The effect of sustainable business models on network density within startup alliances in the entrepreneurial ecosystems of Amsterdam and Eindhoven.

(1)

Faculty of Economics and Business

The effect of sustainable business models on network

density within startup alliances in the entrepreneurial

ecosystems of Amsterdam and Eindhoven.

Author: Thijs van der Wijk

Student number: 11880600

(2)

Statement of Originality

This document is written by Thijs van der Wijk who declares to take full

responsibility for the content of this document.

I declare that the text and the work presented in this document is original and that

no source other than those mentioned in the text and its references have been used

in creating it.

The faculty of Economics and Business is responsible solely for the supervision

of completion of the work, not for the contents

(3)

Acknowledgement

This thesis is the last stage of my bachelor’s degree at the University of Amsterdam. I would like to thank Mr. Roel van der Voort personally for guiding me through the entire process, while entrusting me with a great deal of independence. His suggestions and insights resulted in an enjoyable writing process. Due to the Corona virus outbreak, we had to adjust to an entirely virtual means of communication, mainly via zoom. This turned out to be a rather smooth and effective transition and did not hinder his role as supervisor. Additionally, I would like to show great appreciation for his role in the admission process for my MSc in the coming year.

This research provided me with an opportunity to put the years of studying and the development of skills into practice. While I really enjoyed writing this thesis, the most interesting part was to grasp a greater understanding of the entrepreneurial landscape and the role of entrepreneurial ecosystems on (sustainable) startups and innovation. Therefore, I would like to thank all respondents who have taken the time to complete the survey and contributed to new literature on the role of networks within startup communities. I would like to thank one of these respondents, Mr. N. Schoen, for his contribution by taking the time to extensively describe the networks of sustainable startups and initiatives that are currently being developed in the Netherlands.

I hope that whoever reads this thesis, does so with contentment and curiosity. Kind regards,

(4)

Abstract

This research focusses on the effect of sustainability on network density within startup communities in the entrepreneurial ecosystems of Amsterdam and Eindhoven. Previous research by Neumeyer & Santos, (2018) and Cohen, (2005) suggested that sustainable startups benefit from a denser network because it enhances the diffusion of knowledge about unique challenges that sustainable startups face during the early stages. In the past decade, the potential impact of integrating sustainability has gained traction in the governance of such ecosystems. While it would be interesting to analyze network and sustainable conditions throughout the entire ecosystem, this study focusses on startup communities to improve feasibility.

In this study, members of entrepreneurial ecosystems participated in a between-subjects design (sustainable vs. conventional startups). The outcome variable measured characteristics of the network in startup communities, e.g. network density, network frequency, and times spent networking. While both an ANOVA analysis and an independent-samples T test did not yield significant results, the means of all variables measuring the network characteristics were higher in the sustainable group. This consequentially suggests that sustainable startups do indeed have denser networks. The apparent mean differences between groups, despite having no significant findings, could be the result of the small sample size that was obtained (N=23). Theoretical implementations, limitations and future suggestions are subsequently discussed in this study.

Keywords: Entrepreneurial ecosystems, Sustainability, Sustainable startups, Startup

communities, Networks, Network density, Network frequency, Innovation, Amsterdam, Eindhoven.

(6)

1. Introduction

While the term ‘entrepreneurial ecosystems’, or EE for short, has gained popularity of the past decades, with Silicon Valley being a widely used example, the precise structure that comprises such an ecosystem remains to a large extent unexplained (Audretsch & Cunningham, 2018). It is argued by Van De Ven (1993) that an entrepreneur by itself cannot orchestrate all the resources, institutions, and other business functions into a fully functioning entrepreneurial venture. He proposes that it is the collective achievement of various inter-connected entrepreneurs in the public and private sectors that provide the ecosystem with a framework, which in turn facilitates the emergence of successful startups.

A great deal of the published research has therefore focused on the typology or classification of the different components in these EEs and the information- /communication-channels that make them interact with each other (Stam & Van De Ven, 2019; Roundy & Bradshaw, 2018; Cohen, 2005). Some of the most prominent and obvious actors have already been identified, such as, universities, government policies, and venture capital. However, some research suggests there is a vast amount of undiscovered, smaller components in the ecosystem that could influence entrepreneurial behavior and startup success, such as public transport infrastructure, firm competition & collaboration, and geographic location (Acs, Stam, & Audretsch, 2017).

While the majority of the research has focused on traditional EEs, there is a growing interest in alternative ecosystems. The world has come to an understanding over the past decades that the way we are consuming and producing is unsustainable in the long run (Baldassarre et al., 2017). Sustainability — 'namely, innovation and development patterns that

meet current human needs without compromising future generations' ability to meet their own'

(Brundtland, 1987) — is a path we are just learning to walk. If we wish to keep walking on this finite planet, some level of action is required (United Nations, 2020). One of the main critiques against the traditional economic ecosystems is the exclusive focus on efficient resource allocation and the neglect of social welfare and carrying capacity of the world's natural ecosystems (Daly & Farley, 2003). Therefore, the development of an ecosystem that fosters innovation, production, and other operational practices in a sustainable way could be a feasible measure that minimizes, to some extent, the negative impact that we have on the future of our planet (Nidumolu, Prahalad, & Rangaswami, 2009; Neumeyer & Santos, 2018).

