The potential range of outcomes from investing in a private fund is typically much wider than that from investing in mutual funds or exchange-traded funds.1 In turn, among private funds, venture capital (VC) has the widest potential range of outcomes.2 The early-stage companies in which they invest have the potential to return very large multiples - often enough to generate a significant profit for the fund on their own. But also, approximately 35% of Series A startups fail before they can raise a Series B financing round (more on financing rounds below).
This elevates the importance of selecting the right VC manager. In this article, we’ll break down one element of our approach to fund selection, laying out:
The wider performance dispersion in VC and its implications for investing
The importance of “secretive” networks to accessing high-quality VC deals
How we use network analysis in our VC fund sourcing process
Put simply, the bigger the difference in performance between fund managers is, the more important selecting the right one becomes.
For example, the gap in annual returns over 20 years between the 25th and 75th percentile performing US Small-Cap Growth fund was just 1.2% (see chart below). For multi-stage venture capital funds, the equivalent gap was 18.2%. Picking an upper-quartile manager can make a huge difference to your portfolio performance.
Why is the dispersion so much wider? There are several reasons, most prominently:
Limited information: With less available information, private markets are less “efficient” in terms of setting valuations. What that means is that fund managers may be able to find much greater value in an underpriced asset, but also leaves greater room for error when making that assessment.
Active not passive management: Private fund returns are often a result of active value creation strategies at the companies or assets invested in. That leaves a lot more space for either adding or losing value over time depending on the quality and approach of the fund manager.
Less diversification: Private funds are often narrowly focused on a particular sector, industry, or geography, which can make them more vulnerable to industry-specific or regional risks. In addition, private funds generally invest in a less numerous portfolio of assets than most public funds.
Venture capital funds invest in the earliest stages of company development. As such, it has the most limited initial information - elevating the importance of networks, which we discuss below. This results in venture capital funds having the widest dispersion of returns among private asset classes (see chart below).
So, how do you ensure (or at least increase the probability) that you end up at the top end of this dispersion? We believe that the extent and strength of a manager’s networks is a predictor of persistently strong returns.
Networks are of heightened importance in VC due to the way investments in the asset class are structured.
Venture capital funds typically acquire minority stakes in early-stage companies. Investments are generally structured in rounds of investment (see chart below), with multiple funds often involved in any given round. Below is an illustrative example of the fundraising life cycle of a successful venture-backed company.
Each fundraising round inevitably determines a new company valuation - a process referred to as price discovery - from the sale of a partial equity stake. Often, the new valuation is higher than the old one, and competition for stakes in exceptional companies can be fierce. This competition may be heightened further because companies’ need for capital is typically limited, as is the share of the company made available for investment, with founders generally looking to retain as much control and equity as possible in their business.
The competition between venture capitalists for access to funding rounds for exceptional companies has led to the development of highly “secretive” networks. One of the most famous networks is the “PayPal Mafia,” which includes Peter Thiel, Elon Musk, Max Levchin, and Keith Rabois. This network has backed multiple breakout success stories, such as Uber, Airbnb, Palantir, SpaceX, Square, and Stripe.
The importance of these networks contributes heavily to the wide dispersion of performance among VC funds. Funds with strong networks may gain preferential access to high-potential companies at their earliest stages. This can help drive stronger returns. This success further builds industry networks and improves access to high-quality deal flow, which in turn may improve performance of successor funds, and so on in a virtuous cycle. This is a major factor in the strong persistence of returns in VC.
Though nothing is ever guaranteed, robust networks are therefore a predictive factor for strong fund performance - particularly for first-time fund managers or other scenarios in which performance data is unavailable. However, applying this insight can be difficult. Network data is not readily accessible, particularly given the relatively secretive nature of VC, which means you have to get creative.
Our due diligence process has many stages and examines numerous facets. An established track record of strong fund performance is, of course, a great first data-point to start with when accessing a potential VC investment. When this is not available, we look to use deal-level data from proprietary data sources to analyze, benchmark, and examine a fund’s follow-on rate.
Follow-on rate approximates performance by calculating the percentage of investments in a fund portfolio that receive a second (or follow-on) investment by a given manager. For example, if a manager funds two companies and only one receives a follow-on investment, the manager’s follow-on rate is 50%.
The below chart shows follow-on rates at different stages of the VC-backed company lifecycle across deal return performance percentiles. What it demonstrates is that follow-on rates vary widely, but across all stages of development the deals that generated the best returns (fairly intuitively) attracted a higher rate of follow-on funding. In short: follow-on rates have historically correlated with deal performance.
Follow-on rate is an important metric, but - as discussed above - recognizing the importance of networks in persistently strong performance, we use methods from the field of network analysis to gain greater, hopefully predictive, insight.
Network analysis helps us process and interpret large and difficult to understand datasets. For example, below is a visual representation of the importance (or “centrality”) of a VC firm within the overall investment ecosystem. Each dot represents a venture capital firm with each line between them representing an investment in common.
From this data, we create a “centrality” score to determine the firms most central to the VC ecosystem, which is indicative of the relative strength of their networks. Similar to our analysis above of follow-on rate, we aggregate centrality scores and benchmark managers based on performance. The chart below compares centrality scores with deal performance at different percentile levels for deals at various stages of the investment lifecycle.
This data-point builds on our other insights to help inform our investment decisions.
While individually both follow-on rate and centrality show a correlation with performance, it is important to demonstrate the forward-looking linkages between the two. Specifically, we would want to see a stable positive relationship over time between centrality and follow-on rate in the past and follow-on rate tomorrow - which would in theory, allow us to predict future performance with more confidence.
To examine this, we applied regression analysis to two datasets, one from 2012 to 2017 that included centrality score, follow-on rate, and entry round, and another from 2017 to 2022 including just the follow-on rate. This analysis showed a statistically significant positive correlation between the two data-sets. Put simply, more extensive networks and higher follow-on rates appear to have been predictive of stronger future performance.
The chart below highlights the direct relationship between the two variables of greatest interest within our model, namely centrality in the first five-year period and the follow-on rate in the five years that followed.
Essentially, our network analysis:
Confirms prior research finding strong evidence for persistent performance among VC funds, which heightens the importance of fund selection in the asset class.
Demonstrates that appropriate network analysis of historical data should help with sourcing VC funds, and may increase the probability of good fund selection.
Venture capital is a highly competitive and secretive industry that relies heavily on networks. Our use of the quantitative methods we have discussed in this article is just one - if a highly important - piece of the puzzle in identifying top managers.
We nevertheless find that our proactive data-focused approach offers us a more holistic view on how the venture ecosystem is developing, while potentially giving us an edge over other fund investors in an asset class in which appropriate fund selection is critical.
To discuss further Opto’s rigorous approach to fund selection and due diligence, get in touch at partner@optoinvest.com.