Fixing the problem of startups and their inherent power law
The saying “go big or go home” (or “go big or GTFO”) could very well have been invented in San Francisco — the city with the highest concentration of startups anywhere in the world. In this business you either become a billionaire or like most — go bankrupt.
Instead of such an uneven distribution, what if the startups shared their wealth? What if they owned a stake in each other and if one made it, they all did? What if founders had more of a portfolio risk just as their investors do?
Most founders have probably heard about the Power Law, the 80/20 rule or the Babe Ruth Effect in venture capital — the fact that only a few startups become successful and the rest end up being worth close to zero. These founders recognize that the chances of making a successful startup exit are slim. Let me just quickly sum up the statistics so everyone is on the same page.
A top European seed fund like Passion Capital gets around 1,500 pitches each year, of which they invest in about 20 (3%) — you can check out their infographics here. Out of those 20 companies roughly 10 of them run out of money within the first couple of years. Out of the remaining 10, around 5 become walking dead, 3 become middle cases with a small return and about 2 become a real success with a return of 10x and upwards. Even if you were in the special 3% there is still only a 10–20% chance of a real exit.
The same metrics are true for most seed funds or accelerators like Y-Combinator. Out of the 842 companies they have funded as of this fall, just four of them account for more than 80% of their portfolio value (the 4 +$1B companies or unicorns as they call them). Take a look for yourself — YC’s stats are summarized here.
Paul Graham most likely started Y-Combinator because he had figured out something that other VC’s hadn’t — making accurate predictions on startups is much harder than expected.
Since the dawn of the VC industry in the 1970’s, investors knew that their startup investments should be diversified and treated as any other portfolio like stock or bonds. The question that remained was how diversified such a portfolio should be? Thirty years later and even after the dotcom bust, VC’s were still overestimating their own ability to pick winners.
Why? Because most VC’s had previously been successful entrepreneurs themselves and had so much winners bias from their own experience. They had completely forgotten about regression to the mean — the fact that most startups fail.
Black Swan theory is used to explain the disproportionate role of high-profile, hard-to-predict and rare events.
Did Paul Graham have more insight than most, or perhaps he was just humble enough to understand that these predictions were very hard and that finding black swans required much more risk taking.
Even when picking a great team every time there was no guarantee of success. Bigger things and externalities — also referred to as “timing” — would have an encompassing impact on probability of startup success.
What was Paul’s solution to this problem? Creating an accelerator that would invest a smaller amount of money in a larger amount of startups at an earlier stage. Y-Combinator would use three months to vet their startups, but essentially it was spray and pray. This strategy has since been adopted all over the world and today there are hundreds of accelerators in different formats doing variations of the same simple idea. Seed-DB currently lists 234 programs worldwide — check their summarized accelerator stats here.
From the perspective of the startup investors, this model would partially solve the problem of the power law that high-risk / high-return startups are subject to. With hundreds of investments, the chances of hitting it big with winners like Airbnb, DropBox or Stripe (to name a few that came through YC) were much bigger. Furthermore accelerators had the added benefit of helping startups become more connected and more prone to helping each other in the early phases.
However, from the entrepreneur’s perspective, the accelerators hadn’t solved the problem of the power law. Founders still “only” own a stake in their company and are very rarely involved in other projects simultaneously. Hence they would still be subject to the above statistics — most would fail or barely achieve any success, while the fortunate few would end up with more money than they could ever spend.
For an entrepreneur, having a large stake in just one startup (their own) is akin to having a single lottery ticket — a portfolio with no diversification. Founders adopt this risky position to communicate their certainty of success to investors in order to raise money.
What are the consequences of adopting this risky position? Founders will most likely fail their first couple of startups. The underlying statistics should tell them that they would need to start five seed-funded startups for just a 67% chance of one exit — an almost impossible endeavor for even the most persistent (assuming there is a 20% chance each time and excluding a potential learning curve it calculates to: 1 — (4/5)^5) = 0,672 ≈ 67%).
However, entrepreneurs are subject to the possibility effect, wherein they overestimate their chance of success and hence go forward with this risky business and most likely fail. Since failing will negatively impact a founders perception, most end up abandoning entrepreneurship before they start enough companies to beat the odds.
There is a relatively simple solution to this problem. The idea at first may seem unconventional, but would easily work if adopted by the right people, which is why I’m asking Paul Graham, Sam Altman, David Cohen & Dave McClure to consider the following:
Your accelerator should take a larger share in each of the startups and, in return, give them not only capital but also a stake in all the other startups in their batch.
I’m also asking founders to give away equity in their companies and they will probably ask: “why should my startup give away equity to the other startups? My startup is clearly going to do better than average and become a big exit.”
What if your accelerator had taken an additional 7% share of your startup and, in return, you had gotten 0.1% share of 70 startups with the same profile and quality?
If you were in the same batch with Airbnb you would now own a 0.1% share of a $24B company and your stake would be worth around $24M today (it would be less than 0.1% because of dilution, but this is just an example to make a point). If you were a founder of Airbnb, you’ve still become a billionaire from nothing and the difference in your lifestyle would be negligible. On the contrary, you would be cherished by fellow entrepreneurs.
Accelerators are the ideal candidate for this model, because they choose startups (and indeed founders) that are of a relatively equal quality. Hence founders will feel comfortable and justified in sharing equity with other startups. Continuing in this vein, it could also be argued that it will decrease negative competition and increase camaraderie and the sense of helping other companies to succeed instead of seeing them as a block to their funding at demo day.
Furthermore, it has the added benefit of encouraging collaboration between startups, increasing everybody’s chances of success. Failure with your startup won’t be nearly as financially devastating, encouraging people to continue building new companies and beating the odds.
I hope Y-Combinator, Techstars, 500Startups and all the other accelerators will consider my proposal and at least present this idea to their startups to see if they also agree with the concept. I believe the whole industry would benefit greatly from having more of a share in each other.
If investors are able to minimise their risk, why not those on the front-line pouring in their blood, sweat and tears too? Fundamentally, the lower the risk and higher the reward, the higher the talent pool and the greater the resultant businesses. It’s win-win as far as I’m concerned.
To organizers of accelerators: be the first to implement this concept and you will have a great incentive for startups to join your accelerator.
To founders who are already in an accelerator or thinking of joining one: tell your organizers about this idea and there is a chance it will actually happen.
If none of the accelerators have done this within the next 6 month I will consider doing a crowdfunding campaign to have a lawyer set up a framework and a fund for this purpose.