(7)

Sustainability plays a crucial role in the development of our society, and it is something that has been subjected to extensive research in the past. These studies found many factors that have a relationship with sustainability, and one of these drivers is startup innovation. (Brundtland, 1987; Nidumolu, Prahalad, & Rangaswami, 2009; Neumeyer & Santos, 2018). In the past, it was often assumed that startup rates and innovation of a country were positively linked to one another. However, a recent study found that there is only a positive relationship between startup rate and innovation in countries that are in a later stage of economic development and that it has a negative relationship in developing countries (Anokhin & Wincent, 2011).

Entrepreneurial ecosystems have been the birthplace for many conventional startups, guiding them through their early stages of development. While network theory concerning EEs is still in its infancy stage, according to Neumeyer & Santos (2018), this is especially true for startups that have a sustainable business model that lacks theoretical grounding. Interestingly, in their research, they also found that there is a significant difference between the network density of conventional and sustainable startups. Startups with a sustainable business model have stronger instrumental ties (e.g., funding, evaluation, professional assistance) and societal ties (e.g., family, friends, advice). Therefore, one could say they are more connected. An explanation for this could be that sustainable startups need a stronger network in order to succeed because their business models require them to. As Cohen (2005) theorized in his research, sustainable startups rely on a network of various stakeholders who understand the unique challenges which sustainable startups face. Not only that, but green startups also face more costs because they, for example, have to invest in an EMS system early on (Ammenberg & Hjelm, 2003). Therefore, sharing information and costs with players within their network could be beneficial. The following research questions was thus formed:

Is the network density of sustainable startups greater than conventional startups in entrepreneurial ecosystems in Amsterdam and Eindhoven?

(8)

2. Theories and hypotheses

2.1 Defining Entrepreneurial Ecosystems

Before diving into the empirical theories about the components and networks, one has to understand what an ecosystem is. In abstract, and ecological ecosystem is a ‘biotic

community, its physical environment, and all the interactions possible in the complex of living and nonliving components’ (Tansley 1935). While biology offers us a precise definition, the

definition of a business ecosystem is not well defined yet. Previous research, like Stam (2014), has gradually formulated the Entrepreneurial-Ecosystem-Approach to highlight the roll of entrepreneurship within an ecosystem. It defines EE by its two components. ‘Entrepreneurial’ referring to entrepreneurship, which is the action of an individual to seize an opportunity for innovation to create novel goods and services. This, in turn, brings new value to society. The EE-approach has gained popularity over the past decades due to the big impact EE’s have on their regional environment and other studies often focuss on high-growth start-ups (World Economic Forum, 2013). In these studies, it is argued that this categorization of start-ups is an important driver of innovation, scaling and employment. However, according to E. Stam & Spigel, (2016) this approach is too narrow, because it excludes networks of innovative start-ups or communities consisting of entrepreneurial employees as a form of productive entrepreneurship. This new classification of productive entrepreneurship in an EE is increasingly being emphasized in modern literature (Stam et al., 2012; Crumpton, 2012; Mason & Brown, 2013; Henrekson & Sanandaji, 2014).

The second term, ecosystem, has a more subjective meaning to it, because the boundaries and components of an ecosystem are harder to define, and all actors have not been subjected yet to extensive research. Therefore, it is easier to think of the ecosystem as an interdependent community which is geographically bound (Audretsch & Cunningham, 2018). Social context, like culture and networks, further mold the community into a unique entrepreneurial environment. While the dissection of the term EE helps in explaining what it is, the modern definition of an entrepreneurial ecosystem emphasizes the role of entrepreneurship. It recognizes that the EE needs entrepreneurship for generating and maintaining the ecosystem, instead of just seeing entrepreneurship as an outcome of such an ecosystem (Stam, 2014). Therefore, this paper defines an entrepreneurial ecosystem as a dynamic, stable community of interdependent actors that both influence and are influenced by entrepreneurship within a geographical location.

(9)

Further research by Stam, Romme, Roso, Van Den Toren, & Van Der Starre, (2016) argues that a distinction can be made between two elements within an EE. Firstly, framework

conditions which can either be social or physical. Social conditions consist of formal or

informal institutions, like universities and talent pools. Physical conditions are elements which can enable or constrain human interaction, like infrastructure/public transport or access to buyers, which tells you something about whether the start-up has a target market within the EE. Fundamentally, these conditions are the source of value creation within an EE, but in order to understand how these conditions foster value creation, one needs to understand the systemic

conditions that enable entrepreneurial activities (E. Stam et al., 2016). Systemic conditions and

the interactions between them are, according to recent research, crucial elements within an EE that determine its success. These elements are; networks of entrepreneurs, leadership, knowledge, talent and support services. Firstly, networks of entrepreneurs are vital in an EE because it enables information flows, distributions of knowledge, labour and capital. Secondly, leadership plays a crucial role because helps in building and maintaining a healthy ecosystem. Thirdly, knowledge is an important source in recognizing opportunities and making effective decisions. Fourthly, talent provides an EE with labour resources like students, entrepreneurs and other agents. Lastly, support services which lower the barriers to enter specific markets and thus reduce time to market (E. Stam et al., 2016).

(10)

2.2 Network Theory and Density

Network theorists have been studying exchange relations between networks and the components that use the network (Witt, 2004). Usually, there is an individual or an individual organization, that shares information with another individual or organization. These organizations could be universities, authorities or firms. Whenever the studied entity has more than one of these contacts, it is called a network. According to Van Gelderen, (2017), networks can bring substantial benefits to companies, because of the social capital they represent. These benefits can come in a material form like; financial capital, suppliers or a helping hand. However, they can also be immaterial like; advice, information or experience which could accelerate the growth-stage of start-ups. Giving the importance and benefits of networks in ecosystems, understanding how actors can built and maintain strong network ties could provide start-ups with a competitive edge (Burt, 2018).

