Does your VC have an investment thesis or a hypothesis?

Venture capitalists love to talk investment theses: on Twitter, Medium, Clubhouse, at conferences. And yet, when you take a closer look, theses are often meaningless and/or misleading.

OpenVC is a new, open-source initiative to collect and analyze all publicly available VC theses to help founders more efficiently find the right investors — and vice-versa. For the first time, we are sharing here our initial conclusions. We hope you’ll upload your own thesis to benchmark yourself. We’ve identified six common patterns of how VCs articulate their theses and some best practices in doing so.

Our analysis is based on two complementary datasets:

• 125 theses so far submitted by investors into the OpenVC database.
• 36 theses pulled directly from U.S. VC websites by David Teten and Sam Sabin, co-founder of Hireblue.

Our four primary conclusions:

1. Public theses are often inconsistent with how firms actually deploy capital.
2. VC theses are often so vague that they’re meaningless.
3. We found seven categories of VC theses, plus an eighth: the non-thesis.
4. Investment theses are just hypotheses; the portfolio shows how accurate the hypothesis was.

For the sake of simplicity, we will consider “investment thesis” and “investment criteria” as equivalent terms moving forward, although we argue that the thesis leads to the investment criteria. We summarize how they interrelate in the table below.

1. Public theses are often inconsistent with how firms actually deploy capital

A typical VC thesis: “We invest in tech startups in Europe at an early stage.” However, our experience shows that in many cases “Europe” means a handful of countries, for instance, France, U.K. and Germany; and “tech” means B2B SaaS/fintech or consumer apps.

Thirty-four VC firms in OpenVC call themselves “early stage.” Yet 30% of those don’t actually invest in pre-revenue startups. The phrase is quite ambiguous; we suggest quantifying check size so that your investment preference is clearer.

Almost every VC says that they invest in the “best” founders. However, according to PitchBook Data, since the beginning of 2016, companies with women founders have received only 4.4% of venture capital deals. Those companies have garnered only about 2% of all capital invested. This is despite the fact that the data show you’re better off investing in women.

This lack of transparency results in confused founders who chase the wrong investors. In turn, investors are overwhelmed with poorly qualified opportunities.

2. VC theses are often so vague that they’re meaningless

Christoph Janz from Point Nine Capital wrote on Twitter:

The modal VC thesis is: “We invest in great teams addressing large markets with disruptive solutions.” Who invests in lousy teams addressing tiny markets with outdated solutions? Theses also tend to use the same words across many firms, e.g., “daring” and “bold.”

In particular, in our second dataset, we found a disproportionate number of theses focused on “technical” companies (vaguely defined) and focused on companies attacking “problems of the future rather than the present,” in various permutations of that language.

Why are the investment criteria so imprecise on the VC websites? We have three theories, in descending order of importance:

• Option value. Investors don’t want to be too restrictive and miss out on a deal. However, we’d argue that for most smaller managers who are not brand names, it’s better to be highly identified in your niche than being a generalist. Most limited partners we speak with agree.
• A desire to look “sexy” and politically correct as opposed to being honest. This is probably a major reason. For example, saying publicly, “We invest mostly in white/Asian men who went to Stanford like us” accurately describes numerous VCs, but doesn’t sound very politically correct.
• VCs are afraid to give out their secret sauce. We think this doesn’t make much sense; you can share your criteria without telling the whole logic behind them. Many top-tier VCs share detailed public theses.

3. We found seven categories of VC theses, plus an eighth: the non-thesis