In my previous post On $600B of Y Combinator Startup Success I discussed our analysis of the 271 companies on Y Combinator’s latest Top Companies list, which now total over $600B in value. A recurring theme in our studies of YC startups is just how ‘winners take all’ value creation is in the world of technology startups. Returns in seed-stage venture are power law driven, with just ~5% of startups accounting for the vast majority of investor returns.
While most early-stage venture investors know this, there’s still much debate on how best to go about catching the ‘unicorns’ that really drive returns in our industry. Some investors take a concentrated portfolio approach (8–10 investments per fund) which emphasizes ownership percentage over diversification, hoping that just one big success can return the whole fund. Other investors take a more index or ‘spray and pray’ approach (100+ investments per fund) which emphasizes diversification over ownership percentage, casting a wide enough net to have confidence that several unicorns will be caught.
Since our job at Rebel is to invest in the top 0.1% of the 30,000+ tech startups that apply each year to Y Combinator, the world’s #1 technology startup accelerator, we collect hundreds of thousands of data points on YC startups each year and maintain a sophisticated machine-learning algorithm to help predict tomorrow’s most successful technology unicorns. Being the data nerds that we are, it’s only natural for us to go about answering the optimal portfolio strategy question with some real-world statistics.
So, we estimated the portfolio returns of 146 fellow YC startup investors who have each invested in at least 5 startups on YC’s Top Companies list. We then created a scatter plot chart comparing their total number of YC startup investments over the years (x-axis) with their estimated gross portfolio return multiples¹ (y-axis). Since some of these investors have been investing in YC startups much longer than others, and thus have a more mature startup portfolios, we’ve color-coded each investor according to the average vintage year of their YC portfolio (see legend). We also decided to keep each investor’s identity hidden to protect the confidentiality of our VC peers.
As for the billion-dollar question of which portfolio strategy works best? Our answer is ‘it depends’
It’s clear from the scatter plot that the very best investor returns come from more concentrated portfolios. The top example of this is the investor with the red arrow, which has made just 70 YC investments dating back to 2007, but with a whopping >700x estimated gross portfolio return (they’re a famous top-tier Silicon Valley fund with several mega-hit investments)
However, note that some of the worst-performing investors also had more concentrated portfolio strategies. There are many others with similarly mature YC startup portfolios (see the yellow, green, and orange dots) that performed quite poorly, as illustrated by the cluster in the lower-left corner of the chart.
On the other end of the portfolio strategy spectrum, note the investor with the blue arrow. They’re a well-known Silicon Valley ‘super angel’ fund that pioneered the large Y Combinator-focused portfolio strategy and has a similar 2014 average portfolio vintage age as the red arrow fund. Even though their returns appear poor relative to the red arrow fund, they’re still extremely good in an absolute sense at 73x gross, and still well-above average compared to their peers with similarly aged YC portfolios.
Also note the two magenta dots towards the far lower right of the chart with purple arrows, who represent a new generation of highly-prolific YC startup venture funds. While their returns appear poor so far, their YC portfolios are still very young with an average investment vintage of just 2018, and I expect their dots to gain some altitude in the coming years.
Another way to examine the effectiveness of different portfolio strategies is to plot each investor’s total number of YC investments (x-axis) vs their Top Companies ‘hit rate’ (y-axis). The advantage of this approach is that it treats every top company equally, thus removing the outsized impact that a few massive successes can have on an investor’s return multiple. It also reduces the impact of different YC portfolio maturities across investors, since it takes a startup much less time to become a top company than grow to billions of dollars in value.
Here we can clearly see a regression-to-the-mean effect as investors make more YC investments, as illustrated by the downward sloping curve pattern in the blue circle. But what’s most interesting is that some investors were able to maintain a solid top companies hit rate around ~20% despite making a large number of investments, which effectively reduces the riskiness of their portfolios. I’ve added arrows to highlight the best 5 examples.
The purple, yellow and green arrows are all venture funds founded by YC alumni. The blue arrow is the same Silicon Valley ‘super angel’ fund in the previous chart, and the orange arrow is a YC partner. What all of these investors have in common is a deep understanding of the YC startup ecosystem and strong reputation within it.
These 5 outlier investors together have a 20% average top companies hit rate, an 8% average unicorn hit rate, and an 80x average estimated gross portfolio return multiple¹ while investing in an average of 28 YC startups per year. These are impressive portfolio statistics, especially given the relatively low risk they’re assuming by casting their YC startup bets so widely. (Yes, our portfolio strategy at Rebel is similar)
Perhaps the reason portfolio strategy remains so hotly debated is there is no optimal approach — it really depends on who you are and whether you can confidently land in the top-tier of early-stage venture investors. If you’re confident you can, then a more concentrated portfolio makes sense. However, if you’re new to the game or simply more risk-adverse, then the data suggests you should cast a wide net. This is especially true if you invest primarily in YC startups and have certain selection, access, and/or reputational advantages within this highly curated startup ecosystem.
¹Assumes an average $10M entry valuation, assigns a zero return value to non-Top Companies, and is gross of dilution in later rounds