Over the past decade, Tech In Asia has been actively tracking and reporting on funding news like these. In the process, we have amassed a huge database of investment data focused on Asia, so we know which startups that venture capital firms have invested in at different funding stages (seed as well as series A, B, and so on). Using the data, we would like to share some interesting insights with you.
VC firms are very important players in the startup ecosystem – they help nascent startups raise capital to turn ideas into products. Almost all of the largest startups in Asia today relied on venture capital when they began.
The evolution of Asia’s startup ecosystem
In the last 10 years, the region’s startup scene has grown tremendously with the rise of unicorns such as Grab, Gojek, Lazada, and Tokopedia in Southeast Asia as well as Toutiao and Didi Chuxing in China. Vertex Ventures, East Ventures and Sequoia Capital were among the early investors in these companies.
Why should we care ?
Asia’s share of the global economy has more than doubled over the past 40 years, and it’s expected to continue growing. And according to a study from Bain & Company, Southeast Asia’s investment ecosystem is entering a new phase of growth, producing at least 10 new unicorns by 2024.
Visualizing the data
To help readers dive into the Asian startup scene, we’ve created bite-sized content like infographics and listicles that are updated regularly. Conventional analysis techniques involve examining the total funding raised, checking the number of deals by country and/or verticals, among others. But we have always wondered, is there a more natural way to view the startup ecosystem?
Suppose we have a round of startup funding represented by a graph that looks like this, where each company represents a node and the investments made are the edges (also called vertices/links) in a network:
Simile Venture Partners and East Ventures invest in Tech In Asia’s Series A on Jan. 22, 2013 / Image: Tech In Asia
Connecting all the data points would give us:
A network visualization of Asia’s startup funding / Image: Tech In Asia
A network visualization of Asia’s startup funding in Asia (click to enlarge) / Image: Tech In Asia
Note: Investors outside of Asia are included because we focus on the target country.
Using colors to differentiate VC firms by country and resizing them by their number of links, a few key insights can be glimpsed:
The startup ecosystem in Asia is very well-connected (and looks like abstract art)
VCs typically have a geographical concentration because of physical proximity, except for a handful that have diversified portfolios across the region.
Established VCs such as Sequoia, IDG Capital, 500 Startups,and East Ventures are more prominent due to the number of deals they have.
Each of the major VCs invest in a small cluster of portfolio companies that have no links to others. Are they likely to be up-and-coming startups?
Humanizing VC firms
If a VC invests in a startup, we can assume that information flows both ways – or undirected – between the parties, whereas an investment is directed from the investor to investee. If we treat their link as an undirected connection between companies – much like a social network of friends – we can compute and analyze certain statistical properties in the network based on the relationships.
Financial networks have been studied extensively in the academe. A notable work is a 2007 paper titled “Whom You Know Matters: Venture Capital Networks and Investment Performance” which contends that better-networked VCs experience significantly better fund performance, as measured by successful exits, initial public offerings or M&As.
Here are some of the reasons why:
VCs draw on their powerful network of service providers, such as headhunters, patent lawyers, and investment bankers, to help portfolio companies succeed. This is known as the “venture capital keiretsu effect.”
Better network provides access to better quality deals (select promising startups) and information.
The ability to value-add to portfolio companies (strategic alliance) allows VCs to negotiate more favorable deal terms.
Co-investments/syndication enables VCs to pool correlated signals to reduce uncertainty and to share vertical-specific or location-specific knowledge to diversify portfolios.
But how do we quantify how networked a VC is? Read on to find out.
Analysis of VC network
Drawing on the similarities between VC networks and social networks, a popular theory asserts that on average, people are distanced by approximately “six degrees of separation.” Researchers at Microsoft proved that theory right: They worked out a precise number of 6.6 degrees by studying records of 30 billion electronic conversations among 180 million people in different countries.
We did a similar calculation for the Asian startup scene by using mean path length based on the network above and got 6.2 – that’s the average number of “hops” it takes a company to reach another based on these links.
Social network analysis (SNA) leverages the graph theory to measure the importance of nodes – each representing a company – in a network, which in our case would be focused on VCs. Here are the measures we used to quantify influence, how they’re calculated, and what they mean:
Definition of measures What it means
Degree Centrality – The number of investments made by each company Shows that very connected VC/startups are likely to hold most information – in this case, it’s the number of deals made
Betweenness centrality – The number of times a node lies on the shortest path between other nodes Identifies VC/startups that act as bridges across different communities
Image credit: NVIDIA
In the mini-graph shown above, Node  has a high betweenness because it acts as a bridge – the communities on the right and left rely on it to pass through
Closeness centrality – Represents how close one node is to all the other nodes, a measure of how fast it can get to all the rest of the network Finds the VC/startups that are best placed to influence the entire network in the shortest possible time
Visual Illustration of Different Centrality Measures
2011 paper by Donovan Isherwood and Marijke Coetzee
(Adapted from Cambridge Intelligence)
Using our database, we computed the measures above and came up with a list of VCs with their respective centrality rankings, limiting it to the top 100 firms based on the number of deals (please click to sort). Our goal is to quantify and rank VCs based on their number of investments/deals with startups, so we did not take into account the investment amount, fund size, IRR and other performance metrics. We also noted other limitations to our data below.
