Venture capitalists are experts when it comes to scaling. You might even say that recognising that a startup has the attributes (and technology) necessary to scale is *the* core attribute of a successful venture investor. And you could say with equal conviction that if there's one value-add that VC guys consistently bring to the table, post-funding, it's a highly useful understanding of the tried and tested formula: capital + the right product + the right technology + the right team (+ good timing!) = scale.
So why is it that most VC guys suck when it comes to scaling venture capital businesses? Why do so few scaled-up funds exist, and why does it take twenty years to build them? Why is it that we assume that "ten to twenty companies" is the optimal size for a USD100Mn fund portfolio (versus 500, or 2,000)? Are a small band of local referrers really the best way to build a deal funnel? Is a locked-up 12 year fund really the best way to service a growing base of LPs/investors?
These are the questions we set out to answer as the basis for a strategy for H2, our second fund. Answering them set us on a two year journey... involving dozens of meetings with other VCs, family offices, institutions and accelerators in more than a dozen countries, analysis of data from hundreds of thousands of venture investments from around the world, and the creation of thousands and thousands of lines of code. And here's what we learned:
BIGGER IS BETTER
Out of all of the variables we looked at, the variable most consistently correlated to positive outcomes was the size of the portfolio - not the dollar size, but the number of companies. And while some small portfolios perform extremely well (Union Square Ventures are the rock stars of this space), virtually all of the 75% of VC funds that we found that underperform the NASDAQ consisted of small portfolios. And with respect to the top quartile, the large, diversified VC funds were among the top performers.
Econ grads are reading this and nodding their heads because they know that this is simply modern portfolio theory (MPT) in action - or in layman's terms, diversification. It took us billions of Monte Carlo virtual portfolio simulations to reach the same conclusion. But reach it, we did. Larger portfolios - particularly portfolio sizes consisting of hundreds of different companies - persistently outperform all but a small proportion of small portfolios - sometimes by a factor of 2x to 3x on exit.
Recommendation: Massively scale your portfolio - when it comes to VC, a large number of small bets performs better than a small number of large bets.
EARLIER IS BETTER
We discovered something interesting during our recent trips to China. We discovered that as overheated as later rounds might be in the PRC (relative to the Valley, or pretty-much anywhere else), investors can sometimes buy into earlier rounds at valuations significantly less than the US or European benchmarks. And as we looked more closely at this phenomenon, we discovered that with respect to very-early-stage investing, over-discounting was taking place, leading to attractive valuations, and large multiples, for those willing to take the risk of investing out-of-the-box.
So we ran the numbers, using data from over 300,000 deals*. And discovered that when you look at all the data collectively, the first four rounds are where the returns come from - if you examine post-Series A median returns, you'll find your median plot line starts merging with the PME (public market equivalent) line pretty-quickly... earlier is most certainly better - assuming you are able to scale your platform to find and act on a large enough number of deals (see point one.)
Recommendation: Massively scale your deal sourcing network (i.e. beyond referrals.)
AI IS KEY TO ENABLING SCALE
As I've blogged about previously, there's only a handful of us actively building AI technologies into the venture investment analysis process right now - as an example, YCombinator is using their HAL algos to parse their applications, and Hatcher+ is using DART to parse applications for a wide range of accelerators and angel groups, and also to do continuous performance monitoring of cohort participants (DART stands for Diligence, Accuracy, Responsiveness and Transparency.)
Why is this necessary? Based on our research, we've found that most venture firms and accelerators only accept a mere handful of submissions - typically between 0.5% and 2% - and of these, on average, just 46% of those selected receive follow-on funding. Which means for every company that is funded, roughly 238 business plans need to be sourced, downloaded, read (at 20 pages per deck, that's 4,476 pages), and voted on by the (average) four people on staff charged with doing this.
AI can help this process massively by actively assisting in finding the startups that best fit the investment mandate conditions concerning geography, sector, stage, culture, and diversification. If you're a specialist accelerator with a tight investment mandate, AI technology is a great way to turn a huge, unwieldy deal funnel into a smart deal engine - and make the most out of the human smarts on your team.
Recommendation: Adopt technology, and use it to create a scalable platform.
ETF-STYLE TRADABILITY WILL DEMOCRATIZE AND REVOLUTIONIZE VENTURE INVESTING
VC funds are notoriously long-term in nature. We came across some research by Adams Street recently that showed that more than 50% of VC funds extend their lives beyond fifteen years - and only 7% manage to liquidate in ten years or less.
Unsurprisingly, a lot of people, including many family offices, find the idea of a fifteen-year-plus lockup deeply unattractive. Yet in spite of the revolution taking place within fund regulations, blockchain-based technologies, and exchange-based platforms, these structures persist.
We predict massive changes are ahead. We believe that while fund lifetimes will remain largely the same, the units within them will start trading with greater and greater frequency. LPs will look to exit based on their individual investment mandates (rather than the fund's), and, spurred on by the emergence of ICOs and ETF-style venture fund units, will force venture funds to finally move from inefficient, manual secondary selling processes to more sophisticated exchange-based structures (such as those currently being pioneered by Estonia-based Funderbeam.)
Why is this a good thing? It's a good thing because these kind of structures will enable VC funds to be accessed by investors with shorter investment horizons - including many of the investors we've spoken to over the past years that have avoided venture not because the returns are unattractive, but because the timeframe for liquidation is just too long. Allowing LPs a greater say in determining the timing of liquidity events would do more than any current government initiative to encourage investment in startups - we predict it would generate a flood of interest - and a flood of new capital as well.
Recommendation: Engage a secondaries specialist and/or structure your next fund to enable ETF-like tradability
As anyone who has recently seen our keynote or pitch deck can tell you, these are the four strategies that resulted from our two years of studying our business - and the four objectives that we are most excited about at Hatcher+.
If you agree, we're happy to share. As part of our commitment to reinventing venture capital, we're happy to share our tools, including our co-investment platform, so you can scale your business (or just your deal network.) Just give us a call, or hit the Intercom button in the bottom right of the screen. We don't charge for using these tools - as we've learned from our entrepreneurs (and investors), that's a very reliable way to enable scale. :-)
*Data sources: Hatcher, Red Seer, PitchBook, Venture Intelligence, other.