This is what riveting superhero drama is made of: Company uses futuristic predictive technology to tap into the latent powers of an unsuspecting lot. Said company unites a select few to help solve problems. Company realizes great returns from this exceptional team.
In the case of Bloomberg Beta, the venture investment arm of Bloomberg LP, truth is not far from fiction. On the heels of its 2013 launch, Bloomberg Beta joined forces with Mattermark to use a powerful predictive deal intelligence framework in identifying startup founders even before these professionals could realize they may be founders. And the firm’s impressive portfolio is already paying off with successful investments in popular startups, such as Newsle and Codeacademy.
The joint initiative posed a multi-billion dollar question: Is it possible to apply big data and machine learning to volumes of public data in an effort to identify individuals who have not started a venture-backed company yet are most likely to launch what may become the next WhatsApp or Twitter?
The team devised a study of 1.5 million professionals in the technology ecosystem. The study analyzed leading indicators in this diverse population, taking into account a host of predictive factors, such as education, age, employment history, career roles and seniority, and geographic location. Based on study results, Bloomberg Beta invited 350 of the highest ranking performers to a private program designed to connect and encourage these individuals to explore opportunities to start a company.
Did each one of the 350 unite for the greater startup good? No. Mattermark had predicted that program invitees had a 17% chance of starting a company, so the likelihood that scores of new companies would result from the initiative was low. In any case, the program intent was never that explicit but, in some measure, perhaps simply seeding the idea in the minds of individuals with a proclivity for startup life would be sufficient encouragement.
So what common characteristics did the founder pool share? The results surprisingly ran counter to the media-driven stereotype of nerdy wunderkind domineering the startup world. The study revealed that nearly 40% of venture-backed founders are over 40 years old and only 15% have a Computer Science degree. Management consultants are more than twice as likely to be venture-backed founders than engineers. Slightly more than 40% of venture-backed founders had worked at a venture-backed company immediately before founding their own company. And, contrary to the conventional wisdom that founders previously have been leaders in a company, nearly 70% of venture-backed founders were not in a senior leadership position prior to founding their startup.
Predictive modeling indeed is transforming the business landscape, as companies increasingly incorporate its techniques in their analyses. But determining which movies consumers prefer or the electronics that will be popular next year is one thing. In the case of Bloomberg Beta, using career indicators to bring seemingly venture-minded professionals together with no plan in place and with only the hope of reaching a future venture outcome seems uncomfortably precarious in a world that is already fraught with financial risk.
Still the notion that an algorithmic pattern can predict and help unite future business leaders is fascinating. Initial results seem to corroborate Bloomberg Beta's investment strategy. From Newsle to ThinkUp to REDEF, the venture firm's startup universe is growing to fight business problems.