The should-I-build-this formula
What product managers can learn from the Kelly Criterion and gambling
Watch the video for this important lesson on risk. Transcript below.
Transcript for the video above:
It doesn’t matter if you're building your own product or just launching a feature, before you build anything, you need to know this man’s formula for risk. The should-I-build-this formula.
Naval Ravikant is the co-founder of AngelList and has invested, or gambled as I call it, in over 200 companies.
But there’s a big difference between Naval and your average gambler or angel investor. He wins. He has over 70 exits, including more than 10 unicorns. An incredible record.
Today, we're talking about the Kelly Criterion, what is it Naval?
(Audio clip starting at 0:03) “The Kelly criterion is a popularized mathematical formulation of a simple concept. The simple concept is: Don’t risk everything. … Don’t bet everything on one big gamble. Be careful how much you bet each time so you don’t lose the whole kitty.”
Here’s an example (see video above) of the Kelly Criterion formula for gambling. The simple concept is important here, not the actual formula.
The result tells you what percent of your money you should bet based on what you’ll win and the probability of winning. Pretty simple.
So, this example (see video above) shows if you have a 60% chance of winning, you should only bet 20% of your pot.
Why not bet the whole kitty?
Any gambler will tell you, no, no, no, you might lose a couple hands before you win. So you can’t bet the whole kitty, you’ll go broke! Pretty simple.
Here’s where it becomes useful for angel investors. You might say, Kelly, can I invest 100 grand in a startup that has a 2% chance of returning a billion dollars? I’m using a 2% probability just as an example because the chances of a startup becoming a unicorn are extraordinarily low. Ms. Kelly Criterion will say, yes, just make sure you’re only betting about 2% of your total fund, meaning you need at least 5 million dollars to make that $100,000 bet.
Gambling and startup investments are obvious places to look for risk because many have extraordinarily low probabilities of success.
Well, any product launch is also a gamble and a lot of people making these bets don’t assess risk in the way that your average gambler might.
Some product leaders are basically sitting down at a roulette table and betting the entire company on red. “But have I considered the risk?”
Sidenote: The Kelly Criterion says to only risk negative 5% of your money on red in roulette, meaning don’t risk anything if you’re not the casino.
Unless you factor in fun and free drinks, then I would say risk only a few percent of your pot on each roll, knowing the house always wins in the end.
As a product manager, your job is to figure out what your team should build – and risk is an important criterion for you to consider when deciding what to build.
The gist is this, when people don’t understand risks, they risk too much, the whole kitty. They bet everything on one big gamble that can wipe them out completely.
The Kelly Criterion simply tells us to make any decent-sized bet, you need to reduce your probability of a loss. You don’t need to use the formula, just acknowledge the risks and take steps to mitigate them.
The Kelly Criterion shows that most people don’t understand the risk they’re taking. Marty Cagan kind of says the same thing, All these products get built without thinking about the risks, and that’s what leads to failure. And that’s why product managers should learn these four types of risks (see video above), how to de-risk them and reduce the probability of failure.
I’ll link to my video on discovery for more on this or you can read his book Inspired.
It doesn’t matter if you’re launching a huge product or a small feature, risk is risk. Someone is going to ask, “What steps did you take to ensure that people want this? That they can figure out how to use it? That the engineers can build it and the company can support it? None? Nothing? You bet it all on red?”
Don’t risk your reputation on a bad bet. De-risk before making any product bets, and don’t bet the whole kitty. You got this!