Since networks in general can reap many benefits to its actors, there is a common understanding that a network within an EE can influence a startup's economic and innovative performance (Gilsing et al., 2008). Even the most researched innovative hub, Silicon Valley, has invested heavily in the creation of high-density hubs of innovative activity (Meagher & Rogers, 2004). As discussed earlier, an EE is a community of interdependent components. The network within such an ecosystem is the intangible entity that makes these components interact and disseminate information. Researchers have developed a multitude of quantitative measures to study network activity. A few examples provided by (Witt, 2004) are: “A central person,

e.g. within an information network, has many direct connections to other persons (‘connectedness’), can reach other members of the network quickly, i.e. needs to use few or no intermediate persons (‘closeness’), or is located on the information paths between other persons of the network frequently (‘betweenness’).” Due to time constrictions and the general

feasibility of this research, this study will be looking at the density of self-centered networks in entrepreneurial ecosystems. Self-centered network density allows for an interpretation of the structure of a specific network, because it refers to the degree to which an individual in the network is connected to potential partners in the network (Van Gelderen, 2017). By comparing the network density of alliances of sustainable start-ups and alliances of conventional start-ups, interpretations can be made about the possible interdependence of sustainable start-ups in an EE, as mentioned by (Cohen, 2005). Density is measured by analyzing all players within a network and their relation to one-another. When a network is dense, this means that almost all players in the network know each other and vice versa.

(11)

The role and impact of different levels of network density are still subject to academic research. A study by Gilsing et al., (2008), for example, argues that a higher level of network density will reduce novelty creation. However, it might enhance the build-up of absorptive capacity, which is theorized to have a more significant impact on sustainable startups than conventional ones (Gilsing et al. 2008; Cohen, 2005). Because of the potential trust-building in network alliances, they also argues that it enables the proliferation of triangulation. Moreover, because of the mutual connectedness, richness, and reliability of information within a network, members will also enjoy profits from triangulation.

2.3 Network density as a key element of sustainable start-ups

Neumeyer & Santos, (2018) and Cohen, (2005) mentioned in their research that sustainability and its relation to key elements of EE’s is a subject that requires more research. Due to the growing urgency of sustainable solutions to reduce the impact of society on the vitality of our planet, this research will focus on one of the systemic conditions. Namely, whether one of the sub-categories of networks, network density, is a more prominent element for sustainable start-ups in an EE than for conventional start-ups. If the network of a start-up is proven to be a prominent element in a business model, future research could examine whether “network” should be one of the building blocks in the Business Model Canvas. It is currently too early to deduct, but as stated by Lund & Nielsen, (2014): “it is an interesting

proposition to study whether new network-based business models factoring in openness, peering, sharing, and global positioning, could enable the possibility of enhancing the value proposition while at the same time reducing costs through partnering”.

Neumeyer & Santos, (2018) discovered in their research that sustainable start-ups had a more connected network than conventional ones, meaning that then had a higher network density. However, in their limitations they pointed out that their sample was limited in size and geographical scope. Therefore, this study will add to the current literature by assessing the role of network density in the entrepreneurial ecosystems of Amsterdam and Eindhoven. Furthermore, a larger sample size will be used, because the scope of this research only focusses on one key element of the systemic conditions. While the scopes of the studies by Neumeyer & Santos, (2018) and Cohen, (2005) were explorative in nature, inferences can be made about their results. These inferences led to the following hypothesis:

(12)

Due to the incomplete literature on key elements of EE’s and their relationship to sustainable start-ups, a more thorough analysis will be performed on network density. Although network density is one of the most popular measurements of networks in EE’s, Witt, (2004) proposed more measurements of networks that address certain specifications. Therefore, the specifications that will be added to this study are network frequency and time spent on networking. Network frequency measure how extensive the contact between nodes of the network is, usually measured in number of contact points per week. Time spent on networking is a variable that describes the duration of contact in the network and is measured in hours per week. The addition of these variable, in order to create a more comprehensive analysis, created the following hypotheses:

H2: Within the EE of Amsterdam and Eindhoven, network frequency will be higher amongst sustainable start-ups than conventional ones.

H3: Within the EE of Amsterdam and Eindhoven, time spent on networking will be higher amongst sustainable start-ups than conventional ones.

2.4 Sustainable and conventional business models

Now that the potential role of network density within an EE is described, what rests is a clear classification of sustainable and conventional business models. In the purest form, a business model is a firm's plan that defines its operations in order to create and capture value (Osterwalder & Pigneur, 2013). Although there are a plethora of different business models that startups can deploy, a generalization has to be made in order to clearly define the difference between a sustainable business model, and a conventional one.

There are numerous suggestions by scientists studying this field, and more measurement methods are being proposed every year (Patzelt & Shepherd, 2011). However, one thing most scientists can agree on is that sustainable behavior should at least have some effect on the triple bottom line or TBL. The three items in TBL are environmental, financial, and social outputs (Schaltegger et al., 2016). Therefore, these three items are going to be the guidelines that this research will uphold when assessing whether a startup is sustainable.

