Fooled by Randomness.
I just finished reading Nassim Taleb’s “Fooled by Randomness“. It’s a great read if you like your thought processes to be challenged and are willing to slog through his slightly pedantic style. Having already read his follow-on book, “The Black Swan,” I was familiar with his style; I also knew that I would find theories and advice that are applicable to our industry. Taleb is focused on the options trading market, so much of his advice and counsel is geared towards traders. However, if you read it carefully there is a ton of good advice for those of us in the private equity markets, and for founders of middle-market companies. I certainly would not try to summarize all of Taleb’s theories here—after all, you should read the book—yet there are clearly a couple of key points for our industry.
Taleb loves using Monte Carlo analysis to generate a spectrum of possible outcomes in order to minimize his overall risk. In some cases, however, he does so to FIND risks that he can exploit to his advantage. If you will grant me a little oversimplification here, I can summarize Taleb’s view on this topic with an example from his book:
Consider a janitor that plays the lottery as compared to a dentist that doesn’t. In a hypothetical infinite universe, some janitors will hit the jackpot. However, on average, most janitors won’t. Dentists on the other hand will never make millions of dollars per year — at least not on average (there may be a FEW dentists making big in Hollywood). They’ll likely average $500,000 a year, year-in and year-out. In Taleb’s world, this would tell you that the range of possible outcomes for dentists is very low risk, and not much affected by randomness. The janitors, on the other hand, are successful a much smaller percent of the time. And their individual success—their jackpot—is entirely random.
Middle market software companies are the dentists in this example; venture start-ups tend to be janitors in our world. If you founded your company around serving a very specialized niche market, and you eliminate cash burn by getting early customers to pay for the development of your product, and you develop your market slowly, then you are a middle-market software dentist.
While you might still fail, the odds begin to work in your favor. You aren’t much affected by Taleb’s randomness.
On the other hand, if you are a venture-backed start-up, using venture dollars to build product, burning cash, with the goal of finding customers, then you are the janitor in Taleb’s example. Faced with a high range of outcomes, you are much more likely to be successful based on some random event. Why did Facebook buy WhatsApp? Were they that much better than the competition? I’d bet that some analyst will happily provide you with some expensive hindsight analysis telling you why. I’d wager far more that Taleb would argue that there is more than a little randomness as to which real-time messaging platform Facebook purchased. Over time this will not bode well for the mass of other real-time messaging vendors out there. Most will likely fall into the abyss. And that’s okay. Venture investors plan for lots of zeroes, a few singles and a homerun or two.
It’s Taleb’s janitor model.
Any long-term venture investor that is successful is probably smart and well connected. They are also lucky. Admittedly I am paraphrasing a bit here, because venture returns do not typically follow a Taleb distribution. Under the Taleb distribution you have a high probability of small gains and a low probability of a enormous loss. Venture investing essentially follows a reverse Taleb distribution from an individual investment perspective. A small probability of a huge gain (Facebook buys you for $1 billion) and the high probably of a relatively small loss (your $10 million investment in the competitor that DIDN’T get acquired).
We’ve gotten lucky in our market too, but by and large we would be Taleb’s dentists. No company that we own is going to sell for a billion dollars…ever! But we aren’t going to get any zeroes either. The range of outcomes in our market is much narrower, much less affected by random events. I say “less” affected, because random events affect our companies too. However, our sorts of companies tend to be more immune to the negative affects of randomness due to their inherent structure.
First, they are cash-flow positive businesses with $3+ million of annual EBITDA and little, or reasonable, debt. Since they are able to fund their business activities through their own profits, they are not dependent upon getting money from an outside investor on an ongoing basis. Which inures them against the random possibility that their investor is short of money when they need some.
They tend to invest behind their market growth as opposed to spending their way to grow. This means that they are likely to be almost constantly understaffed and short of other resources (servers, storage, etc.). This also means that their budget will be behind their growth, so random factors contributing to a down year have a much less dramatic impact.
They have a single majority owner, us. So they are not subject to the randomness caused by having several different investors with potentially conflicting investment agendas.
For example, I’m at the end of my fund, and I’d like to sell. You’re at the beginning of your fund and you’d like to hold the investment. Or vice-versa, I’m at the end of my fund and NEED to hold the investment while you are at the beginning of your fund and NEED an exit.
No single customer represents more than 10% of their revenue and EBITDA. So they aren’t destroyed by the loss of a single large customer. While the slowdown of an entire market segment affects them, the combination of a large base of customers with strong cashflows and manageable debt allows them to weather this kind of storm.
Technological disruption can be a huge random event for software companies. Apple releases iOS 8 or Microsoft discontinues Windows XP, or Google buys up your competitor. These events can cripple a start-up (except for Windows XP, there aren’t any startups building in the last 10 years building on XP). Customers of focused vertical software companies tend to be more patient with this type of disruption. They often move slower than the company itself when it comes to technology change. A startup that is built to rely heavily on Facebook’s platform (Zynga comes to mind) is heavily affected by changes in Facebook’s technology ecosystem.
I’m not advocating for one business model over another (venture vs. private equity). They are both valid approaches to investing. Returns for both have been roughly the same over like periods, but the beta on venture has been higher. It’s instructive for founders and entrepreneurs to understand whom they are partnering with. If you’re a company with $3MM – $8MM of EBITDA then you’re better offering partnering with investors that understand your business model and risk environment. We’ll better understand your growth curve and investment cycle than either venture investors or large buyout shops.
We aren’t fooled by randomness, we’re built to withstand it.