Nothing Has To Go Wrong For It To Be Wrong
A 1976 bird study explains why staying in something that's still working can be the most expensive mistake you make.
Most people wait for a signal that never comes. They think quitting requires a reason — a loss, a blow-up, a moment where the thing clearly turns bad. It doesn’t. There’s a mathematical model for exactly when to walk away, and it comes from watching birds eat berries.
The berry bush problem
In 1976, biologist Eric Charnov formalised something foragers already knew instinctively: animals don’t strip a food patch bare before moving on. A bird lands on a berry bush, eats the easy fruit near the surface, and leaves — while the bush is still half full, while the berries left are still good to eat.
Charnov called it the marginal value theorem. The logic is simple once you see it. Every patch starts rich and then slows. The berries get harder to find, smaller, and further apart. The bush hasn’t failed — it’s just decelerating. And the animal’s real question was never “Is this bush still giving me something?” It was “Is this bush still giving me more than I’d get somewhere else, once I account for the cost of moving?”
The moment the answer flips, the bird leaves. Not when the bush is empty. Long before that.
Working and working well aren’t the same thing
This is where it gets uncomfortable for anyone running a business, a deal pipeline, a portfolio, or a career. Most of us were trained to finish what we start and to read “still producing something” as “still worth doing.” That’s the wrong test.
A client, a listing, a market, a role, a strategy — all of them behave like berry bushes. They start rich. Early effort returns fast. Then the curve bends. You’re still getting something out of it. It’s just costing you more time and attention per unit of return than it used to, and there’s a version of you, or your business, that could be getting a better rate somewhere else.
The trap isn’t staying in something bad. Nobody needs a model to tell them to leave a bad deal. The trap is staying in something that’s merely no longer the best use of your time — because it still feels like it’s working. Slower returns don’t announce themselves as failure. They just quietly become the wrong home for your effort.
What the model actually asks you to track
Marginal value theorem thinking isn’t about optimism or pessimism. It’s a rate comparison. Two numbers:
What is this specific thing returning me right now, per unit of time or capital I put into it — not what it returned me at the start, not what it could theoretically return again?
What’s the average rate available to me elsewhere, once I price in the cost of switching — the deal I’m not chasing, the client I’m not prospecting, the capital sitting in something else while I hold this?
When the first number drops below the second, the model says go. Everything after that point is sentiment, sunk cost, or fear of admitting the bush changed.
The hard part is that this requires you to actually track the rate, not the total. A deal pipeline that’s “still bringing in commission” can be quietly bleeding your best hours if you never measure velocity, only volume. A portfolio position that’s “still up” can be dead capital if the market’s diminishing-returns curve has already bent and you haven’t clocked it.
The uncomfortable version of this for operators
Nobody builds a business, buys a stock, or takes a job intending to overstay. It happens because leaving requires admitting nothing went wrong — which feels irrational. It’s much easier to justify quitting a disaster than quitting something that’s merely fine.
But the fine is the actual cost. Fine is the bush still worth eating from while a better one goes unvisited. The birds don’t wait for the bush to fail. They don’t need permission from a crisis. They just run the rate comparison, quietly, constantly, and move when the number says move.
That’s the whole discipline. Not grit. Not loyalty to the thing you started. Just an honest, ongoing read on whether this is still the best use of what you have.
The bush doesn’t have to go bad for it to be time to leave.