Implementation intentions: the if-then plan that beats willpower

An implementation intention is a plan in the form “if situation X arises, then I will do Y”, such as “if I’ve just dropped the kids at school, then I walk for 20 minutes before going home”. Because the when and the where are settled in advance, the moment only has to be recognised, not argued with. A 2006 meta-analysis of 94 tests found that adding a plan like this to a goal had a medium-to-large effect on reaching it, and a much larger review in 2025 found the average effect smaller but reliable.

Why don’t I follow through on my goals?

You probably aren’t short of wanting. Most people who drop a goal meant it when they set it, and the research on this is blunt. Reviewing the evidence in 2002, the psychologist Paschal Sheeran found that across studies, a median of 47% of people who intended to act didn’t, while only 7% of people who didn’t intend to act ended up acting anyway. Psychologists call this the intention–behaviour gap, and most of it is made of people who meant it.

Wanting was step zero. The trouble comes at the moment of starting. Peter Gollwitzer’s 1999 review lists the usual ways it goes wrong: you’re absorbed in something else, caught up in your own thoughts or a strong feeling, or just tired, and the chance to act goes by without catching your attention. Some chances only last a minute. And a decision made on the spot is expensive, because the moment has to weigh itself against every other moment that might suit better, and against every other goal you hold.

That’s a lot of negotiating to do at 6pm on a Wednesday, and the default has the advantage. Unstarted goals don’t go quietly, either; why they come back at night is its own effect.

What the research found

Gollwitzer described implementation intentions in a 1993 paper, as the second of two kinds of intention. A goal intention names an end point: I intend to reach X. An implementation intention serves the goal and names the when, where and how: when situation X arises, I will do Y. His 1999 review in American Psychologist, “Implementation intentions: Strong effects of simple plans”, brought the early studies together.

Some of the first studies ran over a Christmas break (Gollwitzer and Veronika Brandstätter, 1997). Students named a difficult project for the holiday. Those who had already decided when and where to start finished 62% of them; those without a plan finished 22%. In a second study, every student was set the same awkward task, a report on their Christmas Eve written within the 2 days after it, and half were randomly asked to pick a time and a place. Of the 36 reports that came back and could be matched, 71% of the planners had written theirs in time (12 of 17), against 32% of the rest (6 of 19). The numbers are small, and in the first study students chose for themselves whether to plan, but the second assigned the plan at random and pointed the same way.

Gollwitzer’s explanation has two parts. Choosing the situation in advance makes it easier to spot: it stands out when it arrives, even while you’re busy with something else. Linking it to an action makes the start close to automatic, because the decision has already been made and there is nothing left to deliberate. In 1993 he described this as “passing the control of one’s behavior on to the environment”.

In 2006, Gollwitzer and Sheeran pooled 94 independent tests from 63 reports, with 8,461 participants between them. Compared with holding the goal alone, adding an if-then plan gave an average effect of d = .65, which they called “medium-to-large”. (d is the difference between two groups measured in standard deviations; by the usual rule of thumb, .2 is small, .5 medium and .8 large.) It held up in experiments, on objective measures and in unpublished studies, though 79 of the 94 tests were run with university students.

When an if-then plan won’t help

The plan carries a goal. It doesn’t supply one. In a study described in Gollwitzer’s 1999 review, women set themselves the goal of a breast self-examination in the coming month. Among those who strongly intended to do it, 100% of those who also made a when-and-where plan did it, against 53% of those who didn’t, and the benefit appeared only among the strong intenders. His conclusion: “Implementation intentions, however, need to be based on strong goal intentions.” In 2006 he and Sheeran, citing the studies that had tested it, put it more firmly: “strong effects of if–then plans only emerge when participants hold strong respective goal intentions”.

The largest review so far makes the same point and trims the headline. In 2025, Sheeran, Olivia Listrom and Gollwitzer analysed 642 tests and found an average effect of d = .36. Small studies with positive results were over-represented, a sign of publication bias. One standard correction barely moved the estimate (d = .35), but a Bayesian method brought it down to d = .15, which they call “small according to conventional guidelines”. The effect was real by their analysis, and its size varied a great deal from study to study.

Exercise shows where some of that variation comes from. Across 64 tests, if-then plans barely moved physical activity (d = .14). When the authors split the studies that measured it by how strongly people meant to exercise in the first place, the plans helped when intentions were strong (d = .31) and did nothing when they were weak.

So the title needs a footnote. An if-then plan beats a goal held on its own, and it works best for a goal you actually want. If you don’t want to run, a plan about running won’t change that, and the more useful question is what you’d rather do instead.

How to write a strong if-then plan

The cue should be a moment that happens anyway, not a mood. Compare:

• Weak: “I’ll exercise more.”
• Strong: “After I drop the kids at school, I walk 20 minutes.”

The weak one waits for a feeling. The strong one names an event that turns up every school day whether you’re in the mood or not, and attaches one action to it. “When I have time” and “when I feel up to it” are the classic weak cues: they arrive rarely, and never on schedule.