(13)

2018). When talking about the interpretation of financial performance in sustainable startups and conventional startups, conventional startups purely focus on financial performance as a means of generating profit with only intrinsic motivation. When researchers talk about financial performance in sustainable startups, they see it as a component that positively affects social welfare, because a part of the financial success is shared amongst the society in which it operates (Milne & Gray, 2013). In theory all companies share some of their financial success amongst society, and example being the creation of jobs (Schaltegger, 2016). Therefore, in order to create a conceptualization of a sustainable start-up, this paper will be analyzing the start-up’s operations. Cooney, (2009) developed 4 steps to assessing whether a firms is sustainable: 1. The start-up incorporates sustainability in its decision-making process. 2. The start-up offers environmentally friendly products or services, which lower demand for traditional businesses. 3. The start-up is “greener” than its competitors. 4. In the start-up’s business operations, it is committed to the principles of environmental sustainability.

(14)

3. Methods

As mentioned earlier, in order to answer the research question and the corresponding hypotheses, a survey has to be developed, examining: 1. Network density. 2. Network frequency. 3. time spent on networking. 4. The business model’s degree of sustainability. 5. Stage in business lifecycle. 6.Group size and additional information. It is important to mention that the entire scope of this research falls within the entrepreneurial ecosystems in Amsterdam and Eindhoven. Therefore, all respondents will be gathered from within these ecosystems. Amsterdam and Eindhoven were chosen because there is no substantial difference in their knowledge network structure (E. Stam et al., 2016). Additionally, Amsterdam and Eindhoven developed an approach that continuously involves key stakeholders in the regional governance of the entrepreneurial ecosystem, called ‘triple helix’. Both EE’s also differ on some levels, creating an opportunity to analyze differences between the ecosystems. While both ecosystems are famous for the generation of a lot of tech companies, Amsterdam has a lot of online technology companies (e.g. platforms, cyber security and digital marketing). Eindhoven, on the other hand, is known for its high-tech ecosystem with a great number of technology giants fostering innovation and entrepreneurship, like ASML and Philips. These characteristics, in regard with the accessibility of both cities for data collection, resulted in the selection of these cities.

3.1 Sampling approach

In order to formulate a high-quality methodology, purposive sampling was used to generate samples which are solely founders or co-founders of start-ups. Through desk-research and start-up alliances, a sample target of 50 founders of a start-up with a sustainable business model and 50 founders of startups with a conventional business model was preferred. However, after extensive mailing, calling, and reminding for 3 weeks, 56 responses were collected, representing 30 sustainable startups and 26 conventional ones. Furthermore, collected companies are between one and five years old, in order to collect startups only. All founders established their startups in the entrepreneurial ecosystems of Amsterdam and Eindhoven.

One Facebook group that was used is the “Duurzame Jonge 100”, which is a list-maker of the top 100 sustainable entrepreneurs in the Netherlands. From another group, Bewust Amsterdam, an eco-entrepreneur responded and allowed me to use his network in order to reach many sustainable entrepreneurs. To reach founders in Eindhoven, the TU/Eindhoven was contacted, which gave me permission to send the survey to their innovation space, which is an

(15)

open community of students, professors and industry specialist that fosters innovation and entrepreneurship.

Lastly, in order to get a desired number of respondents to ensure validity of the results, GetProspect was used to collect a large email-list. GetProspect is a tool which functions as an email-address scraper plugin for SalesNavigator. SalesNavigator is a tool from LinkedIn which allows the user to filter businesses and employees on the basis of location, job title, years in current position, size of the company, main operations and keywords. Four different entries were used to collect 50 mail addresses of all categories. For every entry, the tags CEO-Co-founder-founder and 1-5 years in current job title were used. The first distinction was then made by adding Amsterdam and Eindhoven to the location tag. The second distinction was then made by adding the keywords “sustainable” and “sustainability”.

3.2 Sustainability measures

• In order evaluate the business model of a start-up, survey questions will be used which are developed and tested by a study by Bocken, Short, Rana, & Evans, (2014). However, a simplification of this survey will be used. The survey is simplified because this shortens the overall survey, possibly increase the response rate. Furthermore, it is not necessary in the scope of this research to examine the degree of sustainability, because just knowing whether a business model is sustainable is enough. In the research by Bocken et al., (2014), they defined different archetypes of sustainable businesses. These archetypes are classified in higher order rankings on the basis of their effect on the triple bottom line. Normally, for each different archetype and extensive method of questioning is applied, resulting in lengthy surveys. However, this survey will simply ask whether the core-business of the startup is tackling on of the nine categories and from there deduce whether a startup is considered sustainable or not. The nine archetypes are:

• Decreasing Carbon Footprint

• Reducing Energy Consumption

• Improving Product Recycling Rate

• Savings due to conservation/improvement efforts

• Environmental performance of suppliers

• Decreasing Supply Chain Miles

• Decreasing Water Footprint

• Improving Waste Reduction Rate

(16)

3.3 Measures for Network Density, Network Frequency and Time Spent Networking

Network density can be complicated to measure, due to the size networks can have. If one, for example, would like to measure the network density of an entire entrepreneurial ecosystem, one would have to map and survey everybody in the ecosystem (Burt, 2018). This would include contacting universities, governments, venture capital, to name a few. Due to the complicated measurement of network density, this study focused on the network density of startups within a startup alliance or community. By reducing the network size, it was hypothesized that it would be easier to contact enough network members of each specific group to give a more accurate proxy of the actual network density of these groups. Multiple startup communities were contacted, and some allowed me to send the survey in their community chat. However, the response rate turned out to be marginal, with 15% at most completing the survey. Due to the difficulty of generating a valid number of inter-alliance respondents, a new strategy was formed. Newly collected respondents were no long required to be from the same startup alliance. The survey was adapted to this new approach by integrating more questions that capture characteristics of the network instead of solely focusing on the direct connections of all group members. These questions resulted in a proxy for network density instead of the actual measure of it. When the respondents provided their group size and the number of members, they have a direct connection with, network density can be calculated by dividing the direct connections within a group by the total number of possible connections. Questions asked to measure network density were, for example, “How many direct connections do you have in the

group?”. In order to generate more details about network density, 5-point Likert-scale

questions were asked, such as, “group members know each other well” or “group members

often share information with one another”.

