LINC’s Matt Guttentag recently presented at the ANDE Metrics Conference on how systems tools can be used by the impact investing and Small and Growing Business (SGB) sectors to develop more effective programming. One part of this was a demonstration of a rapid Social Network Analysis (SNA) of impact investors to identify closely connected communities and the investors who seem to bridge these communities. The presentation is laid out below, and can also be viewed here (recommended for viewing on mobile platforms).
The Overall Network
The map below represents a network of approximately 4,000 investments among over 400 impact investors, scraped from ImpactSpace. You can use the zoom buttons on the right side of the map to zoom in and out, or press the double arrow button to automatically zoom to scale.
Each node (dot) represents one investor. Each line represents an instance for which two investors funded the same company (a “co-investment” for the purposes of this presentation, although this differs from a true co-investment).
Mapping these connections allows us to understand which investors are most central to the network, where there are “clusters” of closely connected investors, and some key ecosystem gaps.
Notice that there is one large group of investors in the middle who are all either directly or indirectly connected to one another through their investments, while many other much smaller groups of investors are around the periphery and disconnected from this larger group.
First, we ran a metric called “Betweenness Centrality” that quantifies the extent to which any given investor lies on the shortest “path” between any other investors in the network. Basically, this tells us which investors are important “connectors” across the whole network.
This map adjusts the size of each investor based on its Betweenness Centrality. We have highlighted the top five investors: Village Capital, the Gates Foundation, Omidyar Network, Investors Circle, and Impact Engine.
This gives us a good sense of the most connected investors overall. But the network seems to have some sub-groups of investors that are particularly closely connected to one another.
Using an algorithm that identifies densely connected “communities” within a network, we discovered 12 sub-groups. We color coded the eight largest subgroups in the map below.
This raises interesting possibilities to understand the nature of investors “cliques” in terms of what types of investors tend to invest together, and where there are gaps, and who is bridging these gaps.
Let’s examine these clusters in more detail.
The largest cluster seems to consist heavily of investors who are either ANDE members or who are part of the larger ANDE community. This includes Village Capital, USAID DIV, GSBI, Acumen, Artha Ventures, Unitus Seed Fund, Shell Foundation, Growth Africa, and LGT.
Indeed, the network shows that ANDE members tend to invest with one another. Using a technique called “Exponential Random Graph Modeling” (ERGM) that determines the probability that a new node with a given characteristic would connect with any other node in the network, we found that a new ANDE member in the network has a 5% chance of investing in the same company as another ANDE member, but only a 1% chance of investing in the same company as a non-ANDE member.
Clusters 3 & 4, on the other hand, seem to include a number of investors more directly in the Silicon Valley and New York VC worlds, such as DBL Partners, Andreessen Horowitz, Tao Capital Partners, and General Catalyst Partners:
Finally, cluster 8 seems to include a number of corporate investors such as Johnson & Johnson Ventures, Monsanto Growth Ventures, Merck Global Health Innovation Fund, and Syngenta Ventures:
Identifying Bridging Investors
Now you can play around with different combinations of clusters to see which investors seem to be “bridging” these different communities through their investments.
In the map below, you can select specific clusters from the drop-down at the top. And using the buttons at the bottom, you can toggle between all connections, and only connections within or between clusters. If this is too small on your screen, you can access a larger version of this map here.
For example, select clusters 2 and 4, and then “Show Only Connections Between Clusters”. You can see that the most significant “Bridgers” here are Village Capital, the Gates Foundation, TONIIC, Investors Circle, and Serious Change.
Do the same thing with clusters 2 and 3, and you can see that Omidyar Network, DBL Partners, and the Dell Foundation seem to be important bridges.
Select clusters 2 & 8, however, and you will see that there are no connections that bridge these communities, seemingly a significant gap!
This rapid analysis is not meant to be exhaustive or authoritative, as it includes only self-reported investment data and does not dive into the many reasons that investors might be funding the same companies. However, it demonstrates the potential for SNA to uncover hidden patterns among investor communities, which could help guide both ecosystem builders as well as entrepreneurs themselves to engage more effectively with the impact investment community. Contact Matt Guttentag (email@example.com) if you are interested in how SNA or other systems tools can be used to build stronger impact investment and SGB communities.