Matrix Partners China (经纬中国)
Shunwei China Internet Fund (Shunwei Capital) (顺为资本)
SMBC Venture Capital
To sift through the clutter, we performed a quadrant analysis that allowed us to develop a list of the most well-connected VCs in Asia, as of September 2019.
Quadrant Analysis: VC Degree, Closeness and Betweenness (click to enlarge) / Image: Tech In Asia
Keen observers will immediately notice that these three scores – degree, closeness, betweenness – are positively correlated. In this case, having more investments (high degree) generally led to a better-networked position.
However, there are exceptional cases of VCs with a large number of investments and low closeness or betweenness scores, or VCs with fewer investments and high betweenness scores, etc.
Why is this so? Are there unique characteristics to these VCs? The short answer is, yes. We constructed a matrix to explain the attributes of different possible segments:
Low degree Low closeness Low betweenness
High degree ✗
High closeness ✗
High betweenness ✗
Here’s a quick recap of what these measures mean:
High degree – Well-connected VCs or startups that are likely to hold most information, i.e., the number of deals made
High betweenness – VCs or startups that serve as bridges across communities
High closeness – VCs or startups that are in the best position to quickly influence the entire network
Connectedness Matrix (click to enlarge) / Image: Tech In Asia
There are some savvy investors with relatively fewer investments but hold a strong position in the network, such as Strive (Gree Ventures), Temasek Holdings, and Goldman Sachs [bottom left]. On the other hand, there are niche investors that have plenty of investments but remain isolated in a bubble of their own [top right]. Keep in mind we have only included the leading 100 VCs, so the rankings are relative and will definitely change if we take into account less active VCs (based on the number of deals).
Here are some areas where network data can potentially help:
Investors (VCs) Startups
Spotting candidates for partnerships or co-investments
Pinpointing startups that could add value in terms of their network of current investors
Identifying cliques or communities of investors within a location or sector
Showing which VCs provide the most social capital
Indicating which VCs can facilitate in geographic expansion
Picking out which VCs can bring in additional capital through syndication
There’s a lot of information to digest, but we’ve only reached the tip of the iceberg. A lot more insights can be derived via further network analysis or metrics such as eigenvector centrality, PageRank, graph density, community detection, and so on, in addition to drilling down to vertical-specific or country-specific analyses. The possibilities are vast, so it all boils down to asking the right questions, as various metrics address different problems.
Limitations and future research
Here are some of the limitations of our current analyses:
An assumption that we made earlier asserts that information flow could go both ways between investors and investees. This may or may not be true in all cases, and it will affect the computation of centrality measures.
Knowing the funding amount for each investor is vital in providing weight to the strength of the relationships. For example, an investor who has put in more money definitely has stronger ties. But both VCs and startups typically keep this confidential, so it’s a challenge to get access to such information.
Data collection, methodology and processes also have their own limitations.
Nevertheless, we believe there’s great potential in the following areas:
Studying the relationships between investors, co-investment/syndications and identify sub-communities – which could be possibly rivals – within larger communities
Investigating whether small-world behavior in VC networks, which researchers have observed this in China, is also happening in Southeast Asia and if the phenomenon affects fund performance
Assessing the performance of VCs currently in Asia that co-invest with those similar to themselves – US-based VCs that did this tended to fare worse, according to a 2012 paper called “The Cost of Friendship”
Conducting similar analyses focusing on the startup network to pinpoint which are the most well-connected ones
Developing a time-based evaluation of how hot verticals like e-commerce, transportation, and fintech in Asia have boomed over the years
Identifying relationships between startups based on similar investors
Modeling relationship data to:
gauge the success potential of startups
use structural balance theory to link these predictions
predict which VCs will invest and possibly work together
figure out how to use this information to your advantage
Overall, this is a very exciting topic, and we’re just getting started. We would love to paint the most accurate picture of the startup scene in Asia. But our insights are only as good as our data, and we acknowledge that there may be lapses/inaccuracies. We aim to continuously improve our methodology and processes to serve the community better.
If you spot any errors or missing data, please feel free to reach out to us at email@example.com.
Note: If you are a researcher and would like to collaborate with us, we’re all ears!”