Venture Capitalists invest in tech, but how much tech do they actually USE in running their own business? From sourcing investment opportunities to choosing which ones will take off, I took a look at the latest trends in AI and ML applied on Venture.
AI and Machine Learning
There’s been a lot of chatter recently about the use of machine learning and AI for sourcing and vetting venture deals. Some of the top examples of this include Sweeden’s EQT Motherbrain, which uses big data to uncover startups. As noted on FastCompany’s article in June 2018:
Motherbrain monitors several million companies using financial data such as funding, web ranking and app ranking data, social network activity, and much more. EQT Ventures continuously adds data about its own assessments of companies in order to train Motherbrain to focus on the right opportunities.
Andreas Thorstensson, partner at EQT told the FT in an interview that 30% of its investment decisions now come through Motherbrain, tho he added
“AI is good for filtering out the noise, but the decision to invest or not will always be about instinct at the end.”
There’s of course GV, where Dan Primack recently covered in “Inside Google’s Venture Capital “Machine”
The first hints of this came in 2013, when then-GV CEO Bill Maris told the NY Times: “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.”
(disclosure:as a former general partner at Google Ventures I’m not able to comment on this).
Other efforts are looking to completely take the human element out of the loop (vs. making recommendations or scores), such as Social Capital’s “Capital as a Service”, a somewhat controversial experiment. Social Capital planned to invest $50,000-$250,000 into 1,000 startups by the end of 2018, but those plans might have changed amid several senior personnel departures. As noted in the Bloomberg article from May 2018 “Impress the Algorithm. Get $250,000“:
There’s an economic logic behind CaaS. Evaluating and backing startups is a labor-intensive process, significantly limiting the number of deals a firm can do. By using software to assess tens of thousands of companies annually, a firm can do each individual deal more cheaply. This makes modest successes worth its time, getting the firm around the need to bet only on companies that could be worth billions. It also eliminates the personal blind spots of its partners.
Another notable effort in play is Correlation Ventures in the US, though not much has been shared on the results so far. The fund leverages predictive analytics for decisions. Correlation currently manages $360 million, and has made 185 investments over the past five years. The fund claims to make co-investment decisions in two weeks and doesn’t take board seats.
Veronica Wu is the Managing Partner at Hone Capital, the VC arm of Chinese PE firm CSC Group. In an interview to McKinsey Quarterly, she described how the firm uses big data to make investment decisions:
We created a machine-learning model from a database of more than 30,000 deals from the last decade that draws from many sources, including Crunchbase, Mattermark, and PitchBook Data. For each deal in our historical database, we looked at whether a team made it to a series-A round, and explored 400 characteristics for each deal. From this analysis, we’ve identified 20 characteristics for seed deals as most predictive of future success.
Based on the data, our model generates an investment recommendation for each deal we review, considering factors such as investors’ historical conversion rates, total money raised, the founding team’s background, and the syndicate lead’s area of expertise.
SignalFire and Inreach ventures are two other examples. This trend also extends to PE funds, according to Pitchbook.
Why don’t we see more funds apply these tools?
The reality is that many funds do, but simply don’t talk about it. It’s safe to assume that KPCB, Sequoia, Accel and others have built their own tools. The downside of this approach is cost and accuracy – it’s expensive to build these recommendation engines (SignaFire said $10 million and InReach mentioned £1 million annually) and employ a large tech team (programmers, data scientists etc). To improve accuracy, funds require large and accurate sets of data that is not so easy to come by.
In addition to AI/ML for the purpose of investments, venture funds are also businesses – they have some specific tasks that can be improved leveraging tech, as well as more general tasks that are common to many SMBs. In the next post I will cover the more traditional tech stack of venture funds, as well as some newcomers to the space.
I'm a former general partner at google ventures, head of Google for Entrepreneurs in Europe and founding head of Campus London, Google's first physical hub for startups.
I'm also the founder of Techbikers, a non-profit bringing together the startup ecosystem on cycling challenges in support of Room to Read. Since inception in 2012 we've built 11 schools and 50 libraries in the developing world.
