A commitment device is a restriction you impose on yourself in advance, deliberately making a future choice harder or more expensive, because you expect to want the wrong thing at the time. Deleting an app, paying for a gym class you’ll lose money on, giving your phone to a housemate and a focus timer that fails your session if you leave it are all versions of the same move.
They work, modestly and not for everyone. Here’s the evidence, including one famous study that has since been retracted.
Why would you deliberately limit your own options? #
Because the you who plans and the you who acts want different things. At 9pm on Sunday, a study block at 8am on Monday sounds reasonable. At 8am on Monday, another hour of sleep sounds better. Economists call this present bias: the same choice looks different depending on how close it is.
A commitment device bridges the gap by letting the planning version of you take options away from the acting version. The classic illustration is Ulysses ordering his crew to tie him to the mast before they sailed past the Sirens, with instructions to ignore anything he said afterwards. He didn’t trust his future preferences, so he removed his ability to act on them.
Does the research support commitment devices? #
Yes, though the effects are moderate and take-up is usually low.
Money at stake changes behaviour. In a field experiment in the Philippines, Xavier Giné, Dean Karlan and Jonathan Zinman offered smokers a voluntary savings account: deposit money for six months, then take a urine test. Pass and the money comes back; fail and it goes to charity. Eleven percent of smokers offered the product took it up, and those offered it were 3 percentage points more likely to pass the six-month test than the control group. The effect persisted in surprise tests at 12 months, which suggests real behaviour change rather than a one-off performance.
Restricting a pleasure to the task works too. Katherine Milkman, Julia Minson and Kevin Volpp tested “temptation bundling” by giving people access to page-turner audiobooks only at the gym. Gym visits rose 51% in the full treatment group and 29% in a lighter version, though the effect declined over the following months. Afterwards, 61% of participants chose to pay for gym-only access to the audiobooks, which is the interesting part: they knew the constraint helped and wanted to keep it.
Both results point the same way. Voluntary constraints move behaviour, the effects are real but modest, and most people don’t take them up when offered.
A word about the study you’ve probably read about #
The most-cited paper on self-imposed deadlines was Dan Ariely and Klaus Wertenbroch’s 2002 Psychological Science article “Procrastination, Deadlines, and Performance: Self-Control by Precommitment”, in which students who set their own spaced deadlines outperformed those who could hand everything in at the end.
It was retracted by Psychological Science in 2026. If you see that study cited as proof that self-imposed deadlines work, including in articles published before the retraction, treat the claim as unsupported by that source. The broader case for commitment devices doesn’t rest on it, which is why the two field experiments above are the ones worth remembering.
Types of commitment device #
| Type | Example | Strength | Risk |
|---|---|---|---|
| Hard, physical | Phone locked in another room or a timed box | Very strong | Can’t get to it in a genuine emergency |
| Hard, financial | Money forfeited if you miss a goal | Strong | Losses can sting more than the goal is worth |
| Soft, social | Telling someone your plan; a coworking session | Moderate | Only works if they’ll actually notice |
| Soft, digital | App blocker in a mode you can’t easily exit | Moderate | Usually defeatable if you’re determined |
| Soft, in-task | A timer that voids the session if you leave | Mild | Easy to ignore if nothing depends on it |
| Bundled | A treat you only get during the task | Mild to moderate | Fades over time |
The pattern in the research is that harder devices work better and get chosen less. A constraint severe enough to be effective is a constraint most people decline.
How do you use stakes for focus specifically? #
- Make the constraint proportionate. The aim is to make the wrong choice slightly annoying, not catastrophic. Big stakes create dread, and dread makes you avoid starting, which is the problem you were solving.
- Attach it to the behaviour, not the outcome. “I do a 45-minute session” is something you control. “I get an A” isn’t, and staking money on things outside your control mostly generates anxiety.
- Make it automatic. A constraint you have to re-impose each morning will lapse. Put the phone in the same drawer every day, or use a timer that applies the rule by default.
- Make the cost immediate. A consequence in six weeks doesn’t compete with a notification now. The most effective stakes bite within the session.
- Add a person if you can. Social stakes are cheaper and often stronger than financial ones. Someone expecting your work at 4pm does a lot of work. Body doubling is the low-effort version.
- Keep an escape hatch for real life. A device that punishes you for answering a genuine emergency call is badly designed. Good ones distinguish between the thing you’re avoiding and the thing you actually need.
Sites like stickK, set up in 2007 by economists Dean Karlan and Ian Ayres with Jordan Goldberg, formalise this: you name a goal, put money on it, and nominate a referee to verify whether you did it. That’s a heavier instrument than most study problems need, but it’s the clearest example of the mechanism.
What a small stake looks like inside a timer #
Work Capy is a focus timer built as a deliberately mild commitment device, and its design choices illustrate the principles above. You start a session and put the phone down. Open another app for more than about five seconds and a card warns you once. Leave a second time and the session is forfeit: it pays no coins and its minutes don’t count toward your total.
The size of the penalty is the point. One slip costs one session, never your streak, your history or anything you’ve earned. Locking the phone never counts as leaving, because that’s the behaviour it’s asking for, and pausing is free, so a real interruption costs nothing if you pause first.
Two further details follow from taking commitment seriously. No purchase and no advert can restore a forfeited session, because a consequence you can buy your way out of isn’t a consequence. And if stakes are the wrong tool for you, a Gentle mode in settings turns the rule off entirely rather than the app deciding on your behalf.
When do commitment devices backfire? #
- When the stake is too large. Anxiety about losing something valuable competes with the work for your attention.
- When they punish the wrong thing. If the rule can’t tell an emergency from procrastination, you’ll stop trusting it and then stop using it.
- When they replace fixing the real problem. If you can’t start because the task is undefined or you’re exhausted, no penalty will help. Stakes are for the gap between intention and action, not for everything.
- When self-criticism is already the issue. Sirois and Pychyl’s account of procrastination as short-term mood repair suggests that feeling bad about a task feeds avoidance. Adding punishment to a task you already dread can make the avoidance worse. In that case, shrink the task instead; there are options in how to stop procrastinating.
Frequently asked questions #
What is a commitment device in simple terms? #
It’s a decision you make now that limits what you can do later, on purpose. Setting up an automatic transfer to savings, giving your phone to someone else during study time, or booking a non-refundable class are all commitment devices.
Do commitment devices actually work? #
The field evidence says yes, with modest effect sizes. A commitment savings product for smokers raised six-month quit rates by 3 percentage points and the effect lasted to 12 months. Restricting audiobooks to the gym raised attendance by up to 51% initially, though that faded. Take-up rates are typically low, which suggests they work best for people who choose them.
Should I bet money on my study goals? #
Usually not as a first step. Financial stakes are a strong instrument and they’re easy to set too high, which creates stress and avoidance. Start with free constraints: the phone in another room, a fixed session time, someone expecting your work, or a timer with a small in-session cost.
Are app blockers commitment devices? #
Yes, and they’re among the more effective kinds because they act before the decision rather than after. Their limit is that most can be turned off if you really want to. We compare the harder and softer options in app blockers vs focus timers.
Do streaks count as commitment devices? #
Loosely. A streak creates something to lose, which is the right shape, but it also creates an all-or-nothing structure that can end a habit the day it breaks. We weigh both sides in do study streaks work?