Recently came across a fascinating Twitter thread from June 2020 by Dave McClure, ex-founder of 500Startups, where he talks about how “investing with conviction” is a myth. This tweet captures his sentiments well:

I agree with several arguments in this thread:

1/ Picking winners in early-stage investing is really hard. Power laws govern the best venture portfolios, driving down the hitting %. Per Horsley Bridge data, even for a top VC firm like Sequoia, ~4.5% of portfolio companies generate 2/3rd of aggregate returns.

2/ Intelligent venture investing, by its very nature, involves making both Type 1 and Type 2 errors. Therefore, even high-conviction deals are likely to exhibit unexpected outcomes, both positive & negative.

3/ There is a lot of hindsight bias in the way investor narratives are created around companies that turned out to be successful – “Look, I had high conviction on this deal & it turned out exactly as I expected. Ergo, I can predict the future”.

So in games like this where outcomes are random & often uncorrelated with the level of effort that goes in, does it make sense to discard the input process?

Based on more than a decade of venture experience, I tend to view it differently. I believe it’s still important to have a rigorous process of building conviction and to keep improving it bit by bit with each experience. Even though eventual outcomes might still be random, this approach helps tilt the playing field a little in your favor every time. Over a long enough time horizon, as one keeps taking more shots at the goal & with continuously improving odds, the hope is that a home run arrives sooner than later.

Particularly at the earliest stages (angel/ pre-seed/ seed), especially with the advent of small check investments ($1-5k via syndicates/ SPVs) attracting a new generation of 1st-time investors, it’s easy to assume that outcomes are randomized & therefore, fall into the trap of doing spray-and-pray that isn’t backed by an intelligent investment process.

It’s important for new angels to first deeply study the asset class & build their personal investment process – areas of expertise, focus sectors, stages, target founder persona, deal flow engine, unique value-add to get into best deals etc. Post which, the odds of success are significantly better.

While being a champion of a “conviction-building” investment process, I also agree with the 3 takeaways that Dave closes the thread with, regarding having enough shots on goal:

Even with the most intelligent investment process, venture investors need to acknowledge their limited picking ability & therefore, keep taking enough intelligent shots at the goal for the odds to work in their favor. Semil Shah of Haystack wrote a great post titled “Shots on Goal” on this idea a while back.

Equally important as portfolio diversification via numbers, is making asymmetric investments – ensuring that the few winning bets have huge outcomes so that even with a high loss ratio, the returns math still works at the portfolio level. The smartest thing a venture investor can do is to befriend the power law, and work towards being on the right side of it!

To summarize, acknowledging the randomness of venture outcomes doesn’t need to be at odds with running a rigorous & continuously-evolving investing process. In fact, such a system should be intelligently designed to account for this randomness, combined with other considerations like power laws, compounding, economic cyclicality etc. Even a few points of “edge” that is systematically created with each experience, can slowly accumulate into a sizable alpha over the long term.