Knowing When to Leave: Deciding Before You Have To

Most CEOs who step down choose to. When researchers at Russell Reynolds interviewed thirty departing chief executives for Harvard Business Review, more than eight in ten had initiated their own exit. The reasons are the interesting part. Not failure, mostly. Something closer to a sense that the timing was right. Pressed for specifics, they said it felt right.

These are people who plan everything. And they leave on instinct.

An exit strategy is the thing that closes that gap. You name, in advance, the circumstances under which leaving would be the right choice. That's all it is. Not a crisis plan, not a letter in a drawer — a decision made while you can still see clearly, so you're not making it later under pressure, with too little information and too much feeling.

There are two versions of it, and only one gets talked about.

The first is the one you'd expect. Something you need in order to do good work isn't there anymore. The second is quieter: you've done what you came to do. The thing you were brought in to build is built. The team runs without you. Nothing has gone wrong at all.

Both are exits. The second one slips past people, because nothing hurts.

So where do you start? With what has to be true for you to work well here. Not what you'd prefer — what has to be true. Autonomy over certain decisions, maybe. A working relationship with one particular person. Room to keep learning. The list is usually shorter than people expect.

Then the harder question: what would tell you you're finished? Most roles don't come with a definition of done. If you were hired for something specific, you can often name it. If not, ask what would have to exist for your staying to add less than it once did.

Now turn each answer into something that could actually happen.

This is where it usually falls apart. If things get bad enough. If I stop growing. Those sound like thresholds, but you can't cross them. There's no day you could look up and say, that happened.

A condition you can use is one someone else could recognize. They don't know you, they're just watching your situation, and they could tell you whether it's been met.

A condition you can use is one someone else could recognize. They don't know you, they're just watching your situation, and they could tell you whether the condition has been met.

The condition doesn't have to be a number. Chip and Dan Heath call these tripwires, and the strongest ones are often patterns rather than metrics.The fourth time an agreement gets broken after you've raised it. The person who championed your work leaving and the role staying empty. Three meetings in a row where you realize the decision was already made. The project you were hired to lead wrapping, and no one naming what comes next.

Write them down these patterns that you are noticing.

Then sort them, because they're not all the same. Some mean leave. Others just mean start paying attention — have a conversation or two, get clearer about what you'd actually want if you left.

Which brings it back to those chief executives. The instinct they described probably wasn't wrong. It was just the only instrument they had.

Naming your conditions doesn't replace that feeling. It doesn't tell you when to leave. What it does is give you something to check your instinct against — a way to ask whether what you're sensing is actually there.

Sources: "The Vital Role of the Outgoing CEO," Harvard Business Review, July–August 2024. Chip Heath and Dan Heath, Decisive: How to Make Better Choices in Life and Work (2013).