Grichnik, D. Kapsa, Should entrepreneurs plan or just storm the castle? A meta Analysis
on contextual factors impacting the business planning–performance relationship in small firms, Journal of
Business Venturing 25 (2010) 24–40.
42
S. L. Brown, K. M. Eisenhardt,. Product development: Past research, present findings, and future directions,
Academy of management review 20(2) (1995) 343-378.
G. D. Bruton, I. Filatotchev, S. Chahine, M. Wright,. Governance, ownership structure, and performance of

IPO firms: The impact of different types of private equity investors and institutional environments, Strategic
management journal 31(5) (2010) 491–509.
G. Bruton, S. Khavul, D. Siegel, M. Wright, New financial alternatives in seeding entrepreneurship:

Microfinance, crowdfunding, and peer‐to‐peer innovations, Entrepreneurship Theory and Practice 39(1)
(2015) 9–26.
F. G. Castro, J. G. Kellison, S. J. Boyd, A. Kopak, A Methodology for Conducting Integrative Mixed

Methods Research and Data Analyses, Journal of Mixed Methods Research 4 (2010) 342-360.
S. M. Clark, D. A. Gioia, Jr. D. J.Ketchen, J. B. Thomas, Transitional identity as a facilitator of

organizational identity change during a merger, Administrative Science Quarterly 55(3) (2010) 397-438.
A. Contigiani, D. A. Levinthal,. Situating the Construct of Lean Startup: Adjacent “Conversations” and
Possible Future Directions, Industrial and Corporate Change (2018) forthcoming.
W. A. Cram, S. Newell, Mindful revolution or mindless trend? Examining agile development as a

management fashion, European Journal of Information Systems 25(2) (2016) 154-169.
J.W. Creswell, V.L.P. Clark, Designing and Conducting Mixed Methods Research, 2nd Edition, London:
SAGE Publications Ltd (2011).
S. De Cesare, M. Lycett, R. D. Macredie, C. Patel, R. Paul, Examining perceptions of agility in software

development practice, Communications of the ACM 53(6) (2010) 126-130.
F. Delmar, S. Shane, Does business planning facilitate the development of new ventures?, Strategic
management journal 24(12) (2003) 1165-1185.
D. Diderot., Diderot’s selected writings, edited by Lester G. Crocker, Collier-Macmillan, 1966.
W. Drover, L. Busenitz, S. Matusik, D. Townsend, A. Anglin, G. Dushnitsky, A Review and Road Map of
Entrepreneurial Equity Financing Research: Venture Capital, Corporate Venture Capital, Angel Investment,
Crowdfunding, and Accelerators, Journal of Management (2017) forthcoming
https://doi.org/10.1177/0149206317690584.
K. M. Eisenhardt, M. E. Graebner, Theory building from cases: Opportunities and challenges, Academy of
management journal 50(1) (2007) 25-32.
K. M. Eisenhardt, Building theories from case study research, Academy of management review 14(4) (1989)
532-550.
T. Eisenmann, E. Ries, S. Dillard, Hypothesis-Driven Entrepreneurship: The Lean Startup, Harvard Business
School Entrepreneurial Management Case (812-095) (2012), last accessed on July 30th, 2018 at:
https://hbr.org/product/hypothesis-driven-entrepreneurship-the-lean-startup/812095-PDF-ENG
G. Fisher, Effectuation, causation, and bricolage: a behavioral comparison of emerging theories in

entrepreneurship research, Entrepreneurship: Theory and Practice 36(5) (2012) 1019–1051.
43
Forbes, You can’t teach entrepreneurship, Forbes (2012) last accessed on July 30th, 2018 at:
https://www.forbes.com/sites/babson/2012/05/24/you-cant-teach-entrepreneurship/#7a1cb2ab9917
D. L. Frederiksen, A. Brem, How do entrepreneurs think they create value? A scientific reflection of Eric
Ries’ Lean Startup approach, International Entrepreneurship and Management Journal 13(1) (2017) 169-189.
R. Garland, The mid-point on a rating scale: Is it desirable, Marketing bulletin 2(1) (1991) 66-70.
W.B. Gartner, A conceptual framework for describing the phenomenon of new venture creation, Academy of
Management Review 10 (1985) 696–706.
A. Ghezzi, A. Cavallo, Agile Business Model Innovation in Digital Entrepreneurship: Lean Startup
Approaches, Journal of Business Research (2018) forthcoming, https://doi.org/10.1016/j.jbusres.2018.06.013
A. Ghezzi, M. N. Cortimiglia, A. G. Frank, Strategy and business model design in dynamic

