Why your startup needs to impress an algorithm
An article in Axios, “Scoop: Inside Google’s Venture Capital “Machine’”, discusses the investment decision-making process at GV, formerly known as Google Ventures, Alphabet’s $2 billion venture capital investment arm, providing details about how the “machine”, an algorithm fed with market data, investment databases, information on founders, investors, financing rounds, progress, etc., and that then comes up with a red, green, or an amber traffic light decision that tends to be more negative than positive.
Axios says the algorithm was initially developed as a decision support system, but evolved into a de facto automated investment committee. Bill Maris, founder and the company’s first CEO, is quoted:
“We have access to the world’s largest data sets you can imagine, our cloud computing infrastructure is the biggest ever. It would be foolish to just go out and make gut investments.”
The algorithm, in fact, is already a veteran: The New York Times mentioned it in 2011, as did Business Insider in 2015 during the first year of the company’s investments in Europe. The idea of a company with a $2 billion investment arm making decisions based on the output of an automated algorithm reminds me of PreSeries, the algorithm created by BigML (I’m a strategic advisor to the company) and Spain’s Telefónica, which I wrote about in August 2015. The algorithm was created to decide on the winner of the Telefonica Open Future startup competition, and has already been used several times, even leading to a spin-off, an automated platform to discover, evaluate and monitor the early stages of investments. In the case of GV, however, everything indicates that whether it has used the algorithm or not, the company has been pulling out of projects in the seed capital phase in favor of others in more consolidated companies.
Axios says that GV is not the only investment company using algorithms, it mentions Sweden’s EQT Ventures’ Motherboard, along with Social Capital’s Capital as a Service. The concept is gaining traction, meaning an increasing number of startup projects, often the result of months or years of hard work and sacrifice, being analyzed in moments by an algorithm, the result of which will possibly rule it out of further investment rounds. In other words, startups should no longer be concerned about impressing a series of investors, but instead an algorithm, in the same way that taxpayers and anybody seeking a bank loan must. Over time, an algorithm can be refined to improve its results by being fed more data, and more decisions with verifiable results, effectively automating the work of investors and making it less intuitive and more scientific. I wonder if all this means that we’ll see fewer black swans in years to come…
(En español, aquí)
