Smart VCs all have a “Portfolio Construction Model”. It’s basically fancy math that lays out if I raise a fund that’s X mil dollars, I think I should invest in Y seed cos, Z series A cos, do this many pro-ratas, etc to generate some expected return.- Brian Ma
It took me 15+ years to finally understand a VC’s “portfolio construction” and how it affects their decision to invest. Founders, here’s what you need to know. Warning: this is 400+ grad level stuff. Scroll to the end if you just want the TLDR.
1/ Background: I’ve founded 3 venture funded companies and raised many many rounds of capital. I’ve always struggled to understand how pro-rata decisions are made or how fund cycles affect a VC’s process and speed. Here’s the secret.
2/ Smart VC’s all have a “Portfolio Construction Model”. It’s basically fancy math that lays out if I raise a fund that’s X mil dollars, I think I should invest in Y seed cos, Z series A cos, do this many pro-ratas etc to generate the some expected return.
3/ After deciding num of co’s and stage to invest in, fancy math has many key inputs to optimize returns, but the main ones are: – Round size and graduation rates per stage – How much to follow-on on each stage – Whether to recycle capital when co’s exit
4/ The output of the model is something called “TVPI” or total value paid-in, which for simplicity is how much you return to investors from the money they give you. You can generally think of 3x as you’ve approximately tripled an LPs capital.
5/ Lets do examples. If I raise $10M to invest in seed stage companies. $150k each, I’ll invest in ~50 companies. Assume some graduation rate (50% co’s get to next stage, at $8M post, etc), the model expects a return of 7.11x TVPI. A “good” fund returns 3-5x.
5a/ It turns out the most sensitive part of the model is what your assumptions are about pre-seed, seed, A etc grad rates, valuations, exit rates, etc. This is why VC’s pay lots to keep an eye on benchmarks on all these rounds in different geos, sectors, etc.
6/ The next big decision is how to much to allocate to follow-ons. These are called “reserves”. In my example model. If I decide to exercise my pro-rata to keep my ownership on every round, my TVPI drops to 4.43x. Ouch.
6a/ If I exercise only 20% of the time (keep picking the best cos), then TVPI is 6.12x. What if I pick the best co’s, but only in the subsequent 2 rounds, not all the way to IPO? 6.8x TVPI. You can see how this gets complicated fast.
7/ While reserves are interesting, picking ability actually matters more. So obvious in hindsight, but VCs need to pick good companies. In our example, if 75% of companies get follow-ons, you get the 7.11 TVPI. If 50% suddenly it drops to 3.71x.
8/ The last ‘basic’ idea then is recycling. If a company does really well (or really badly) and gives you back money within the first X years, do you put that back into more companies? Turns out recycling is generally very good.
8a/ In example above, if you recycle in the first 3 years up to 20% of the fund, you get 7.24x TVPI. Recycle for 5 years, you get 10.7x TVPI. This again assumes you keep picking the same quality companies.
9/ The biggest ‘ah-ha’ point here is Smart VC’s will compare their initial models to how their portfolios are doing right now. They’ll do this quarterly, and it’s THIS that affects their ‘mood’ in how they invest.
9a/ ie. If I’m generating higher TVPI then I modeled because I had a great early exit, I might be subtly more lax in investing, maybe less aggressive, more focused on my next fund. If I’m behind, I’ll be more aggressive.
9b/ Timing also matters a lot. If I’m early in my fund, I need to get companies under my belt, if I’m in the middle or late in the fund, I’m probably majority deployed.
9c/ The biggest “it depends” factor is actually the pro-rata reserve portion. It’ll probably be hard to get a VC’s pro-rata strategy from them without a very very good relationship, but you can at least ask how much reserves they have, what % deployed.
10/ So what should you really care about? As a founder, I really think the answer should be not much. Focus on a great product and making customers happy, and you’ll get funded by the best VC’s.
11/ But if you want to show off your mostly not very useful knowledge of portfolio construction to VCs, ask: – “How many companies do you do every year at our stage” – “How much have you reserved for follow ons” – “How far are you into this fund”
12/ That’ll give you some indication of how aggressive they might be at making a term sheet and how much you can depend on them for future rounds.
Bonus: If you’re really confident, ask them “how’s this fund doing? What’s the TVPI?” 🙂 Then DM me on twitter. Happy fundraising!
Member discussion: