Do It Tomorrow and Commitment Contracts: One Narrow Place They Fit

What each one is actually for

Do It Tomorrow

Do It Tomorrow, from Mark Forster's 2006 book of that title, starts from an unusual diagnosis. Forster's claim is that most people who feel disorganised are not disorganised — they are taking on more work than there are hours, and no system can fix a volume problem, it can only make it visible. So the system is built to make it visible. You write tomorrow's list today, at the end of the day, from what actually arrived. Then you draw a line. Tomorrow, that is the list, and anything new that comes in goes onto the following day's page instead of interrupting. Because the list is closed rather than open, you can see at the start of the day whether it is doable, and over a week or two you can see whether you are systematically finishing them or systematically falling behind — which is the real number Forster wants you to have. He pairs it with a working method he calls little and often: rather than trying to finish each item, you make a pass down the list doing a bit of each, then another pass, which keeps long tasks from blocking everything behind them.

Commitment Contracts

A commitment contract is a deliberate reduction of your own future freedom. You state a goal in terms specific enough that someone else could verify it, you name that someone as a referee, and you put up a stake — usually money, held by a third party, forfeited if the referee says you did not do it. Websites built for this, of which stickK is the best known, add one refinement that turns out to matter: you can direct the forfeited money to a cause you actively oppose, which people find considerably more motivating than losing it to charity. The logic is not that you need punishing. It is that you are two different people at two different moments — the person on Sunday who wants to run four times this week, and the person on Wednesday at six in the morning — and Sunday-you has no other way to bind Wednesday-you. The contract is Sunday-you making Wednesday's easiest option expensive, in advance, while Sunday-you is still the one in charge.

What you get out of pairing them

Here is the reason to be careful with this pair rather than enthusiastic about it. Commitment contracts work by making failure cost something, and that only produces a good outcome if the target is genuinely under your control. Almost nothing about a normal working day is: a task list is a forecast, and by three in the afternoon somebody has escalated something, a call has run long, and a thing you thought was an hour was four. Attach money to finishing that list and you have not bought yourself discipline. You have bought yourself a strong daily incentive to write shorter lists, to define done loosely, and to be furious at colleagues for existing. But the failure mode of Do It Tomorrow is real too, and it is very specific: the system depends entirely on your honouring the deferral, and it dies at the moment you decide something is too urgent to wait — because that decision is available every hour, it feels responsible every time, and after a fortnight of small exceptions the line under the list means nothing. That specific behaviour, unlike list completion, is completely within your control and takes a second to verify. It is exactly the kind of thing a stake is good at. So the honest offer of this pairing is narrow: not more force applied to your workload, but a small cost attached to the one habit that holds the whole structure up.

What Do It Tomorrow rests on

• Implementation intention closes intention action gap randomised trials
  Pre-committing to a specific situational cue paired with a specific response lets the behavior fire automatically when the cue occurs, without depending on in-the-moment willpower.
• Closed lists reduce overwhelm practitioner report
  A fixed bounded list of tasks rather than an ever-growing open list reduces the felt overwhelm of an infinite to-do list.
• Deferred capture protects commitment practitioner report
  Writing newly arriving tasks onto tomorrow's list rather than acting on them immediately protects today's already-committed plan from being derailed by reactivity.

What Commitment Contracts rests on

• Accountability to others increases follow through observational studies
  Having another person witness or enforce progress raises the likelihood that a commitment is kept.
• Loss aversion strengthens commitment randomised trials
  Framing a stake as a loss to avoid rather than a gain to earn increases adherence because losses loom larger than equivalent gains.
• Precommitment constrains future choice randomised trials
  Binding oneself in advance to a costly consequence removes the option to yield to in-the-moment temptation.

Where they actually meet

The single place these genuinely meet is a rule about arrivals, and it is worth being precise about it because the rest of the overlap is illusory. Forster's rule — new work goes to tomorrow, not today — is a decision made once, in advance, that then applies automatically to a whole category of future situations you have not seen yet. You do not evaluate each incoming request on its merits at the moment it arrives; you have already ruled on all of them. A commitment contract is doing the same thing with a different enforcement mechanism: a decision made once, applied to future moments, deliberately removing your discretion at the point where discretion is unreliable. Both are betting that the version of you who is calm and unhurried should get to bind the version who is not, and both accept a real cost for that — Forster's cost is that a genuinely urgent thing waits a day, the contract's cost is money. Where they part company is scope. Forster is protecting a whole day's plan from reactivity. The contract is enforcing one named, verifiable behaviour. Blur those and you get the version that backfires; keep them distinct and you get one useful join.

The principle underneath both

A decision made in a cool moment is a better decision than the same decision made in a hot one, and the only way to make the cool decision stick is to remove your ability to reopen it when the hot moment arrives. That is what both practices are instances of, and it is much older than either. What is distinctive here is where each one places the removal. Forster removes it procedurally — there is nowhere to write the new task except tomorrow's page, so the reopening has no physical home, which is a quiet and frictionless kind of enforcement that also means nothing stops you if you simply decide to override it. The contract removes it by price — you can absolutely reopen the decision, and it will cost you two hundred pounds. Procedural enforcement is cheap and leaky. Priced enforcement is expensive and tight. Which one is right depends entirely on whether the behaviour in question is one you actually control, and that question is the hinge for everything on this page.

What that makes the real goal

Stated so it belongs to neither source: making a decision you have already made hold up against the specific later moment that is designed to overturn it. Notice that this puts the emphasis on identifying the moment rather than on strengthening the resolve. Every plan has a characteristic point of failure and it is usually singular and predictable — for a closed list it is the mid-afternoon request from someone senior; for an exercise goal it is the first cold morning; for a spending limit it is the second drink. The objective is not general discipline, which nobody has ever successfully installed. It is knowing which single moment kills this particular plan and putting something in that moment's way. Once framed that way, the design question becomes concrete: what is the cheapest obstacle that would actually hold at that specific point, and does it need to cost money at all?

Other practices that serve the same goal

Several other approaches serve that objective, and a couple of them carry better evidence than either practice here. Self-imposed deadlines are the closest cousin: Dan Ariely and Klaus Wertenbroch ran a study in which students could set their own deadlines for three papers, and those who spaced them out performed better than those who left everything to the end — but the striking finding was that people set their deadlines suboptimally even when free to choose, which is a useful caution about trusting your own contract design. Implementation intentions, the if-then plans studied by Peter Gollwitzer, do the same job without any stake at all: you name the exact situation and the exact response in advance, and a meta-analysis of ninety-four independent tests found a substantial effect on whether the behaviour occurred. That is worth trying before you put money on anything. Temptation bundling, tested by Katherine Milkman and colleagues, permits some enjoyable thing only during the activity you avoid, which changes the price of the moment rather than the price of failure. Removing the option physically — the phone in another room, the card not saved in the browser — is unglamorous and often more reliable than any agreement, because it does not depend on you honouring anything. And plain accountability, a named person who will ask, sits underneath commitment contracts anyway: in most contract designs the referee is doing a large share of the work, and it is not obvious how much the money adds on top.

A tradition that already does this

There is a tradition that already fused a closed daily list with a binding commitment, and it ran for fifteen centuries. The Benedictine Rule, written in the sixth century, structures the monastic day into fixed hours with fixed occupations, and the crucial feature for this comparison is that the day's contents are settled by the horarium rather than by what arrives — the day is closed before it begins, in exactly Forster's sense, and a request that appears at eleven does not restructure it. Sitting underneath that schedule is a vow, made once, publicly, with real and irreversible consequences for abandoning it, which is a commitment device in the strict sense. What the Rule shows, and what is worth borrowing, is which layer carries the binding. The vow is attached to remaining in the life, not to completing each day's work. Nobody is bound to finish; they are bound to stay inside the structure that decides what finishing means. That is precisely the distinction this page is arguing for, arrived at independently and stress-tested over a very long period, and it is the strongest available argument that the join described here is the right one and the obvious join is not.

