The 37% Rule: How to Decide When to Stop Wondering and Start Deciding

Earlier this year, I read Algorithms to Live By, a book that explains how to use insights from computer science in daily life. One of the rules is the 37% rule. It’s an important rule because it’s broadly applicable. But I had forgotten about it until I listened to the author on the Software Engineering Daily podcast.

The 37% rule says that if you have a decision to make, you should spend 37% of the amount of time you have. This is called exploration period. Seek advice, learn about demands of the role, and meet some candidates.

After you’ve explored, you transition to exploit. Exploit in this context doesn’t mean to take advantage of or misuse, but to seize the opportunity and capitalize on your knowledge.

How do you do that? You pick the best candidate you’ve met so far or meet very shortly thereafter.

For example, assume you need to hire a salesperson or a marketer. If you give yourself 90 days, spend the first 33 days and then pick the best choice immediately after that. You’ll find the optimal candidate 37% of the time, which is an odd coincidence. This strategy, however, is the optimal one.

Computer scientists call this the multi-armed bandit problem. A one-armed bandit is another name for a slot machine. If you go to a casino and want to optimize your winnings, how do you pick the right slot machine? You spend 37% of your time figuring out which machine is lucky and then spend the rest of the time with that machine.

Once you look for this pattern, you’ll see it many places. Children try many different things but elderly people focus on the things they love. Each are on opposite ends of this continuum - called the explore vs exploit tradeoff.

You visit a restaurant frequently because you love it. How many dishes should you try before settling on a favorite? 37%.

We see it in startups all the time. For example, an early-stage startup spends some amount of its life figuring out the right sales strategy. It might take a few months or a few years to identify the correct playbook, ideal customer profile, and product to sell. Then the business transitions to scaling. The first period is the explore period and the second the exploit period.