Startup success can be predicted by employees’ connections, study finds
UK researchers construct massive, interconnected global business network, find that young companies near its center in early stages more likely to succeed financially
Luke Tress is an editor and a reporter in New York for The Times of Israel.
The success of startup companies is heavily influenced by their existing professional connections, a study has found.
Researchers in the UK constructed a massive visual network showing connections between companies and their employees, and found that the closer a company was to the center of such a network, the greater financial success it tended to achieve.
The study used the model to develop an algorithm to predict companies’ success that was two to three times more accurate than existing methods used by venture capital firms.
The study, by Queen Mary University of London, was published in the online journal Scientific Reports last week.
Traditional business evaluation models are difficult to apply to startups because the young companies lack historical data, the authors note. Venture capitalists and private investors tend to base their evaluations of startups on the quality of the people involved and the potential of the markets they’re entering. The evaluation process is subjective, labor-intensive and largely ineffective.
“Traditionally historic reports on sales, growth or market size are used to predict future success but with startups this level of data usually isn’t available,” Dr. Lucas Lacasa of Queen Mary University said in a statement. “Instead measures such as the qualifications and attributes of founding entrepreneurs are used, which can be subjective as well as labour-intensive.”
The study proposed a data-based framework for evaluating the potential of young companies. The researchers used a worldwide network of professional relationships among startups in what they said was a novel approach to assessing the companies.
The massive network “provides the backbone and the channels through which knowledge can be gained, transferred, shared, and recombined,” the study said.
Employees and advisers who move between companies bring experience, expertise and different technologies, and investors, lenders and board members accumulate and apply their business acumen and funding.
The study drew data from the Crunchbase business information website, which compiles global data on companies’ leadership, investments, founding events, mergers, acquisitions, trends and other news.
The network comprised 41,830 companies in 117 countries, and 135,099 links between the firms.
Financial success was measured by funding, acquisitions and initial public offerings.
The research looked at 26 years of data, spanning from 1990 to 2015, on companies and individuals. They connected people to companies according to their professional roles in the firms, and then examined how individuals were connected to multiple companies.
Individuals often change companies for better opportunities, creating a flow of information and know-how, benefiting their new employers and likely moving the receiving company into a better position in the larger business network. The individuals’ former companies were not likely to be significantly harmed by the move, however, the authors said.
In one example cited by the study, Airbnb hired an individual who had previously worked at Google, Twitter and Mozilla. This movement likely increased the flow of knowledge between the companies, and moved them closer together within the larger network.
A company was considered more central to the network if it was linked more closely to other companies. Apple was always among the top 10 companies in the network, while younger companies including Facebook, Airbnb and Uber all moved quickly into central positions after being founded, possibly due to an influx of venture capital into startups in recent years.
Startups that were more central in the network at an early stage were more likely to succeed long-term by acquiring other firms, getting acquired, or going public, the authors said.
“We propose that this novel, data-driven approach could complement existing screening approaches used by investors and we anticipate that further refinements could improve the prediction accuracy even more,” Lacasa said.
The model was less accurate during periods of economic crises, such as during the market crashes following the dot-com bubble in 1999-2001 and the US housing bubble. The reasons for this were unclear, the authors said.