Using the closed list to give a contract a target that cannot inflate

Commitment contracts fail in a predictable way, and it is almost never about willpower. It is that the goal was badly specified and the specification could move. Write your goal as be more productive and there is nothing for a referee to check. Write it as finish my work each day and you have bound yourself to a quantity that other people control, which means you will either lose money for things that were not your doing or start quietly redefining what work means, and the second one is worse because it corrupts the instrument silently. A closed list is unusual in a working life because it fixes the target before the period starts, in writing, at a moment when nobody is pressuring you. That is a genuine property, and it is the property a stake needs. Here is the concrete move if you want to use it. Write tomorrow's list tonight and photograph it, or timestamp it in whatever app you use. That photograph is the contract's scope — not the work you did, the list you committed to at nine hours' distance from the day. Then, and this is the part that keeps it safe, set the stake on the list being closed rather than completed. Closing it means one of two things happened to every line: it got done, or it got explicitly moved to a named later day with a reason. What you are paying to prevent is items evaporating — the ones that neither get done nor get rescheduled and simply stop being mentioned, which is how a backlog becomes invisible. Keep the stake small, five or ten pounds a working day, and run it for two weeks rather than indefinitely. Two weeks is enough to find out whether the list is honest, and past that the contract stops teaching you anything and just becomes a tax on a bad week.

Putting a price on the exception that dismantles the system

Forster's system has one load-bearing behaviour, and if you have tried it you know exactly which one. It is not writing the list and it is not the little and often passes. It is what you do at half past two when your manager asks for something today. The rule says it goes on tomorrow's page. The rule feels wrong in that specific moment, every time, and it feels wrong in a way that sounds like responsibility rather than like weakness, which is why it wins. Nothing in the book stops you, because the enforcement is procedural — there is no place to write it, so you just do not write it anywhere and do it instead. Ten of those in a fortnight and the closed list is a fiction, and worse, you have lost the one number the system existed to give you, since the count of what you finished no longer means anything. This is where a stake is genuinely well-matched, because unlike list completion, the exception is fully in your control and takes three seconds to check. The concrete version: name one person as referee, agree a stake you would actually mind — for most people that is more than five pounds and less than fifty — and define the failure precisely as any item entering today's list after the line was drawn. Not as finishing the list. Not as productivity. One binary fact about one behaviour, verifiable by looking at the page. Then build the escape valve into the rule rather than leaving it to the moment, because a rule with no valve will be broken on the day something genuinely does need doing at three. Forster's own answer is an audit trail: a genuine emergency goes on the page in a different colour, and you get one a week without penalty. What you have then is a system where the exception still exists but costs something to use, which is the difference between a boundary and a wish.

Twenty to three on a Wednesday, and the message says any chance you could look at this today

You are eleven items into a list of sixteen, and it was going unusually well. The message is from someone you like, it is not unreasonable, and it will take about forty minutes. Everything in you is already reaching for the yes, and here is what is actually happening in that reach: you are not weighing today's list against this request, because today's list is not in front of you as a quantity — it is behind you as eleven done things and ahead of you as a vague five. The request is vivid and the list is not, and vividness wins every time. What you would say if you were being honest is not I have capacity, it is I do not want the small social cost of saying tomorrow. That is a real cost and it is worth naming rather than pretending you are making a scheduling decision. Now the version with the stake in place. You have agreed with a colleague that anything entering today's page after the line was drawn costs you twenty pounds to a political party you cannot stand, and you get one free emergency a week, marked in red. Sitting at twenty to three, the arithmetic is suddenly explicit rather than felt. Is this the week's emergency? Almost certainly not, on a Wednesday, for a forty-minute favour. So you type the sentence you would not otherwise have typed: yes, first thing tomorrow, it is on the list. And the thing worth noticing is what that costs, honestly. It costs a small amount of social friction. It cost you nothing in outcome — they get it Thursday morning rather than Wednesday evening, and the difference to them is close to zero. The stake did not give you willpower. It made a decision you were making implicitly, on feel, into a decision you had to make explicitly, with a number attached, at the exact moment the implicit version was about to win. That is the whole mechanism, and it is smaller and less impressive than it sounds — which is why the honest use of it is one narrow behaviour and a fortnight, not a permanent regime over your working life.

When this is worth reaching for

Skip this until a closed-list system has actually been tried; a stake has nothing to enforce against a habit that was never attempted. Once one has been tried, watch for a specific degradation pattern: it worked for about a week, and it degraded through exceptions rather than through abandonment — one reasonable-sounding item let back in at a time. That pattern is the signal this combination answers. If the lists stopped being written altogether, the problem is elsewhere and a stake will not touch it; but if they kept being written and things kept being let in, that is a boundary that needs a price rather than a system that needs replacing. It is also relevant when the exceptions come from one identifiable source — a particular colleague, a particular kind of request — because then the contract can be written narrowly around that source and the rest of the working life stays unpriced. A third case: short, defined periods where holding the line actually matters, like the four weeks before a deadline that everything else keeps eating into. A stake with an end date is a much better instrument than a stake without one; this combination is at its best as a temporary measure with a specific target, not as a way of running a job.

Where this does not apply

The main thing to say is the one from the top of this page. Do not put a stake on completing a daily task list. A day's workload is not a quantity you control, and pricing failure at something you do not control produces gaming rather than performance — shorter lists, vaguer items, and a running resentment at anyone whose request lands on your desk. That is not a hypothetical risk of the combination, it is the most natural way to combine them, which is why it is worth stating first rather than last. Second: if the reason your list does not get finished is that you have more work than hours, this pair will make it worse and Forster would say the same. His system is diagnostic — the count of unfinished lists is meant to tell you that you must shed work or renegotiate — and a stake converts that signal into pain without changing the underlying arithmetic. Fix the volume first. Third: commitment contracts appear to work poorly for people who are already prone to punishing themselves for shortfalls, since the mechanism adds an external penalty to an internal one that is already firing, and there is a real difference between an incentive and a scheduled reason to feel bad. Fourth, and this one is empirical rather than cautionary: the largest trials of deposit contracts have found that when people accept them they work well, but far fewer people accept them than accept a reward-based alternative. If you find yourself reluctant to actually put money down, that reluctance is informative, and the answer is probably a smaller stake or a plain if-then plan rather than talking yourself into it.

What the evidence does and does not support

The two halves of this pair sit at opposite ends of the evidence spectrum, and the gap is unusually wide. Commitment contracts have been tested in real randomised trials with real money and real outcomes. Xavier Giné, Dean Karlan and Jonathan Zinman ran a voluntary deposit contract for smokers in the Philippines and found meaningfully higher quit rates, verified by biochemical testing rather than self-report, among those offered the contract. Heather Royer and colleagues found gym attendance effects from a commitment device that persisted after the incentive ended. Scott Halpern's team, publishing in the New England Journal of Medicine, compared deposit-based and reward-based smoking cessation programmes and found the deposit contracts substantially more effective among people who accepted them — but that far fewer people accepted them in the first place, which is the finding most worth carrying away, because it means the technique's real-world value is limited less by whether it works than by whether you will actually use it. Do It Tomorrow has nothing comparable. It is one experienced practitioner's system, set out in a trade book in 2006, and its distinctive claims — that a closed list reduces the feeling of being overwhelmed, that deferring arrivals protects the day — have never been tested against an open-list control, which is why our own grading marks them anecdotal. The one thing shared between them, pre-deciding a response to a future situation, does have strong support in the if-then planning literature, but that support attaches to the shared mechanism rather than to either practice as taught. Nothing has tested the specific join described here, and testing it would be straightforward: recruit people running a closed-list system, randomise half to a small stake on the no-additions rule, and count additions over eight weeks. That study does not exist.