Hunting for Unicorns: Venture Capital and Startups
#probability #correlation #regresstomean #fallacies #induction
Everyone wants to win big, but if you’re going to take a big risk you’d better believe you have a pretty damn good shot. Remember that one time you played the claw machine and won? You just knew you would win, you could feel that your chances were better. Bull. We humans are prone to overweighting our chances of success based on how we emotionally feel about something, or worse, basing our predictions of future successes on instances of pure luck in the past. We are full of biases, and we would do best to avoid them whenever possible, relying on hard evidence instead of our “guts.” This is the same for venture capital investments, and finding a metric to analyze the thousands fund-hungry startups for future success is key if a firm ever hopes to find the next Uber or AirBnb, the next Unicorn.
A few weeks ago, I took a trip to 500 Startups with many of my peers to get a run-down of what to look for in new startups. After the visit, we were given access to 100 startup applications to review. After the day was done and I had looked through hundreds of these startups, I realized:
- There are a lot of shitty startups out there
- Sometimes something you think is dumb shit turns out to be Uber
- You’re gonna feel real bad after that, but hey, live and learn
This effect is another issue of us humans. When we fail, we regret it big time. And although it can feel incapacitating it really is best to get up and try again, this time all the wiser.
Overall, it is still the most important to consider the facts in VC decisions. After a big success, there will most likely be a big failure. This statistical principle is called regression to the mean and explains why making only investments that paid off well in past will fail as a strategy. While the culture in the silicon valley may make it seem like success is overflowing, this is really just another human bias (availability bias) because all the startups we can think of are the ones advertised and used widely, and the only ones advertised and widely used are the exceptionally successful ones. Take this statistic to heart:
Despite the recent surge in startup growth and VC investments, 80–90% of startups fail.
That’s not so bad you say, that means 10–20% succeed? Nope. This really means that just some didn’t completely bomb, most of that 10–20% is barely scraping by or is just OK. The funny thing about VC though is that it only rewards the investor if the company hits the jackpot, so just “not failing” isn’t enough.
Out of the 1300 companies 500 startups chose themselves, 250 (19%) are doing well, 25 (2%) are great, and 3 (0.23%) are absolutely killing it (Unicorns).
The recent growth in the Silicon Valley is hugely important, but it seems that quite a bit of people are just riding the bandwagon with everyone else, where everyone thinks the money is. Just because everyone else is doing it though, doesn’t mean it’s best. The graph above shows a correlation between the 2008 market crash and the VC invested, they both decrease. In the future, it is statistically likely that the VC market will crash again, it’s only a matter of when.
But hey, let’s not get too negative — the market right now is pushing a lot of good developments out and creating a faster paced model for businesses, one that might just be able to iterate into the solutions to the world’s biggest problems.
A good sense of induction principles is required to judge startups, being aware of biases, and realizing that decisions are temporary and must reviewed again later. A level and logical head is important, but I cannot deny emotion is important as well — every startup has to come in looking like, feeling like a unicorn. If there is one thing true about startups, it’s that doubt is the #1 killer — if the founders don’t believe they are the hottest shit ever to exist, they will never succeed. Persistence and faith, however, offer hope.