Network frequency is a variable that gives more insights in the nature of the network. It is means of measuring the intensity or extensiveness of networks, according to Witt, (2004). Questions for network frequency are, for example, “how many times per week do you contact

somebody from the group?”. Lastly, time spent network is another way to address certain

network specifications mentioned by Witt, (2004), which tells us more about the intensity and duration of network activities. Question to measure time spent networking are, for example, “How many hours per week are you in contact with somebody from the group?”.

(17)

4. Results

First, this chapter explains the preliminary analysis. Data was recoded, variables were tested for reliability and descriptive statistics were calculated. Second, it presents a thorough analysis of the main effects.

4.1 Data cleaning, recoding and correlations

The data was first checked for missing values. In order to answer the next question, respondents had to answer the previous one. Therefore, there was no missing data in the dataset. Multiple variable had to be calculated by merging one or more variable. In order to test the first hypothesis, three variables were used to asses network density. Network density itself was calculated by dividing the number of in-group direct connections by the group size. A network density proxy was computed by reverse coding inverted questions, combining all network-density-related Likert questions, and then dividing that by the number of questions. Then last variable did not have to be adjusted. The second hypothesis used most variables as they were presented. A novel variable to analyze H2 was created by dividing a network frequency variable by the group size, which is strongly correlated with network density but generated slightly more significant results. A possible explanation could be that it amplifies network density. For example, if a network is denser, the frequency of contact might not increase at a constant rate, but rather towards the exponential spectrum of a line. Lastly, the variables that measured time spent networking were not altered except for Likert-scale variables.

As shown in table 1, most variables correlate significantly, indicating that they related to one another and might be useful in testing the hypotheses. A possible theory to why that is, is because all variables measure aspects and characteristics of networks, therefore they might be related to one another. The exception is the variable which measures the length of an average conversation. This could be due to the fact that the length of a conversation is to a large extent

(18)

conclusions can be made. Variables 1-3 are used to test the first hypothesis and had a Cronbach’s Alpha (8 items; α = .928), indicating high internal consistency. Reliability of variables 4-6 which were used for hypothesis two, generated a Cronbach’s Alpha of (5 items; α = .891), also indicting high internal consistency. Lastly variables 7-9 resulted in a Cronbach’s Alpha (3 items; α = .322). Consequently, after recoding the variable, which was in hours, to minutes, the Cronbach’s Alpha was (3 items; α = .754).

While these dependent variables all correlate with one another, there was no significant correlation found between the independent variable, sustainable/conventional (now called SustConv), and all outcome variables. This could have been due to the small sample size (N=23). A suspicion for the existence of a relationship between SustConv and the outcome variables, despite the correlation matrix suggesting otherwise, is that after interpreting the different means of the groups, a clear pattern arises where sustainable startup founders essentially always score higher than their counterparts. It is also notable that all Cronbach’s Alpha’s of SustConv and the dependent variables of the three hypotheses suggest a positive relation. The Cronbach’s Alpha for H1, H2 and H3 are 0.631, 0.772 and 0.672 respectively.

In order to perform an independent-samples T-test or an ANOVA, certain assumptions must be met. For instance, the dependent variable must be normally distributed within each group. The Kolmogorov-Smirnov and Shapiro-Wilk test for normality suggested that the variables were not normally distributed. However, this could be due to the small sample size that was used. The Q-Q plots were evaluated to see if most data points were centred around the baseline, which was the case. Therefore, one can conclude that the variables are normally distributed and fit for an analysis.

(19)

4.2 Analyses of the results Table 2. Independent-samples T test results

Table 2 shows the results of the independent-samples T test for all variables testing the three hypotheses. The first three variables correspond to hypothesis 1, the second three hypothesis 2 and the last three hypothesis 3. Only one of the variables shows a significant effect, which is the variance of the minutes of contact with the group each week. This means that the data points around both respective means are spread out unevenly. Since there is no significant difference in the means of both groups, this statistic can be discarded, since it could only be of use to interpret the possibility of an incorrect conclusion when the means would significantly differ from one another. All others not being significant leads to the conclusion that a there is no significant difference of network density, network frequency and time spent networking for startups which have a sustainable business plan and startups with a conventional one. Interestingly, however, there are some findings from the analysis that enable a deduction of certain traits and possible differences between groups. For example, all mean differences are negative, which indicates that all means of sustainable startups are higher than the means of conventional ones, since mean difference is calculated by subtracting the means of the second group (sustainable) from the first group.

