← Back to the library

    Implementation intentions

    holds up

    Researchers have tested this many times, in different places, over many years, and they keep finding the same result. This one is safe to trust.

    In 2002, Paschal Sheeran pooled the existing meta-analyses on the gap between what people intend and what they actually do. Across the studies he gathered, roughly half the people who had formed a positive intention did not go on to act on it. The intentions are usually sincere. The problem is that a goal without a trigger depends on you noticing the right moment, remembering, and deciding all over again while tired, busy, or interrupted. The fix turned out to be very small. In one exercise study, a motivational programme built on risk and self-efficacy left only about a third of participants exercising weekly. A second group did the same programme and then wrote one sentence naming exactly when and where they would exercise. 91 percent of them did. Gollwitzer's explanation is that an if-then plan hands the job of starting to the situation itself. You stop guarding the goal in your head, and the cue does the remembering. When Tuesday at seven arrives, the decision is already made, so nothing gets renegotiated in the moment. Across those 94 tests, from vitamin tablets to cancer screening to reports written over holidays, the same small sentence keeps producing the same lift.

    The gap this closes is initiation, the space between deciding and starting. Most plans die there, and most advice attacks it with more motivation, which the exercise study shows is the weak lever. The stronger lever is removing the decision from the moment. The format matters: if situation, then action. If it is Tuesday at seven, I run before opening the laptop. If the meeting ends without a decision, I send the summary with a deadline in it. If the client goes quiet for a week, I call on day eight. Teams can hold these as standing rules, which is what good operating cadences amount to. Two honest limits. The big effect sizes come from discrete, nameable actions, and the further the outcome sits from a single act, the smaller the lift, which is why health-domain meta-analyses find more modest gains than the headline figure. And the plan spends motivation rather than creating it. That used to be an inference from the theory, and it has since been measured. A 2024 review covering 642 independent tests found that the effect holds across cognitive, affective and behavioural outcomes, with sizes running from d = .27 to d = .66 depending on what is being measured, and that it is larger where the plan uses the if-then form, where the plan gets rehearsed, and where the person is already motivated toward the goal. So if you do not actually want the thing, naming Tuesday changes very little. What the plan removes is the tax of deciding again every time.

    Read this against
    Affective forecasting

    Implementation intentions say that tying an action to a concrete cue, if this then that, raises follow-through by a medium-to-large margin across 94 studies. Affective forecasting says the goal you are committing to may rest on a misprediction of how the outcome will feel. So the pre-commitment that makes you act may be locking in a forecast that was wrong. The intentions entry is honest about this on its own: the plan spends motivation rather than creating it, and if you do not actually want the goal, naming Tuesday changes nothing. What separates them is which failure you are in. If the problem is starting, intentions are the lever. If the problem is that the goal was chosen by imagination rather than by sampling, an if-then plan just gets you there faster.

    Source: Peter Gollwitzer, Implementation Intentions: Strong Effects of Simple Plans, American Psychologist, 1999. The meta-analysis: Gollwitzer and Sheeran, Implementation Intentions and Goal Achievement, Advances in Experimental Social Psychology, 2006, 94 independent tests, d = .65. The size of the gap: Sheeran, Intention-Behaviour Relations: A Conceptual and Empirical Review, European Review of Social Psychology, 2002, a meta-analysis of meta-analyses. The exercise study: Milne, Orbell and Sheeran, British Journal of Health Psychology, 2002. The current synthesis: Sheeran, Listrom and Gollwitzer, 2024, 642 independent tests, with effects across cognitive, affective and behavioural outcomes ranging from d = .27 to d = .66, and larger effects where the plan used the if-then format, where it was rehearsed, and where the person was highly motivated toward the goal. Domain-level meta-analyses in physical activity and eating find smaller but still positive effects.

    The book, if you want to go further

    Rethinking Positive Thinking

    Gabriele Oettingen, 2014

    From Gollwitzer's closest collaborator: why positive fantasy alone reduces follow-through, and how to pair the wish with the obstacle before writing the if-then plan.

    The Saturday letters

    Every Saturday, one idea worth your time, with an honest account of how far the evidence really goes and where it stops. Free, a few minutes to read, and it leaves you with questions rather than instructions.

    You will get one email asking you to confirm this address. Nothing arrives until you click it.

    Draw your own card. It does not take long, and it rewards taking your time.