I like the thinking behind Yael Hochberg, Alexander Ljungqvist, and Yang Lu’s research on predicting venture investment return. Their study, “Whom You Know Matters: Venture Capital Networks and Investment Performance” shows how, historically, Angels and VCs that make a habit of co-investing outperform funds that tend to participate in more limited syndicates. As TechCrunch explains:
A venture firm’s network in the study was defined as being made up of all the other venture firms who co-invested with it in funding rounds. The more co-investors a venture firm has, the better its network. The better its network, the better its overall returns. The correlation between the size of a venture firm’s network and its returns may have something to do with better access to deal flow, talent, advisers, potential customers, and potential exits.
If this is true, then who are the most connected venture firms and angel investors today? Vijay Dondeti , a graduate student in bioinformatics, applied the analysis in the Hochberg paper to about 2,700 investors in CrunchBase who participated in over 3,300 startup funding rounds between 2006 and 2008. He scored each investor based on how well connected they are to other investors as well as how well-connected their co-investors are to other investors. “In summary,” says Dondeti, “to get a high score, you need to co-invest often with others that also co-invest often.”
But I think the “Whom You Know” study is missing an important correlation. Networked investors tend invest earlier, often at the seed stage, and seed investments, on average outperform the VC investment class. Small funds and Angels need to co-invest, because they don’t have huge funds. Without a huge fund, an investor runs the substantial risk of being unable to meet the needs of the portfolio company. This is why small investors ban together to make sure they can invest a meaningful amount, while still keeping funds in reserve. Small funds make their money on the upside; large funds often make substantial money on their management fees. The result is not a big surprise; small funds tend to obsess on helping their portfolio. Don’t get me wrong, I respect several large funds. There are several great large funds that have a proven track record both participating at all stages and also adding value beyond simply writing a check. DFJ and Sequoia are good examples, and are ranked high on Vijay Dondeti's ranking.
So I salute Yael Hochberg, Alexander Ljungqvist, and Yang Lu. I hope there is a next version of their study where they try to statistically separate Alpha from “stage” from Alpha from “network” (a hard task given the opaqueness of venture returns). As for Vijay Dondeti, I salute you also…your analysis was excellent. I don’t claim your ranking of me (at 63) above Cisco (at 65) is perfect, but if someone knows a better statistical predictor of venture/angel performance… please speak up.
BTW, when reviewing the "Whom You Know Matters" study in the Journal of Finance I noticed an outcome tree for 16,315 venture investments from 1980 to 1999, following their outcome through 2003. For my own edification, I created a pie chart showing outcomes.
Only 18.5% of all deals resulted in an exit over the course of the
dataset. So the takeaway for Angels: if you can't see at least a 10X+
return on investment, don’t make the investment, unless you have some
method of avoiding the other 81.5%. If you have a method of avoiding
exitless investments… please speak up.
Perhaps the folks at Sharespost will help create exit opportunities for early shareholders? Til then I am sticking to co-investing with a 10X+ hurdle.