(20)

4.3 Hypothesis 1 – Network Density

H0: µsustainable - µconventional = 0 H1: µsustainable - µconventional ≠ 0

As seen in table 2, the F-statistic for the first three variables is not significant, indicating that we can assume that the variance on both groups is equal. Additionally, the t-statistic of the test for equality of means is also not significant, indicating that there is no significant difference in the means of both groups. It is, however, interesting to note that while the f-statistic is not significant for the Likert-scale questions that measured the Network density proxy, the t-value of that variable is almost significant (p = 0.067) if an alpha is taken of α = 0.05. This would mean that there is a significant difference in the means of both groups. The questions that were used to generate the NetDensProx, were subjective questions about their respondent’s perception of in-group network density. For example; “everybody in the group knows each

other well” and “I often ask for information in the group”, were used. Since the mean difference

is 0.784, we can deduce that they score almost one point higher, on average, on these Likert-scale questions. Although not significant, the same can be said for Network density and Network percentage. In for network density, the sustainable group had a means that was 0.113 higher than the other group, translating to a 11.3%, and a percentage score of the other variable, which intended to measure the same condition, of 10.9%.

4.4 Hypothesis 2 – Network frequency

As seen in table 2, both the test for equal variance and equal means are not significant, implying that there is no difference in the variance of both groups and the means. However, like mentioned in the paragraph above, all means of the sustainable group turned out to be higher than their counterpart. Notably, the f-statistic and t-value of network frequency were almost significant, corresponding to (p = 0.060) for the f-statistic and arguable (p = 0.066) or (p = 0.063) for the t-value. The reason for both statistics just barely being significant, could be attributed to the samples size of 23, the sustainable group having 12 respondents and the conventional group 11. The sample size seems to be the biggest bottleneck of the analysis, resulting in no significant findings. If, for example, the sample size would be twice as big, all the values that are on the border of the α = 0.05 level, might become significant. The value of

(21)

4.5 Hypothesis 3 – Time spent networking

Like mentioned before, the only significant finding of the last three variables, is the f-statistic of the average conversation length with (p = 0.022). Therefore, we can assume that there is an inequality of variance in both groups, but the null hypothesis cannot be rejected, and the means of both groups remain statistically indifferent. All means of the sustainable group, just like in the other variables, were higher than the means of the conventional group. The average member of a sustainable startup group spends an extra 18 minutes on actively networking each week, 30 minutes more on contacting their alliance and the average conversation length is 5 minutes longer than in the conventional groups.

While almost all findings are not significant, in an effort to still provide novel insights, the discussion section will now switch to a more explorative approach in evaluating the data.

As there were no significant findings in the main hypotheses of this study, there were also no significant differences found between the respondents from Eindhoven and Amsterdam. The sample size simply has to be substantially larger before and inferences can be drawn from those distinctions. Additionally, the mean differences between cities were not as spread out as the differences between sustainable startups and conventional ones. Therefore, no further analysis was performed on the inter-city differences.

(22)

5. Discussion & conclusion

Table 3 summarizes the hypotheses that were derived from the research questions. Previous literature on networks in entrepreneurial ecosystems pointed out that sustainable ventures often had a greater network density than their counterparts Neumeyer & Santos (2018). Unfortunately, it was not the case in this research. However, that does not mean it can be discarded. Due to time constraints, the contractibility of founders and the pandemic that is shaping the year 2020, this research was subjected to some limitation which will be discussed later on. Before talking about limitations and conclusions, it is interesting to take a look at the group statistics of the variables and interpret the results.

One of the most interesting observations that can be made in the analysis, is that all means of the variables were higher in sustainable group. Of course, there is no significance, so one cannot rule out that this happened merely due to chance. But because of the consistency of the results, it could he hypothesized that; indeed, something is going on. The consistent results in mean difference and the fact that all variables, except conversation length, correlated with one another leads to the suspicion that the sample size might be the inhibiting factor. While 56 people responded to the survey, only 23 were useful. 33 respondents had to be deleted because their company was 6+ years old, they were not part of a startup group, or they were managers. Research suggest that, while it is still possible to analyze, a sample size of N<30 is undesirable. For further research, it would be interesting to keep this study largely the same, but drastically increase the samples size. As show in table 4, interesting information could be derived from the functioning of network within startup alliances in entrepreneurial ecosystems.

(23)

5.1 The effect of sustainable and conventional business models on network density

Interpretations of the first three variables measuring network density, show that on average the sustainable group has a Network density of 0.43, which is a 34.4% increase of the network density of conventional startup groups. The Network density was calculated by dividing the number of direct connections they had by the group size. The following variable, measuring the percentage of group members, confirms the inference that the network of the sustainable group is about 10% more dense. The variable Network Density Proxy was formulated by summing up all Likert-scale questions that address network density and dividing that by the number of questions. It was a 5-point scale going from strongly disagree to strongly agree. Therefore, although not significant, one could infer that the conventional group was somewhat indecisive on how well all group members know each other, scoring 3.3 which is slightly above neither agree nor disagree. The sustainable group, on the other hand, moves more towards the positive side of the scale, implying that they somewhat agree on that most group members know each other. Although the means are quite interesting in this case, the relatively large standard deviation shows that the entries of the respondents are widespread in all three variables, meaning that there is a lot of variation within the groups.

(24)

5.2 The effect of sustainable and conventional business models on network frequency

The three variables measuring Network Frequency show the same pattern as did the variables for network density. Although not significant, the sustainable group contacts their alliance almost twice as much as their counterparts. The Density of Network Frequency variable cannot be evaluated the same way as network density, because their scales differ. When network density measures density on a scale from 0.0 to 1.0, network frequency can possibly exceed that. You cannot know more people from your startup group than there are people in the group. However, if one would contact almost every group member once a week and one group member twice a week, the value would exceed 1.0. Still, an inference that can be made from that variable, is that they contact their group members, relative to the group size, about twice as often as conventional startup founders. The proxy of density of network frequency was also measured on a Likert-scale, but the scale had the headings; Never – Once a month – Once a week – 2-6 Times a week – Daily. Although this scale is not distributed evenly, one could deduce that sustainable founders, on average, contact their group members more towards once a week, while conventional founders do so once or twice a month.