telecommunications industries: A study on Italian mobile network operators, Technological Forecasting and
Social Change 90 (2015) 346-354.
B.Glaser, A. Strauss, Grounded theory: The discovery of grounded theory, Sociology The Journal Of The
British Sociological Association, 12 (1967) 27-49.
A. Gustafsson, J. Qvillberg, Implementing lean startup methodology – An evaluation, Master thesis,

Chalmers University of Technology, Sweden (2012), last accessed on July 30th, 2018 at:
http://publications.lib.chalmers.se/records/fulltext/164603.pdf
K. Hartman, A summary of the book ‘The Lean startup – How today’s entrepreneurs use continuous
innovation to create radically successful businesses, (2013) last accessed on July 30th, 2018 at:
http://www.kimhartman.se/wp-content/uploads/2013/10/the-lean-startup-summary.pdf
P. Heiramo, Lean startup is a business development framework, Futurice (2013) last accessed on July 30th,
2018 at: https://futurice.com/blog/lean-startup-is-a-business-development-framework
M. A. Hitt, R. D. Ireland, S. M. Camp, D. L. Sexton, Strategic entrepreneurship: Entrepreneurial strategies

for wealth creation, Strategic management journal 22(6‐7) (2001) 479-491.
H. Holmström, B. Fitzgerald, P. J. Ågerfalk, E. Ó. Conchúir, Agile practices reduce distance in global

software development, Information systems management 23(3) (2006), 7-18.
R. D. Ireland, M. A. Hitt, D. G. Sirmon, A model of strategic entrepreneurship: The construct and its
dimensions, Journal of management 29(6) (2003) 963-989.
R. B. Johnson, A. J. Onwuegbuzie, L. A. Turner, Toward a definition of mixed methods research, Journal of

mixed methods research 1(2) (2007) 112-133.
B. Jullien, Participation constraints in adverse selection models, Journal of Economic Theory 93(1) (2000) 1-
47.
W. R. Kerr, J. Lerner, A. Schoar, The consequences of entrepreneurial finance: Evidence from angel
financings, The Review of Financial Studies 27(1) (2011) 20–55.
W. R. Kerr, R. Nanda, M. Rhodes-Kropf, Entrepreneurship as Experimentation, The Journal of Economic

Perspectives 28(3) (2014) 25-48.
44
T. Kiura, O. J. Bosch, N. C. Nguyen, S. Shirasaka, T. Maeno, Creating a New Business Through Applying
the Systems‐Based Evolutionary Learning Laboratory Approach, Systems Research and Behavioral Science
31(6) (2014) 696-707.
S. Kraus, I. Kauranen, Strategic management and entrepreneurship: friends or foes, International Journal of
Business Science and Applied Management 4(1) (2009) 37-50.
V. Krishnan, K. T. Ulrich, Product development decisions: A review of the literature, Management Science,
47(1) (2001) 1-21.
P. Lorange, Corporate planning. Englewood Cliffs, NJ, Prentice Hall (1980).
J. Meredith, Building operations management theory through case and field research, Journal of Operations
Management 16(4) (1998) 441–454.
J.F. Molina-Azorin, Mixed methods research in strategic management: impact and applications,
Organizational Research Methods 15(2012) 33-56.
S. Nambisan, Digital entrepreneurship: Toward a digital technology perspective of entrepreneurship,

