What is a commitment device, and why does it work?

You can argue with yourself forever. Here's why willpower tends to fail right when you need it, and what works instead.

When you try to make yourself sit down and work or study, there are two of you in the argument: you right now and you tomorrow. Right-now you wants a break. Tomorrow you wants it all done already. And the decision always gets made right now, which is why "I'll do it tomorrow" wins so often.

Advice about willpower usually boils down to building more of it. There's a more reliable approach: instead of winning the argument every single time, set things up in advance so that backing out costs more than following through.

Why willpower isn't enough

Economists call it present bias. The effort is felt immediately, the payoff arrives later, and in the moment of choosing, a future payoff weighs less. Tonight it's half an hour of work against one more episode, and the work loses. Tomorrow the same scene replays.

That's why reminders do so little. A reminder tells you about the task, but it doesn't change the one thing that matters: putting it off is still the better deal. A to-do list with no consequences is a list of regrets with checkboxes.

Commitment device meaning

The fix is older than economics. Odysseus knew he would hear the sirens and want to go to them, so he had himself tied to the mast in advance. He didn't count on willpower at the critical moment. He removed the easy way out ahead of time.

That's a commitment device: a choice you make now to limit your options later. It doesn't make your future self any stronger. It changes the prices your future self will face.

Commitment device examples from research

The best known study on the subject is even named after Odysseus: Tying Odysseus to the Mast. Bank clients in the Philippines were offered an account they couldn't withdraw from until they reached a date or an amount they had picked themselves. The account paid no extra interest. All it offered was the restriction. A year later, the people who opened it had saved 81% more than the control group.

The second is Put Your Money Where Your Butt Is. Smokers were offered a contract: deposit money for six months, then take a nicotine test. Pass and the money comes back. Fail and it goes to charity. The effect was modest: people offered the contract were about 3 percentage points more likely to pass. But it was still there a year later, in a surprise test. For a habit as stubborn as smoking, three durable points is a lot.

Services like stickK and Beeminder run on the same idea: you agree in advance that money leaves if the goal isn't met. And plenty of everyday habits are commitment devices too:

• leaving your credit card at home before walking into a store where you tend to overspend;
• keeping your phone in another room while you work;
• telling a friend you'll send them a draft by Friday;
• booking a class instead of planning to "go to the gym sometime this week".

Why paying upfront doesn't work

A gym membership looks like a commitment device. You've already paid, after all. Economists Stefano DellaVigna and Ulrike Malmendier tested that with data from three US health clubs. Members on contracts of more than $70 a month went an average of 4.3 times a month, which works out to more than $17 per visit. They could have paid $10 a visit with a ten-visit pass.

The catch is simple. The membership money is already gone, so skipping a workout costs nothing on the evening you're deciding whether to go. Only something that makes skipping expensive at the moment of choice actually works. That's why StarDo doesn't ask for anything upfront: the cost shows up exactly when a task doesn't get done.

Why money beats willpower

Kahneman and Tversky showed something simple: a loss hurts more than a gain of the same size feels good. Losing ten dollars stings more than finding ten dollars pleases.

Once a promise has a price, the moment of choice gets a second weight on the scale, and that weight is felt now rather than someday. That's exactly what was missing: the cost of putting things off moves out of the future and into the present. The amount doesn't have to be big. It has to be noticeable.

How StarDo turns this into a to-do list

• You pay nothing upfront. Finish on time and you pay nothing at all, and your Stars stay with you. Miss the deadline and you pay what you promised.
• You choose the amount, starting at 5 Telegram Stars.
• 48 hours to pay. Once a deadline passes, you still have two days to pay and keep your honor rating intact.
• A promise kept becomes a star. Everyone gets their own sky, and it grows by exactly as much as you actually finish.
• Nothing is left to chance. Only you affect the outcome, and there's no way to get back more than you put in. The best possible outcome is paying nothing.

How to use a commitment device without it backfiring

A good amount is one you don't want to pay but wouldn't be afraid to. If the thought of paying makes your stomach drop, the number is too high: fear works worse than mild annoyance.

Start at 5 to 10 Stars and raise it once the mechanism feels familiar. Different tasks deserve different amounts, and not by importance. What matters is how badly you'd like to wriggle out of this particular one. "Read the chapter before the seminar" might be worth 10 Stars, while the report you've been dodging for three weeks deserves more.

Attach the promise to a specific action with a deadline, not a vague intention. "Solve ten problems from chapter two by Thursday" is easy to check. "Study more" isn't.

One more rule: if the same task falls through three times in a row, the amount isn't the problem. The task is too big and needs to be cut into pieces with an obvious first step. Our guide on how to stop procrastinating covers how.

What a commitment device won't do

It won't do the work for you, and it won't replace a plan. It won't rescue a task you don't actually need either: if something has been hanging around for months and brings nothing but dread, deleting it is more honest than paying for it. And if you haven't wanted to do anything at all for weeks, that's a reason to talk to a doctor, not to raise the amount.

One note about the research itself. The most famous study on self-imposed deadlines (Ariely and Wertenbroch, 2002) was retracted in 2026: a replication failed, and questions were raised about the data. For years it traveled through articles about procrastination, including ones about apps like ours. We don't cite it. Everything above is what held up.