5.3 The effect of sustainable and conventional business models on time spent networking

Lastly, the three variables measuring Time Spent Networking, show that sustainable startup founders contact the group members almost almost 20 minutes longer every week. A difference was made between contacting the alliance and actively networking, meaning the attempt to actively strengthen one’s network. In that case, the findings suggest that sustainable founders’ network 30 minutes more every week. This would be in line with the Neumeyer & Santos (2018), stating; that startups with a sustainable business model have stronger instrumental ties (e.g., funding, evaluation, professional assistance) and societal ties (e.g., family, friends, advice). Therefore, one could say they are more connected. Maybe these founders focus more on actively networking, because they face unique challenges that require them to have strong social ties in order to increase their competitive advantage. The average length of conversations was also longer in the sustainable groups, but the increase was relatively negligible.

(25)

5.4 Limitations and recommendations for further research

One of the biggest limitations of this study was the sample size. Over 300 startup founders were contacted via phone, mail and Facebook, and a response rate of about 19% is close to the average response rate of similar surveys. Partly due to the reduction from a sample size of 56 to 23 usable respondents, the desired significance level was not reached. When dealing with low samples sizes, inputs that deviate a lot from the population means can drastically alter the sample mean that is eventually formed. In the short timeframe of this research, convincing founders of startups turned out to be harder than I initially though. Some founders that were contacted to ask them about the study often noted that most founders do not take the time to take surveys because there are plenty of other things to do. This could especially be true during the pandemic we are facing, which leads to many founders having to save their companies from going bankrupt. On another point, after lengthy contact with the University of Eindhoven I was allowed to share my survey in their Innovation Space community, which increased the number of respondents from 29 to 56. I would recommend future research to focus more on gaining respondents through organizations like that, instead of Facebook groups and scraping emails.

A second limitation is the possibility of the conventional startup finding not being generalizable all conventional startups. About 30% of the conventional startups were working in online marketing, decreasing the heterogeneity of that sample group. While this does not definitively change the outcome of the study, future research could focus on generating a more heterogenous conventional group to make the results of their study more generalizable.

5.5 Conclusion

As mentioned by Acs, Stam, & Audretsch, (2017), the obvious components of entrepreneurial ecosystems are researched more and more often, but there remains a vast number of smaller components that could contribute to the effectiveness and structure of EE’s.

Network density, although not proven significant in this study, does appear to have positive relationship with a certain level of sustainability of a company and is a widely under-researched topic. Novel measures were generated that allowed for a more thorough explanation about the characteristics of networks in startup communities. The averages of sustainable startups scored higher on all variables that were used. When the samples size would be increased and the analysis still yields the same results, one could infer that sustainable startups could be benefit more from denser network than their counterparts. Universities could then, for

(26)

networks. This would simultaneously combat one of the main critiques against EE’s, which is the neglection of social welfare and carrying capacity of the world's natural ecosystems (Daly & Farley, 2003). By fostering these startups specifically and creating a combined focus of increasing startup success and increasing the number of sustainable startups that are established in ecosystems, this implication could be a feasible measure that minimizes, to some extent, the negative impact that we have on the future of our planet Neumeyer & Santos, (2018).

This research also successfully elaborated on the suggestions given in Witt, (2004), where he argues that novel means of measurements for networks have to be formulated. To complete a thorough analysis of network density of entrepreneurial ecosystems, the sheer size of these ecosystems constrains the possibility to accurately measure them. When measures for characteristics of networks are simplified and the scope of studies is more focused on small components, this research suggests there is still a lot to be discovered about the functioning of certain networks.

(27)

6. References

Acs, Z. J., Stam, E., & Audretsch, D. B. (2017). The lineages of the entrepreneurial ecosystem approach. Small Business Economics, 49(1), 1–10.

https://doi.org/10.1007/s11187-017-9864-8

Ammenberg, J., & Hjelm, O. (2003). Tracing business and environmental effects of

environmental management systems?a study of networking small and medium-sized enterprises using a joint environmental management system. Business Strategy and the Environment, 12(3), 163–174. https://doi.org/10.1002/bse.357

Anokhin, S., & Wincent, J. (2011). Start-up rates and innovation: A cross-country examination. Journal of International Business Studies, 43(1), 41–60. https://doi.org/10.1057/jibs.2011.47

Audretsch, D. B., & Cunningham, J. A. (2018). Entrepreneurial ecosystems: economic, technological, and societal impacts. The Journal of Technology Transfer, 44(2), 313– 325. https://doi.org/10.1007/s10961-018-9690-4

Baldassarre, B., Calabretta, G., Bocken, N. M. P., & Jaskiewicz, T. (2017). Bridging sustainable business model innovation and user-driven innovation: A process for sustainable value proposition design. Journal of Cleaner Production, 147, 175–186. https://doi.org/10.1016/j.jclepro.2017.01.081

Burt, R. S. (2018). Network Disadvantaged Entrepreneurs: Density, Hierarchy, and Success in China and the West. Entrepreneurship Theory and Practice, 43(1), 19–50.

