The Tyranny of Venture Capital Math
Venture capital is a simple business but it’s not an easy business. When a new analyst or associate joins us at AgFunder, it usually takes 9-18 months to calibrate to this reality. The most frequent mistake I see founders and new VCs make is that they focus too much on the idea. They focus on the utility of the product or service without thinking about how big it could. For VC funds, outcomes generally need to be really big, especially when you factor in the high failure rate of early-stage startups. This puts ultimately puts investments into two buckets: those that are “venture backable” and those that are “not venture backable”.
To help founders and new investors get their bearings and think more critically as they evaluate opportunities, I created a simple fund model to fill it out with a defensible set of assumptions. I’ve added the exercise below.
Exercise: Build a Fund Distribution Model
In the [Fund Model] tab, build a financially defensible model to generate VC-level returns.
For the top two performing outcomes, please sanity check those investments by building up a defensible series of investments in those companies in tabs [Company 1] and [Company 2].
SPOILER: READ AFTER THE EXCERCISE
Post-Excercise Discussion
In the preamble, I purposefully left out a lot of guidance because I want to see how the person thinks natively about defensible returns without leading them too much. The first thing that a person should do is to understand what return expectations are for venture capital funds. A 2x fund is not going to be interesting to LPs and unless you’re Sequoia or First Round, no one’s going to believe you can generate a 4x or greater fund so the general target is to model a fund that can return 3x.
The next thing a person should do is is to go and find base rates. Chris Dixon and Alex Graham have compiled some nice data that capture loss-ratios and outcome multiples. In the fund model above, the 80x return option is really a red herring because it’s generally not defensible to assume one of your investments will generate an 80x return, nor is it defensible to assume that your loss ratio will be much lower than the industry average. In Alex Graham’s article above, he shows graphs from Correlation Ventures that only 0.4% of all investments generate a 50x or greater return but even 10x or greater returns are rare (4%-6%).
For my own-back of the napkin thinking, and because we often go early, I generally assume that 65% of the time any investment I make will be a write-off (0x-1x) which means I need to be looking at very large outcomes to make up for the inevitable losses. In a simplified model, if I have a $50m fund with $40m of deployable capital after fees, and 2/3rd of that capital will effectively go to zero, then it means that I need to generate $150m (3x) from $13m of investment. Here, my average winner needs to be ~10x. In a portfolio of 10 companies, with an average investment of $5m, I need three companies to each generate a $50m outcome. In other words, I need to be aiming for each investment to return the fund.
And this gets to the heart of what I call the tyranny of venture capital math: not everything that can be invested in is investible and this may be the hardest part about venture capital. Every day I meet eager founders developing technologies that can make the world a better place, but the tyranny of the math is a forcing function on venture capital that greatly limits the types of investments that we can make. In effect, every investment has to have the potential to become a category-defining company.