Entrepreneurship Theory and Practice 41(6) (2017) 1029-1055.
S. Nerur, R. Mahapatra, G. Mangalaraj, Challenges of migrating to agile methodologies, Communications of

the ACM 48(5) (2005) 72-78.
A. Osterwalder, Y. Pigneur, Business model generation: a handbook for visionaries, game changers, and
challengers, John Wiley & Sons (2010).
K. Popper, Conjectures and refutations: The growth of scientific knowledge, Routledge Classics, Vol. 17
(2014).
S. R. Porter, M. E. Whitcomb, The impact of contact type on web survey response rates, The Public Opinion
Quarterly 67(4) (2003) 579-588.
B. Power, How GE applies lean startup practices. Harvard Business Review 13-5 (2014), last accessed on
July 30th, 2018 at: https://hbr.org/2014/04/how-ge-applies-lean-startup-practices.
M. Rappa, Business models on the web. Managing the Digital Enterprise, on-line guide, North Carolina State
University (2001), last accessed on July 30th, 2018 at: http://digitalenterprise.org/models/models.html.
E. Ries, The lean startup: How today's entrepreneurs use continuous innovation to create radically successful
businesses, Random House LLC. (2011).
E. Ries, J. Euchner, Conversations: What Large Companies Can Learn from Start-ups: An Interview with
Eric Ries. Research-Technology Management 56(4) (2013) 12–16, In: Martinez, M. S. (2016) Good
Practices of the Lean startup Methodology: Benefits, challenges, and recommendations , Master thesis, Aalto
University, Finland last accessed on July 30th, 2018 at:
https://into.aalto.fi/download/attachments/13426465/sci_2016_salasmartinez_marc.pdf.
S. D. Sarasvathy, Causation and effectuation: toward a theoretical shift from economic inevitability to
entrepreneurial contingency, Academy of Management Review 26(2) (2001) 243–263.
S. Shane, S. Venkataraman, The promise of entrepreneurship as a field of research, Academy of management

review 25(1) (2000) 217-226.
45
K. Schwaber, J. Sutherland, The scrum guide, (2011) last accessed on July 30th, 2018 at: http://scrum.org/.
A. Strauss, J.Corbin, Basics of Qualitative Research: Techniques and Procedures for Developing Grounded
Theory, 2nd ed. Thousand Oaks, CA: Sage (1998).
D.J. Teece. Business models, business strategy and innovation, Long range planning 43(2) (2010) 172-194.
G. Walsham, Interpretive case studies in IS research: nature and method, European Journal of information
systems 4(2) (1995) 74-81.
K.E. Weick, The Social Psychology of Organizing, Addison-Wesley: Reading, MA (1979).
J.P. Womack, D.T. Jones, Lean Thinking: Banish waste and create wealth in your corporation, Simon and
Schuster (1996).
X. Yang, S. L. Sun, X. Zhao, Search and execution: examining the entrepreneurial cognitions behind the lean
startup model, Small Business Economics (2018) 1-13.
R. K. Yin, Case study research: design and methods, Applied social research methods series, Thousand
Oaks, CA: Sage Publications, Inc. (1984).
Appendix 1 – Survey sections and questions
Survey section Questions Values

1. LSAs adoption 1.1 Did your digital startup adopt Lean Startup Approaches (LSAs)? – i.e. Boolean
Customer Development, Lean Startup (Yes, No)
1.1.1 Which of the LSAs did you adopted the most?– i.e. Customer List of items
Development; Lean Startup; both
1.2 What are the main reasons and motivations behind your choice to List of items
adopt/not to adopt LSAs?
1.3 How did you get to know about LSAs? List of items
1.4 What are the steps you followed to implement LSAs? List of items
1.4.1 What LSA concepts, tools and models did you use? List of items
1.5 At what stage of the digital startup’s development did you use LSAs? List of items
2. LSAs results 2.1 How long did the LSA implementation process last? Open-ended
2.2 How much did the LSA implementation process cost? Open-ended
2.3 Overall, are you satisfied or dissatisfied with the results obtained from Likert scale
adopting LSAs?
2.3.1 Are you satisfied or dissatisfied with the results obtained from the Likert scale
adoption of specific LSA concepts, tools and models?
3. LSAs advantages 3.1 What are the most significant advantages you experienced relating to List of items
and disadvantages the adoption of LSAs?
3.1.1 What are the most significant advantages you experienced relating to List of items
the adoption of specific LSA concepts, tools and models?
3.2 What are the most significant disadvantages you experienced relating List of items
to the adoption of LSAs?
3.2.1 What are the most significant disadvantages you experienced relating List of items
the adoption of specific LSA concepts, tools and models?
4. LSAs combined 4.1 Did you use any other models, tools or approaches in combination with Open-ended
with other LSAs to develop your startup?
entrepreneurial
theories, approaches
and tools
46
Appendix 2 – Interview sections and questions
Interview section Semi-structured questions Values

