The Golden (Selection) Ratio
When we decided we wanted to determine what selection ratio we should adopt - i.e. how many deals we should fund out of every thousand in order to create a 1,350 company portfolio for our H2 fund, it was obvious where to look. There are several VCs that have been at this for decades, and our accelerator partners, some of which have parsed tens of thousands of applications (and interviewed literally hundreds of founders on their way to constructing high-quality cohorts) were great targets for this question.
So we asked them for the numbers. And we included a few crowdfunding platforms for good measure. Here's what we found. We found that most of the generalist VCs we spoke with, as a rough rule of thumb, take about 10 out of every thousand deals analysed to their IC (Investment Committee) - and of these deals, about half, or 0.5% of the deals in the initial funnel, get funding and make it into the portfolio*.
This isn't true for everyone - the more highly specialised or sector-focused portfolio managers tend to accept more deals from their referral networks, simply because the size of the pool is less (imagine yourself as a specialist medical device or hardware accelerator program manager, and you'll quickly be able to visualise why it is necessary for them to push to 3% of deal flow, or even 5%.)
Interestingly, there was little to no geographic bias among the best firms - we found the same ratio in India, China, and the US. Most operators were using, unconsciously or not, a selection ratio of roughly 1%.
The anomalies? Family offices doing direct investment into tech - not all, but certainly many were over-selecting. And some were over-selecting to a dangerous level - we've spoken to some family offices that appeared to have selection ratios approaching, or even exceeding, 20% - or one in five deals. There is, of course, a chance that the referral network being used is so good that all the filtering has been done prior to the selection process... and we've certainly observed that. But we've also observed the opposite - a too-large number of direct investments being made on the basis of poor-quality deal flow.
Luckily, that problem is not only solvable, it gives us fund managers a reason to exist - and is probably the core reason why funds (and technologies) such as ours find investors. 2 and 20 is a small price for an exceptionally well-curated, high-quality deal flow we think.
Why is the selection ratio important? We thinks if you're building a large portfolio, or an index fund, it's vital to maintaining a level of quality that is consistent with industry standards.
We also think it's important for reasons downstream of the selection process - to use one obvious example of another "one in one hundred" ratio, according to Hatcher+ research (and CB Insights, and others), large portfolios produce unicorns, or companies with a valuation that grows to in excess of one billion dollars, roughly one in every one hundred externally-funded startups, which correlates to 1 in 20,000 business plans at the tightest generalist VCs, and one in every 2,000 for the most specialised accelerators.
*It has to be said that sometimes this isn't because the deals didn't get the thumbs up, it might simply be because the IC was meeting one week too late relative to another investor looking at the deal - such is the way sometimes with the more traditional, once-a-month IC meetings.
The Best Investor Deck Ever|
The Most Unrewarding, Misunderstood, Underrated Job on the Planet|
Venture Capital Simplified - The Rule of 5X|
The AI-Powered, Highly-Automated, Global-Diversified, Exchange-Tradable VC|
AI and Machine Learning: The Poincare Five Step Process|
|The Power of Venture|
|US Immigrants Are Vital To VC Returns|
|Hatcher+ v2.0 Release|
|The Golden (Selection) Ratio|
|The Time-Traveling VC|
|We Created a Venture ETN - Here's Why|
|The Crane's Neck - Using Data To Improve Startup Portfolio Construction|
|Leveraging industry expertise – best practices from the accelerator universe|
|8 Easy Places To Start Up|
|Your Startup Should be in an Accelerator|