The Explore/Exploit Tradeoff: When should I Leap?

I recently got a question from a reader asking about if and when they should move from working on their current business to a new venture.

I’ve gotten a bunch of versions of this question over the years (often from myself): when should I leave this job for another job? When should I leave this job to do my own thing? When should I leave my business to take a job or start another business?

In short, when should I leap? I think leaping is a good metaphor here because that sort of out of control sensation when you are in the air and hope you can stick the landing is the same as when you make a big shift. It’s exciting and scary at the same time. I think it’s psychologically, and, if you will allow me to sound like a cliche, spiritually important to move towards new challenges, but timing matters.

So, how do you know when to leap?

The first and most important thing to know about leaping is that there is no right answer that is knowable a priori.

You are making a decision in which the future forks and both paths are uncertain, so if you wait for the perfect answer then you’ll be waiting forever. One way of thinking about it that has been helpful for me: the explore/exploit tradeoff (also called the multi-armed bandit problem).

Here’s a short version of the problem: imagine walking into a casino and deciding to play the slot machines.

There’s a row of ten slot machines, each of which has a different probability distribution of rewards when you pull the lever. Some machines pay more than the other machines, but you’re not sure which machine has the highest return.

If you knew the best machine in advance, you would obviously just pull that lever all day long, but you don’t know.

The only way to find out is to start pulling levers, pay close attention, and keep track of what works and what doesn’t.

Here’s where the tradeoff comes in: when you choose to pull a lever you haven’t pulled before, you get new information about that option. That information is valuable in finding the best paying slot machine.

But, it also has an opportunity cost: you’re not pulling the lever you currently think will give you the best return. There’s a risk that the lever you pull will return less than what you would’ve brought in pulling the lever that you currently think is the best one.

This new information is valuable, but it comes at a price, an opportunity cost.

So how do you know whether to pull the existing lever again or try a new one?

The details get complex, but the broad strokes of the answer are not that surprising: you should start doing some amount of exploration – it’s very unlikely that the first machine you pull will be the best paying one. Once you have done some amount of experimentation, you should focus more on exploiting. If you have done a lot of exploration of the different machines, it’s less likely you’ll make a significant breakthrough in discovering the best machine.

The multi-armed bandit problem is pretty easy to parallel with how careers work. You walk into a big wide world with all sorts of industries and jobs and opportunities and you start working on one but constantly find yourself wondering “Is this the highest leverage thing to work on?”

The only way to know is to:

Observe other people playing those games and see how it goes for them.
Try other things yourself and see how it goes.

If you see someone with a similar skillset doing something that is working a lot better than what you are doing, you should probably start exploring it a bit. This is especially true early in your career. In general, I would say most people focus on exploiting too soon.

“Learn in your 20s and earn in your 30s” is pretty decent advice. If you are going to be working for 40-50 years, spending at least the first ten in exploration mode makes good sense. So it’s worth picking jobs early on that pay you enough, but give you more flexibility to explore new areas. New areas tend to have the most opportunity because they are growing and there are less well known and so the competition is lower.

Everyone knows that you can make a lot of money in investment banking or management consulting and so those roles have lots of competition. If you told someone 5 years ago that you were learning to code bitcoin or blockchain applications then you would have sounded like an idiot… right up until you sounded like a genius.

At the same time, some people stay in explore mode for too long. VC Mark Suster relates the story of an early 40’s friend that went to Harvard Business School and had worked at 3 successful startups:

Whenever he calls me he must think I’m a broken record. I always say, “Dude (I live in SoCal now!) — it’s time to EARN. Stop dicking around with another number 2 job (he always gets offered the number 2 job). It’s time for you to be in the driver’s seat. Either start a company or go somewhere where they need a CEO.”

If you really want to earn you need to be in the top 3–4 in the company. Best to be a founder. Very few people can do this. It’s a rare skill. Be realistic about your skills, background and ideas.

Said another way, he was telling his friend that it was time to stop exploring and starting exploiting.

I am primarily casting this in economic terms, but it’s equally applicable to what other qualities a job provides. Explore which roles provide you the most overall “wealth” in the broad sense: afternoons off to go to the gym or play with your kids, then at some point you have to move into exploit.

If you’re interested in more, here’s some places to start reading more about the explore/exploit tradeoff:

Last Updated on September 27, 2020 by Taylor Pearson

Essays

Greatest Hits (Sorted by Theme)