Venture Haves and Have Nots


Various commenters have recently observed that there appears to be a small number of very hot startups that can easily raise large amounts of capital at very high valuations, and a larger number of not as hot startups that face more difficulty in raising money. Eric Little and Herb Fockler at Wilson Sonsini were kind enough to share some population statistics drawn from their financings database, in response to discussions from my previous post (“VC Bubble” a Reflection of Public Markets). More specifically, we wanted to see if that data supported the notion that venture financings have dichotomized into a world of Haves and Have Nots. If the data showed average valuations were climbing faster than median valuations, that would be consistent with an increasing dichotomization.

Positive Skew Distribution

Since the lowest valuation in a venture financing can’t go below zero and the highest valuation is unbounded, we expect the average to be higher than the median. Below is a graphical reminder of what a positive skew distribution looks like.

Generic positive skew distribution example.

The ratio of the average to the median value gives us a measure of how far out those outliers on the right hand side are. The higher this ratio, the more pronounced the Haves vs Have Nots separation is. And, if the Haves vs Have Nots issue was becoming more pronounced over time, we would expect this ratio of average valuation to median valuation to go up over time. Below is a plot of those ratios drawn from the WSGR financings universe. Each point in this plot represents the average pre-money valuation of the financings WSGR was involved in for the given quarter and the given financing sequence, divided by the median pre-money for the same group.

Ratio of Average to Median Pre-Money Valuations in WSGR Financings Universe. Higher ratio indicates more pronounced positive skew — a few high value outliers.

As expected, we see the ratios are generally well above 1.0. The average ratio across all 80 quarterly data points is about 1.9, which indicates a very pronounced positive skew to the population. This says that venture financings generally have a small number of hot companies commanding very high valuations and a larger number of less hot companies getting significantly lower valuations.

Looking at the data over time, it’s hard to draw any strong conclusions. We humans are given to seeing trends where there is just statistical noise. The apparent trend of rising Series B dichotomization over the last few quarters may be evidence of an increasing dichotomization between the Haves and the Have Nots, or it may just be statistical noise. The fact that the Series C+ financings don’t show this behavior makes it harder to believe there is a truly causal increasing Haves vs Have Nots effect.

Finally, in the table below, we see something that does seem statistically significant for this data set: the Haves vs Have Nots effect increases with financing stage sequence, which doesn’t seem too surprising. When an Uber is raising Angel money, it is just another intriguing startup story, and it commands less of a premium relative to all other financings. But, when they are raising a $1.2B Series D, they have the staggering financial metrics to prove they are the hottest company around.

Summary statistics from plot above.

Summarizing, this data leads us to three interesting conclusions:

  1. Venture financings have always been about Haves and Have Nots, with Average Pre-Money Valuations for the few hot companies far outstripping the valuations of the large number of less hot companies.
  2. The Haves vs Have Nots effect is more pronounced with later sequences of financings, with a high degree of statistical certainty.
  3. There is only weak statistical support for the idea that the Haves vs Have Nots effect has increased over the last 20 quarters.