The 2025 review also compared the plans themselves, and a few results are worth borrowing:

• Keep the if-then shape. A calendar entry tells you when; an if-then ties the moment to the act. Plans written as if-then did better than plans written as a schedule of when, where and what (d = .43 against .29).
• Name a place as well as a time. Time and place together did better than a time alone (d = .46 against .25). “Tuesday” is weaker than “Tuesday, when I get off the train”.
• Say it once. Plans that people went over at least once did better than plans they never rehearsed (d = .45 against .33), and going over them more often added nothing.
• Then leave it alone. In studies where people wrote their plans down and could look at them again, the plans worked less well (d = .31 against .44), which the authors link to deliberation getting in the way of an automatic start. Their summary: “It matters that participants are motivated to act, that plans use a contingent format, and that the plan is rehearsed but not recorded.”

Plans for obstacles

The same shape works for the moment a plan usually breaks. Pick the obstacle you can see coming, and decide your response while you’re calm: “if I want to skip the run, then I put my shoes on and run for 5 minutes.” The smaller version keeps the cue and the action tied together on the days the full version won’t happen.

The research covers this too. In studies described in Gollwitzer’s 1999 review, students did arithmetic while clips of award-winning adverts played at random. A plan to ignore the distraction whenever it appeared helped across the board; a plan to try harder whenever it appeared helped only when motivation was low or middling. His advice: plan to ignore distractions rather than to step up effort.

In the 2006 meta-analysis, plans that protected a goal already under way had an average effect of d = .77 across 21 tests, against d = .61 across 52 tests of plans for getting started. And the “then” can be a line of self-talk: in the 2025 review, plans whose response was to tell yourself to keep calm did well across 14 tests (d = .69).

Is habit stacking the same thing?

Mostly. Habit stacking is the popular name for an if-then plan whose cue is something you already do. James Clear’s formula in Atomic Habits (2018) is “After/Before [CURRENT HABIT], I will [NEW HABIT].” He calls habit stacking “a special form of an implementation intention”, and spells out the difference: “Rather than pairing your new habit with a particular time and location, you pair it with a current habit.” Clear credits the method to BJ Fogg, whose Tiny Habits programme calls the existing routine an anchor, and the name to S.J. Scott, from Scott’s book of that title. Fogg’s approach, and why it starts so small, has its own post.

Treat habit stacking as a practice: the studies above tested if-then plans, not habit stacking under that name. The reasoning carries over, though, because a routine you already have is a cue that turns up every day without a reminder. The kettle, the commute, the laptop closing.

Gollwitzer wrote in 1999: “Apparently, implementation intentions create instant habits.” He meant the start: a plan can make an action follow its cue about as fast as a habit does. A habit that holds up without the plan takes repetition, and how long that takes is a separate question.

If-then planning examples

Area	Weak	Strong
Exercise	I’ll exercise more.	If I’ve just dropped the kids at school, then I walk 20 minutes before going home.
Exercise (obstacle)	I’ll push through when I don’t feel like it.	If I want to skip the run, then I put my shoes on and run for 5 minutes.
Writing	I’ll write every morning.	If I’ve sat down at my desk with the first coffee, then I open the draft before I open email.
Writing (obstacle)	I’ll stop checking my phone.	If I reach for my phone mid-paragraph, then I put it face down and finish the paragraph.
Admin	I’ll sort out my receipts.	If it’s Friday and my last meeting has ended, then I photograph the week’s receipts and file them.
Calls	I should call my mum more.	If I’m walking home from the station on Sunday, then I call my mum.
Calls	I’ll book the dentist.	If I’ve just finished lunch on Monday, then I call the dentist and book the check-up.
Calls (obstacle)	I’ll stop putting off difficult calls.	If I feel myself putting the call off, then I tell myself to keep calm and dial.

Pick a goal you already mean, find the moment in tomorrow that will happen whether you’re ready or not, and finish the sentence: if ___, then I will ___. Say it to yourself once, in full. The deciding is done, and tomorrow only has to notice.

Sources

• Clear, J. (2018). Atomic Habits. Avery. Quoted from the excerpt on the author’s site, “How to Build New Habits by Taking Advantage of Old Ones”. https://jamesclear.com/habit-stacking
• Gollwitzer, P. M. (1993). Goal achievement: The role of intentions. European Review of Social Psychology, 4(1), 141–185. https://doi.org/10.1080/14792779343000059
• Gollwitzer, P. M. (1999). Implementation intentions: Strong effects of simple plans. American Psychologist, 54(7), 493–503. https://doi.org/10.1037/0003-066X.54.7.493
• Gollwitzer, P. M., & Brandstätter, V. (1997). Implementation intentions and effective goal pursuit. Journal of Personality and Social Psychology, 73(1), 186–199. https://doi.org/10.1037/0022-3514.73.1.186
• Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. Advances in Experimental Social Psychology, 38, 69–119. https://doi.org/10.1016/S0065-2601(06)38002-1
• Sheeran, P. (2002). Intention–behavior relations: A conceptual and empirical review. European Review of Social Psychology, 12(1), 1–36. https://doi.org/10.1080/14792772143000003 (figures as summarised in Gollwitzer & Sheeran, 2006)
• Sheeran, P., Listrom, O., & Gollwitzer, P. M. (2025). The when and how of planning: Meta-analysis of the scope and components of implementation intentions in 642 tests. European Review of Social Psychology, 36(1), 162–194. https://doi.org/10.1080/10463283.2024.2334563