By four in the afternoon, decisions you’d have made carefully at nine get made carelessly, or not at all. Most people who run something recognise this immediately.
The usual explanation for it is wrong, or at least much shakier than it’s presented. Which matters, because the wrong explanation leads to the wrong fixes.
What you’ve probably been told
The popular account is that willpower works like a fuel tank. Every decision draws down a finite reserve of self-control, the tank empties as the day goes on, and eventually you’re running on fumes. It’s why the story about tech executives wearing identical clothing every day became so widely repeated — conserve the fuel for decisions that matter.
It’s a tidy model. The research underneath it has not held up.
The original experiments, from the late 1990s, found that people who exercised self-control on one task gave up sooner on an unrelated task afterwards. A 2010 meta-analysis reported a large, apparently robust effect, and that’s the figure that entered popular culture.
Then it was tested properly. Re-analysis correcting for publication bias produced an estimate close to zero. A 2016 registered replication across 23 laboratories and more than 2,000 participants — with the method agreed and published before any data was collected — found no reliable effect. A later multi-lab attempt found effectively nothing either.
The researchers behind the original work dispute this, arguing the replications didn’t successfully induce the effect they were testing. That argument continues. But the honest summary is that the fuel-tank model is not established science, and anyone presenting it as settled is overstating it.
What does hold up
Here’s the part I find genuinely interesting: the real-world evidence is stronger than the laboratory evidence.
Studies of clinicians working through a session of appointments find that prescribing patterns shift measurably as the session goes on — later appointments produce more of the quick, default option and fewer of the ones requiring explanation. Similar drift shows up in other settings where people make many comparable judgements in sequence.
So something real is happening. It just may not be a depleting energy store. More plausible candidates: attention narrows, motivation to expend effort declines, and the appeal of whatever is easiest rises. Practically, the difference matters less than you’d think — but it does change which fixes make sense.
Why the distinction changes the advice
If willpower is fuel, the strategy is conservation: minimise trivial decisions to save capacity for important ones. Hence the identical shirts.
If what’s actually happening is drift toward the easy option under accumulating load, the strategy is different. You don’t conserve capacity. You change when and how the decisions happen.
Which produces better and more practical advice:
Put consequential decisions early. Not because you have more fuel at nine, but because you’re more willing to do the effortful version. The decision that needs you to think rather than default belongs in the first part of your day.
Don’t decide anything that matters at the end of a long stretch. The specific risk isn’t a bad decision — it’s the easy one. Renewing the contract rather than renegotiating it. Saying yes because saying no needs a conversation.
Break sequences of similar judgements. The drift shows up most in long runs of comparable decisions: interviewing six candidates, reviewing twenty invoices, working through a queue. Split it, or accept that the last ones get less than the first.
Convert recurring decisions into rules. Not to save willpower — to remove the decision. “We don’t take clients under X” made once is a policy. Re-litigated each time, it’s a decision you’ll eventually get wrong when tired.
Notice the drift rather than resisting it. If you catch yourself about to approve something at five because arguing feels like too much, that recognition is the whole intervention. Move it to tomorrow morning.
The owner-specific version
One thing worth adding, because it’s the part that applies specifically to running your own business.
Most of your decisions are unforced. Nobody assigns them. They arrive because you’re the one who notices things, and there’s no natural boundary on how many you take on.
An employee’s decision load is largely set by their role. Yours is set by what you happen to look at — which means the volume is, to a considerable extent, self-inflicted and therefore adjustable. Every process you document, every spending limit you set, every decision you push to someone else is a decision that doesn’t reach you at all.
That’s a structural fix rather than a personal one, and it does far more than any amount of managing your own energy.
The short version
Your judgement is worse late in the day, and the evidence for that comes from real workplaces rather than the laboratory studies everyone cites.
The fix isn’t to conserve willpower like a resource. It’s to put the decisions that need real thought where you’re still willing to do it, to turn repeated choices into standing rules, and to reduce how many reach you in the first place.
And to be suspicious of yourself at five o’clock — not because you’ll decide badly, but because you’ll decide easily.
More on what to do when you can’t decide at all and how to stop decisions reaching you.
