The Essential Factors for a Succesful Startup Ecosystem; A Quantitative Study Across European Cities

Master thesis research by Thijs Waanders

In June 2016 Thijs Waanders graduated from the Tilburg School of Economics and Management at Tilburg University. He followed the master Strategic Management and wrote his master thesis in the area of startup entrepreneurship.

In recent years, many industries have been disrupted by so called startup companies. These startup companies share a number of characteristics in the sense that they are usually small scale, privately funded and fast growing. Local governments have been trying to attract startup companies in order to stimulate economic growth and create new jobs, yet research on this topic proved to be fragmented. It either focused on case-studies in which two local ecosystems were compared, or was aimed at several well-known ecosystems like Silicon Valley, Berlin or London.

Previous studies identified many relevant factors, ranging from accessible markets, to cultural support. Although the goals of previous research were the same,  identifying crucial factors that make up a startup ecosystem , a lack of meta-analyses resulted in a highly fragmented pool of findings.  This thesis set out to identify the critical factors that contribute to a successful startup ecosystem by comparing many European cities.

The first part of the thesis deals with identifying and standardizing previous research, by means of a comprehensive literature review. The critical factors identified by other researchers often overlapped. Consequently, this study restructured these into a new set of factors, namely; (1) Market conditions, (2) Human capital, (3) Funding, (4) Entrepreneurial support network, (5) Legal framework and (6) Education.

Data was obtained from several sources, including startup databases (Angellist and Dealroom) and government sources (Eurostat and statistical bureaus of local governments). In order to obtain and process these large sets of data, programming techniques were used. Several Python and VBA scripts were developed, which allowed for web scraping and processing of large numbers of raw data. A final sample of 283 European cities, each with a number of variables was obtained and used in the statistical analysis. This final set of six factors described above was then statistically tested by means of several regression models.

Three of the proposed factors, namely Funding, Education and Entrepreneurial support network were found to play a significant role in startup ecosystems. The hypothesized effect of city size on the success of a startup ecosystem (measured as the number of startups with funding), had to be rejected.

One of the primary takeaways is that automated research techniques can and should be used to analyze large datasets. This is not only more efficient, but also allows the researcher to identify and take into account factors that would otherwise be overlooked due to a lack of structured data. An example of this principle that was applied to this thesis, was the analysis of the factor Entrepreneurial support network. This factor was measured based on the number of followers per city on startup minded social networks ( and, which was retrieved using their API’s (application programming interfaces) and web scraping techniques.  Another key takeaway is the power of a large-scale analysis. By taking a large number of smaller (lesser known) cities into account, city size as a moderator could be tested. The hypothesis stating that the size of a city has a significant effect on the success of a startup ecosystem could be rejected, which is an interesting conclusion in itself and allows for further research.