Network centrality is a key driver of venture capital fund performance and of our Investor Mosaic models. Given the asymmetric nature of information in the VC market, the more networked you are (with the right investors and ecosystems players), the better fund performance.
And so given this, understanding which investors are more or less likely to invest with one another in a syndicate can be quite useful.
To try and answer this question, we used a technique called Association Rules and applied it to our historical data set of venture capital fundings at all stages in the last 5 years.
Association Rules are used extensively in many industries, especially retail, where they are used to calculate the change in probability of adding an item to a shopping cart, given the items that you already have. As an example, it helps calculate whether you are more or less likely to pick up a six pack of beer if you already have tortilla chips and salsa in your cart, than you were if you did not have those items in the cart. If we apply this analogy, individual investors are items and funding rounds are shopping carts and we thus calculate the change in probability of an investor being in an funding round with another investor.
We wanted to share some initial results that we found interesting – first, some of the positive results –
The way to read this table is “If Investor A participates in a funding round, the likelihood of Investor B participating in the same round increases by [Lift]”. This does not make a distinction between which investor came to the table first but rather is an ex-post analysis. Thus, in the table above, DAG Ventures is 37.62 times more likely to invest in a round where Benchmark Capital is also present.
Similarly, we found negative effects as well, that is, presence of certain investors in round decreased the likelihood of co-investment by certain other investors. Here are some of these negative effects –
Thus, the data suggests that if Draper Fisher Jurvetson invests in a round, the likelihood of NEA co-investing goes down 0.60x (40%). This can likely be attributed to different investment theses including sector, size and round focus rather than an acrimonious relationship. Nevertheless, it is interesting to see stalwarts of the industry like DFJ and NEA co-investing less frequently with each other than the baseline.
** Note – The results shared above are only to demonstrate the applicability of this technique to this kind of data and are not meant to be an authoritative statement about the co-investment relationships or friendships (or lack thereof) between venture capital firms.
We’ve previously reviewed investment syndicate relationships of firms here —
- Andreessen Horowitz and Sequoia Capital
- Google Ventures
- Union Square Ventures
- Syndicates among tech’s 59 most active firms
- Healthcare VC syndicates
- Felicis and Floodgate (micro-VCs)