(28)

Cohen, B. (2005). Sustainable valley entrepreneurial ecosystems. Business Strategy and the Environment, 15(1), 1–14. https://doi.org/10.1002/bse.428

Cooney, S. (2009). Build A Green Small Business. Profitable ways to become an ecopreneur. McGraw-Hill Publishing House. Geraadpleegd van

https://www.researchgate.net/publication

Crumpton, M. A. (2012). Innovation and entrepreneurship. The Bottom Line, 25(3), 98–101. https://doi.org/10.1108/08880451211276539

Daly, H. E., & Farley, J. (2003). Ecological Economics: Principles And Applications (1ste editie). Geraadpleegd van

https://www-sciencedirect-com.proxy.uba.uva.nl:2443/science/article/pii/S0921800905003277?via%3Dihub Gilsing, V., Nooteboom, B., Vanhaverbeke, W., Duysters, G., & van den Oord, A. (2008).

Network embeddedness and the exploration of novel technologies: Technological distance, betweenness centrality and density. Research Policy, 37(10), 1717–1731. https://doi.org/10.1016/j.respol.2008.08.010

Henrekson, M., & Sanandaji, T. (2014). Small business activity does not measure

entrepreneurship. Proceedings of the National Academy of Sciences, 111(5), 1760– 1765. https://doi.org/10.1073/pnas.1307204111

Mason, C., & Brown, R. (2011). Creating good public policy to support high-growth firms. Small Business Economics, 40(2), 211–225. https://doi.org/10.1007/s11187-011-9369-9

(29)

Meagher, K., & Rogers, M. (2004). Network density and R&D spillovers. Journal of Economic Behavior & Organization, 53(2), 237–260.

https://doi.org/10.1016/j.jebo.2002.10.004

Milne, M. J., & Gray, R. (2012). W(h)ither Ecology? The Triple Bottom Line, the Global Reporting Initiative, and Corporate Sustainability Reporting. Journal of Business Ethics, 118(1), 13–29. https://doi.org/10.1007/s10551-012-1543-8

Neumeyer, X., & Santos, S. C. (2018). Sustainable business models, venture typologies, and entrepreneurial ecosystems: A social network perspective. Journal of Cleaner

Production, 172, 4565–4579. https://doi.org/10.1016/j.jclepro.2017.08.216

Nidumolu, R., Prahalad, C. K., & Rangaswami, M. R. (2009). Why sustainability is now the key driver of innovation. IEEE Engineering Management Review, 43(2), 85–91. https://doi.org/10.1109/emr.2015.7123233

Nijkamp, P. (2003). Entrepreneurship in a Modern Network Economy. Regional Studies, 37(4), 395–405. https://doi.org/10.1080/0034340032000074424

Osterwalder, A., & Pigneur, Y. (2013). Business Model Generation. Hoboken, NJ, Verenigde Staten: Wiley.

Patzelt, H., & Shepherd, D. A. (2010). Recognizing Opportunities for Sustainable Development. Entrepreneurship Theory and Practice, 35(4), 631–652. https://doi.org/10.1111/j.1540-6520.2010.00386.x

Roundy, P. T., & Bradshaw, M. (2018). The emergence of entrepreneurial ecosystems: A complex adaptive systems approach. Journal of Business Research, 86, 1–10.

(30)

Schaltegger, S. (2016). Business Models for Sustainability. Organization & Environment, 29(3), 264–289. https://doi.org/10.1177/1086026616633272

Stam, E., Romme, A., Roso, M., Van Den Toren, J., & Van Der Starre, B. (2016). Knowledge triangles in the Netherlands : an entrepreneurial ecosystem approach. NARCIS. Geraadpleegd van https://research.tue.nl/nl/publications/knowledge-triangles-in-the-netherlands-an-entrepreneurial-ecosyst

Stam, E., & Spigel, B. (2016). Entrepreneurial Ecosystems. Geraadpleegd van file:///Users/thijsvanderwijk/Downloads/rebo_use_dp_2016_1613.pdf

Stam, Erik. (2014). The Dutch Entrepreneurial Ecosystem. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2473475

Stam, Erik, & van de Ven, A. (2019). Entrepreneurial ecosystem elements. Small Business Economics. https://doi.org/10.1007/s11187-019-00270-6

Tansley, A.G. 1935: The use and abuse of vegetational concepts and

terms. Ecology 16, 284—307. (2007). Progress in Physical Geography: Earth and Environment, 31(5), 517–522. https://doi.org/10.1177/0309133307083297 United Nations. (2020). Transforming our world: The 2030 agenda for sustainable

development. Geraadpleegd van

https://sustainabledevelopment.un.org/content/documents/21252030%20Agenda%20f or%20Sustainable%20Development%20web.pdf

Van De Ven, H. (1993). The development of an infrastructure for entrepreneurship. Journal of Business Venturing, 8(3), 211–230. https://doi.org/10.1016/0883-9026(93)90028-4

(31)

Van Gelderen, M. (2017). Giving and taking in networking. Enterprising Competencies. Geraadpleegd van http://www.enterprisingcompetencies.com/

Witt, P. (2004). Entrepreneurs’ networks and the success of start-ups. Entrepreneurship & Regional Development, 16(5), 391–412.

https://doi.org/10.1080/0898562042000188423

World Economic Forum. (2013). Entrepreneurial Ecosystems Around the Globe and Company Growth Dynamics. World Economic Forum. Geraadpleegd van

The effect of sustainable business models on network density within startup alliances in the entrepreneurial ecosystems of Amsterdam and Eindhoven.

Faculty of Economics and Business