The Data Driven VC
The VC model needs an update — We are about to update ourselves.
Many models have been said to bring innovation/disruption to the VC industry in the past decade. Venture Builders, Operating VCs, Crowdfunding and now ICOs are some examples of it.
And yet many of the top VC firms have been using the same approach to decision making for decades and claim that Venture Capital is a boutique business, a product crafted by artisans trained to leverage instinct, relationships and emotions.
The data shows that this is changing. We have already seen early indicators that AI systems can be surprisingly successful at picking VC investments. Increased availability of data at almost no cost in combination with breakthrough developments in the field of AI create a fertile ground for data driven strategies.
On top of that, as some analysis of VC returns show, randomness and geometric growth play a relevant role in the distribution of outcomes, which creates an exceptional opportunity for improvement using a data driven approach, particularly, for early stage VCs. As an example of that,
if an early stage VC manages to increase the number of home runs by 200–300 basis points, it will double its returns.
By the way, statistical models have already become a necessary tool to make a wide range of decisions in VC, from building the fund model to managing reserves for follow-on investments using Monte Carlo simulations.
We are hiring Data Scientists — Here is Why.
At Nauta we think that there is a wave of data driven innovation coming that has the potential to shake the VC business. Firms like Index or Google Ventures added data scientists to their teams. Others like Correlation Ventures or EQT have started from scratch with a data driven approach.
This is the technology stack that a data driven VC could eventually build:
- The predictive engine: A system that can assess the probability of a potential investment being successful using predictive models (note that “success” needs to be quantified, which is not obvious in VC). For that, the system has to be able to identify causality (not only correlations) between successful companies and attributes or factors that can be spotted at the time of investing, the latter being a condition sine qua non for this machine to be of any use. This is a big challenge because of the long feedback cycles in VC — It takes 3 to 5 years to find out if an investment is a success.
- The dealflow engine: A system that will automatically gather, enrich, analyze and prioritize a vast amount of data of potential investment opportunities from a number of sources. Using a well-defined set of criteria that needs to be very carefully chosen, the engine will come up with a short list of potential investment opportunities to start tracking or to reach out to — This is probably the low hanging fruit because the hurdle is low and the ROI is quasi-immediate.
- The dynamic reserves planner: Using an evolution of the predictive engine, Monte Carlo simulations and basic VC rules, the engine should be able to calculate the optimal distribution of reserves for follow-on investments and plan capital calls accordingly. You can think of it as an inventory optimization platform as used in the agile retail but applied to VC.
We will soon open a Data Scientist position at Nauta in Bcn. Please, get in touch if you would like to know more.
— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
About myself: I’m VC @ Nauta Capital. I’d love to blog more often but I only do it when I’m not doing DD or helping our portfolio ;-) email: email@example.com