Seed to Series A graduation rate in Germany
This post is part of a series covering the German and European VC industry by numbers. To read the starting post with a general overview of VC in Germany and sources used click here.
In this post I present the results of some research I did on one critical point for an early stage investor: graduating Seed companies into Series A.
If a company gets a Seed what are the chances it gets a Series A ? And how long does it take?
See some kind of pattern ? Let me draw it in the chart below.
One interesting stuff I observed from the data is a correlation between Seed and Series A lines. Data gives a nice visualization of Seed deals building up a pipeline of Series A. Digging a bit I found out that Series A correlate with Seed at most after 3 quarters (to know more about this check the appendix for statistics lovers).
Data shows that you can expect 3 quarters lead time from Seed to Series A
But what about the graduation rate ? I proceed and calculate the graduation as:
#of Series A in Quarter X+3 / # of Seed in Quarter X
Data shows that the median graduation rate is 42%.
Take it with a pinch of salt because of data quality and low statistical significance.
Data shows that graduation rate decreased with time. While it does not make sense to look at single observations I believe that the trend makes sense and validates the way I calculated the graduation rate. Looking back at the chart with #of Seed and Series A we observe an explosion of Seed deals in 2013 and 2014 only partially followed by an increase of Series A deals. This also reinforces the notion of the abundance of Seed capital in the German market.
Appendix for statistics lovers:
Here is how I came up with the 3 quarters time to Series A.
For starter I calculated the correlation between the number of Seed and Series A as they are. I get a correlation coefficient r= 0,344796. No correlation. (r=0 is completely uncorrelated data, r=1 is perfectly mirrored data)
Secondly, I built a “build-up time” or as you would say in econometrics a “lag”. I did this for +1,+2,+3,+4,+5 quarters. Then I proceeded and calculated the correlation with the different lags. The lag with the highest correlation is +3. See below a table of the data with a 3 quarter “build-up time”.
With this data structure I can compare the way Seed correlate with Series A after 3 quarters. With this data I get r= 0,659055. Nice! So Seed and Series A are indeed correlated with a 3 quarters lag.