1. LSAs adoption  How would you describe your LSA adoption and implementation Open-ended
process? What were the steps you and your startup followed? And
when did you adopt and implement LSAs?
 Can you comment further on your reasons for adopting/not adopting
LSAs?
 How extensively did you use LSA concepts, tools and models? Is
there any particular element of the LSAs that you mostly used/mostly
did not use? Why?
2. LSAs results  Can you describe or justify the reasons that determined the length of Open-ended
your LSA implementation process?
 Can you describe or justify the reasons that determined the cost of
your LSA implementation process?
 Can you further comment on the results you obtained from the
adoption and implementation of LSAs as a whole? And of the results
related to specific LSA concepts, tools and models?
 Can you further explain why you were satisfied/dissatisfied with the
adoption and implementation of LSAs?
3. LSAs advantages  Can you discuss in more detail the main advantages you feel you Open-ended
and disadvantages obtained through the adoption and implementation of LSAs?
 Is there any specific LSA concept, tool or model that determined such
advantages?
 Can you discuss in more detail the main disadvantages you feel you
suffered from because of adopting and implementing LSAs?
 Is there any specific LSA concept, tool or model that determined such
disadvantages?
4. LSAs relation with  Can you discuss in more detail whether you used any other Open-ended
other entrepreneurial entrepreneurial model, tool or approach in combination with the LSAs
theories, approaches to develop your startup?
and tools  Why did you use these additional models? What benefits did you gain
from this combined use?
47

Jur 12

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Jur 12

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Jur 12

Uploaded by

Copyright:

Available Formats

Ghezzi, Antonio

Forthcoming in Technological Forecasting and Social Change

Please cite as:

(Denis Diderot, 1753)

1.1. Reasoning behind the study

process (Contigiani and Levinthal, 2018).

entrepreneurial theories and logics.

1.2 Purpose and research questions

launching startups (Research Question 4).

qualitative research method based on 32 semi-structured interviews, to disclose additional insights.

startups to adopt this promising - but still scientifically under-investigated - approach.

2.1 Discovery and creation theories of entrepreneurial opportunity and action

behaviour, subsequently followed by strategic or “advantage-seeking” behaviour, in order to integrate the

indicate that opportunities can be discovered or created.

interaction with customers and subsequent iteration (Fisher, 2002).

precursor, Customer Development (Blank, 2007).

2.2 Lean Startup Approaches

Development (Blank, 2007) and the Lean Startup (Ries, 2011).

four sub-steps (see Figure 1).

assigned to a business model, as shown in Table 1.

Pillar Parameter Description

Figure 1. Customer Development (Blank and Dorf, 2012).

build/measure/learn loop” (Ries, 2011).

et al., 2012), thereby introducing a third element.

acquire customers (Gustafsson and Qvillberg, 2012; Heiramo, 2013).

Figure 2: Lean Startup Process Steps (Eisenmann et al., 2012)

Levinthal, 2018) that needs to be backed by empirical investigation.

3.1 Survey on Digital Startups

startups associations and the startups themselves.

options, to come up with the final version of the survey.

The full list of survey questions is given in Appendix 1.

the sample across different fields of application.

Field of application Number of % of sample Average age (in years)

3.2 Interviews with significant digital startups

1955; Yin, 1984).

strong similarities or differences when compared to other startups.

startups contacted in the qualitative research step).

average 1 hour and 8 minutes.

to ask for clarification.

analysis, and stored in a software archive.

Selection of codes and concepts Theme

conducted on a subsample of the survey’s respondents.

4.1 Lean Startup Approaches - adoption

Research item Results

Table 4. Quantitative results on RQ1 – LSA adoption

launch and development of their startup.

need to test that out”.

ASAP and lean startup is just what I need”.

need to burn less cash”.

4.2 Lean Startup Approaches - results

Table 5. Quantitative results on RQ2 – LSA results and outcomes

terms and as a percentage of the capital raised.

– as the CEO of a digital wearables startup admitted.

to know about your company and its market”.

4.3 Lean Startup Approaches - advantages and disadvantages

Research item Results

Table 6. Quantitative results on RQ3 – LSA advantages and disadvantages