I just got done reading Jeff Bezos’ annual letter to shareholders, which is outstanding as it always it. As I finished it, I spent a few minutes thinking about it. He references Amazon’s style of “portfolio management”. He doesn’t call it that of course, but this passage got me thinking about it. Since I wrote a post earlier in the week about portfolio management, I thought using Bezos’ letter would allow me to expand on a few other random thoughts. But here is just one clip from many valuable nuggets that are in the letter:
Bezos has always gone for the home run ball at Amazon, and it’s worked out tremendously for him and for shareholders. Would this type of swinging for the fences work in investing?
I’ve always preferred trying to go for the easy bets in investing. Berkshire Hathaway is an easy bet. The problem though (or maybe it’s not a problem, but the reality) is that the easy bets rarely are the bets that become massive winners. Occasionally they do—Peter Lynch talked about how Walmart’s business model was already very well-known to investors in the mid 1980’s and it had already carved out significant advantages over the dominant incumbent, Sears. You could have bought Walmart years after it had already proven itself to be a dominant retailer but also when it still had a bright future and long runway ahead of it.
So sometimes the obvious bets can be huge winners. But this is usually much easier in hindsight. After all, Buffett himself couldn't quite pull the trigger on Walmart in the mid 1980's, a decision he would regret for decades. At the annual meeting in 2004, he mentioned how after nibbling at a few shares, he let it go after refusing to pay up:
[drizzle]“We bought a little and it moved up a little and I thought maybe it will come back a bit. That thumbsucking has cost us in the current area of $10 billion.”
So sometimes obvious bets can be huge winners. But many times, the most prolific results in business come from bets that are far from sure. Jeff Bezos has always had a so-called moonshot type approach to capital allocation. The idea is simple: there will be many failures, but no single failure will put a dent in Amazon's armor, and if one of the experiments works, it can return many, many multiples of the initial investment and become a meaningful needle-mover in terms of overall revenue.
Amazon Web Services (AWS) was one such experiment that famously became a massive winner, set to do $10 billion of business this year, and getting to that level faster than Amazon itself did. The Fire phone was the opposite–it flopped. But the beauty of the failures at a firm like Amazon is that while they are maybe a little embarrassing at times, they are a mere blip on the radar. No one notices or cares about the Amazon phone. If AWS had failed in 2005, no one today would notice, remember, or care.
So this type of low probability, high payoff approach to business has paid huge dividends for Amazon. I think many businesses exist because of the success of a moonshot idea. Mark Zuckerberg probably could not have comprehended what he was creating in his dorm room in the fall of 2004. Mohnish Pabrai has talked about how Bill Gates made a bet when he founded Microsoft that had basically no downside–something like $40,000 is the total amount of capital that ever went into the firm.
“Moonshot” Strategy is Aided by Recurring Cash Flow
One reason why I think this approach works for businesses and not necessarily in portfolio management is simply due to the risk/reward dynamic of these bets. I think a lot of these bets that Google and Amazon are making have very little downside relative to the overall enterprise. Most stocks that have 5 to 1 upside also have a significant amount of downside.
I think lost dollars are usually much more difficult to replace in investing than they are in business, partially because businesses usually produce recurring cash flow. Portfolios have a finite amount of cash that needs to be allocated to investment ideas. Portfolios can produce profits from winning investments, and then these profits can get allocated to other investment ideas, but there is no recurring cash flow coming in (other than dividends).
Employees, Ideas, and Human Capital
Not only do businesses have recurring cash flow, they also have human capital, which can produce great ideas that can become massive winners. Like Zuckerberg in his dorm room, Facebook didn't start because of huge amounts of capital, it started because of a really good idea and the successful deployment of human capital (talented, smart, motivated people working on that good idea). Eventually, the business required some actual capital, but only after the idea combined with human capital had already catapulted the company into a valuation worth many millions of dollars.
There was essentially no financial risk to starting Facebook. If it didn't work, Zuckerberg and his friends would have done just fine—we would have most likely never have heard of them, but they'd all be doing fine.
If AWS flopped, it's likely we would have never noticed. There would be minor costs and human capital would be redeployed elsewhere, but for the most part, Amazon would exist as it does today—dominating the online retail world.
Google will still be making billions of dollars 10 years from now if they never make a dime from self-driving cars.
So I think this type of capital allocation approach works well with a corporate culture like Amazon's. Bezos himself calls his company “inventive”. They like to experiment. They like to make a lot of bets. And they swing for the fences. But the cost of striking out on any of these bets is tiny. And you could argue that any human capital wasted on a bad idea wasn't actually wasted. Amazon—like many people—probably learns a ton from failed bets. You could argue that these failures actually have a negative cost on balance—they do cost some capital, but this loss that shows up on the income statement (which again, is very small) ends up creating value somewhere else down the line due to increased knowledge and productive redeployment of human capital.
So I think there are advantages to this type of “moonshot bet” approach that works well within the confines of a business like Amazon or Google, but might not work as well within the confines of an investment portfolio. This isn't always the case—I recently watched the Big Short (great movie, but not as good as the book) and the Cornwall Capital guys used these types of long-shot bets to great success. They used options (which inherently have this type of capped downside, unlimited upside risk/reward) and they turned $30,000 into $80 million. But I think this would be considered an exception, not the rule.
I think most investors have a tendency to arbitrarily tilt the odds of success (or the amount of the payoff) too much in their favor with these types of long-shot bets. They might think a situation has 6 to 1 upside potential when it only