Diversification Doesn’t Work (in VC)

It has become conventional wisdom that the secret to success in investing can be summed up in one word: diversification. While that perspective has driven trillions of dollars into low-cost index funds, it fails in seed-stage venture investing.

Study after study has shown that holding a simple passive index of equities (mirroring the S&P 500 say, or the Russell 3000) consistently outperforms active fund managers, after accounting for fees. Modern portfolio theory suggests that investors are best off owning a portfolio comprising the entire market. Diversification is strength.

Surprisingly, this thinking has made its way into the venture capital market as well. Many observers have noted that venture returns are driven by a very small number of very successful companies. A fund that bought an early stake in Facebook can see every other investment go to zero, and still do well. Therefore, the thinking goes, VC managers win only when they get into those few winners. And the best way to do that, especially at seed stage? Hold a large, diversified portfolio! Thus, we see a new breed of seed investors and incubators who take small stakes in many, many startups. 500 Startups, Y Combinator, Tech Stars, etc etc etc.

Diversification does not work for funds making seed investments. The reason is obvious: the expected return in venture capital is terrible. The Kauffman Foundation’s 2012 report on the VC market demonstrated that the majority of funds underperform the Russell 2000 (a public small company benchmark), and only a handful beat the Russell 2000 after fees. In other words, the average VC underperforms the public market, without the benefit of liquidity.

Holding a huge portfolio of venture investments will approximate the performance of the entire venture market. The more holdings, the more closely you’ll approach the performance of the sector overall, and that performance is not good. A highly-diversified VC is designed to be a bad investment.

If the alternative is concentration, we face another conundrum: where to concentrate? If I need to assemble a portfolio with just a few investments that outperform, how do I select those investments?

One perspective is that you have to have access to “the best deals.” This perspective assumes that you know a priori which venture investments are going to deliver the best returns, and that the challenge is getting those companies to take your money. In this worldview, the key differentiators are brand, network and relationships. The best companies (i.e. the hardest deals to get into) want money from the best investors (i.e. the highest-profile funds). I might be convinced that such a strategy can work for growth stage companies (series B+, where sales are ramping exponentially), but only for the Sequoias, Andreessens and Union Squares of the world. Such a model is very hard to replicate for new entrants, and in any case, is unlikely to work at seed stage.

A better way to build a concentrated portfolio of seed-stage investments is to find companies where you, the investor, can actively add value. That is, find companies that you can invest in, then play an active role in making the companies more valuable. You buy in at the “pre value-add price” and exit at the “post value-add price.” If you make an investment and then materially improve the value of the company, you capture outsize returns. Even better, you capture returns that are not available to anyone else.

Executing this strategy requires specialization in an underserved niche, and requires unique capabilities that are relevant to that niche. It also requires doing hard work “in the trenches” with entrepreneurs. It is insufficient to simply share your network with your portfolio — really making a difference means working really hard on behalf of the founders you back. That work leverages your invested capital, reduces risk, and makes everyone (you, your founders, your LPs) more successful.