VC Type II, Part II

People have come up with reasonable answers to my previously posted “Type II” problem with respect to VCs. In essence, I asked whether, given defensible assumptions, you might expect a skilled VC to, purely as a statistical fluke, post four straight failing investments. The answer, of course, is Yes.

The way I approach this is to imagine that for a skilled VC, 2-3 of 10 venture investments are outright flops, while 1-2 are big hits. The rest of the investments are small exits and walking wounded, not the sort of thing that creates a career. With these likelihoods in mind, there is small but perfectly plausible chance that your favorite VC could have a run of small-sample bad luck.

Now, there is a confound here, which is the amount of time involved. You don’t walk onto the playing field with ten completed investments. Instead, you make investments serially, with, on the high side, perhaps four having been completed by your third year as a typical venture investor.

And that’s one part of the problem. Not only is it conceivable that a skilled venture investor could come out of the gate with four straight flops, it is even more likely that there is a combination of flops and no data, which hangs the VC in information limbo, making it awfully tough to raise the next fund.

What does all this mean?

  • If you fund a new venture fund once, it is highly likely you will have to fund it a second time if you don’t push the GPs to invest quickly. You want data within four years, and you want every partner to have done at least 3-4 deals.
  • There is tension about investing speed and numbers in first-time funds. You want your GPs to get out of the gate quick enough to get you some results, but you don’t want that to compromise quality.
  • There is tension on fund size in first-time funds. While LPs understandably don’t want to put a lot of capital at risk in first-time funds, the fund has to have enough money for the GPs to run out and make enough investments for some species of quasi-statistical defensibility. Too small a portfolio — across the fund, and on a per-partner basis — and you have no idea whether it’s simply a small-number artifact, either way.

These musings were mostly precipitated by a recent conversation with a major LP agonizing about whether to re-up with a first-time fund — and whether to let go one of the partners who seemed to be underperforming.