I’m currently putting the finishing touches on a new study about women-founded venture-backed companies in the United States, which should be out in a couple weeks with the Center for American Entrepreneurship. One of the things I looked at is exit rates—the share of companies either being acquired or doing a IPO—by the gender composition of founding teams. To do this, you have to group companies into cohorts and follow them over time—in this case, by the year of first financing and after eight or ten years down the road. The challenge with doing this type of work is that it requires many years’ worth of data and by definition you have to ignore the most recent cohorts (which everyone is the most excited about).
A colleague who reviewed a draft of the study challenged me on the eight- and ten-year exit lag from first financing because the time to exit has gone up in recent years. That’s a fair point, but I only have data going back to 2005 (the oldest first-financing cohort in my data), which constrains my ability to look over longer time periods. It is still an important exercise, and most critically, the results of the comparative analysis between women-founded and non-women-founded companies wouldn’t change much by having more data. And, that’s what I’m most after in the report.
But that did get me to thinking: just how much longer is it taking for venture-backed companies to exit?
I pulled the data from PitchBook to get an idea. The charts below show the time (in years) from first financing to exit, for U.S.-headquartered venture-backed companies that were either acquired (including buyouts) or that issued shares through an initial public offering (IPO), listed by the year of exit. The results are interesting.
For acquisitions, the time to exit has increased across the board, and in particular since around 2013 when we see a sharp uptick in at the average, median, 10th, 25th, and 75th percentiles (which starts moving up earlier, in 2011). The average time to exit by acquisition is now 6.3 years from first financing, the median is 5.4 years, the 75th percentile is 8.9 years and the 90th percentile is 12.4 years. Those same figures in 2005 were, respectively, 4.6 years, 4.7 years, 5.7 years, and 7.4 years.
We also see a widening dispersion of exit times since around 2005. The spreads at the middle and the bottom-half of the distribution haven’t really changed much—the difference between the median (50th percentile) and the 25th or 10th percentiles is more or less the same, and the spread between the 25th and 10th percentiles has actually tightened a bit. The spread between the average in the median did widen after 2008, but it has been steady since 2013.
However, these changes pale in comparison to the widening of the spreads driven by the top of the distribution. Consider these data points. The split between the 90th percentile and the median was 2.7 percentage points in 2005, but by 2018 it widened to 7 percentage points. The split between the 90th and the 10th percentiles was 5.9 points in 2005 and it is 10.4 points today.
Moving to IPOs, we see some similar patterns to acquisitions but also some major differences. The data here are noisier, due to a smaller number of observations—since 2005, there were nearly 7,800 acquisitions compared with less than 700 IPOs.
The overall trend here is a tale of two worlds. On the upper end, we see a significant increase in exit times since 2005, as the 90th and 75th percentiles increased from 8 and 6.1 years, respectively, to 13.6 and 10.7 years in 2018. These increases at the top have pushed the average up too—from 4.8 years in 2005 to 6.6 years today.
On the other hand, the median, 25th percentile, and 10th percentile are all down from 2005, which has had the effect of widening the spread in the middle—creating a bifurcation in time to exit for IPOs. This has all occurred due to post-2012 activity because each of these series increased substantially between 2005 and 2012, but have fallen sharply between then.
***
What has been learned here? First of all, we’ve seen a general increase in time to exit for venture-backed companies since 2005, and more specifically after around 2013. For both acquisitions and IPOs, this has been driven substantially by an increase in the companies taking the longest to reach an exit. There are major differences at the lower end (faster exits) of the market, where time to acquisition has increased slowly and uniformly, but time to IPO has collapsed (speeded up) after around 2012.
Going back to the original motivation of this post—to see whether looking at eight or ten year time lags to venture-backed exits captured a reasonable share of the market—I feel pretty good about those choices, particularly in the face of the data constraints imposed by time. Even today, ten-year-lags to exit from first financing will capture most of the acquisition and IPO events. And, in the case of the study I’m working on, these seem to be more or less consistent across the founder types.
Finally, taken in context with some work I did a few months back on funding and valuation trends, an interesting picture emerges—one that shows fewer companies getting funded, those that do are raising more capital, reaching higher valuations, and taking longer to reach an exit, compared with the past. To partly compensate for the capital and time commitments, investors in certain—though not all—segments of the market are taking a larger ownership stake in these companies.