Breaking the Zuckerberg Myth: Successful Entrepreneurs Have 10 Years of Prior Employment

Abstract

With the advancement of Big Data technologies, and the creation of new, open source Machine Learning and Data Science tools, we now have new methods and approaches to obtain quantitative entrepreneurship research results. This approach benefits from a seemingly inexhaustible source of information, i.e. data. In this paper, we apply data science techniques from UC Berkeley's Data-X framework to research socio-economic characteristics and traits, related to successful entrepreneurs, and their link to early stage venture success. Such techniques include the use of web scraping scripts, data cleaning, the creation of data pipelines, and visualization using open source Python libraries (e.g. Scikit-learn, Pandas) as well as other tools commonly used in the Machine Learning development stack. This paper offers two types of results of the performed quantitative study, that is in the intersection between Machine Learning and Entrepreneurship Research (Fig 1.). First that the Machine Learning development stack offers powerful tools to produce entrepreneurship research results, and this might be a new paradigm in this field of research. Second, we can now quantitatively verify from public data sources that, contrary to popular belief, a successful entrepreneur is not a young, college drop-out that had a genius idea in their dorm room. Instead, the most predictive trait of entrepreneurial success is the work experience of the founder, i.e., the number of years the founder has been employed before starting their company. Optimally they will have been employed for 10 to 12 years.