Implementation intentions: if-then plans that stick

"I will read more" is a goal. "If I sit on the train, then I open the book before the phone" is an implementation intention. Peter Gollwitzer's claim, developed across decades of social psychology, is that the second sentence does work the first sentence cannot. It pre-decides the when and the where, so that the situation can trigger the action instead of waiting for a fresh burst of will.

That sounds small. It is small. The intention-behaviour gap is also small-looking on paper and large in January. Thomas Webb and Paschal Sheeran showed that interventions which change intention still leave a lot of behaviour unmoved. An if-then plan is one of the few tools aimed at the gap rather than at the slogan.

A sentence with a hinge

The hinge is the word if. It forces you to name a situation. Without a situation, you are rehearsing a personality. With a situation, you are writing a cue. Wendy Wood and David Neal's account of habits is that context retrieves the response. Gollwitzer is giving you a way to nominate that context in advance, before automaticity has had time to grow.

What belongs on each side of the hinge

The if should be an event you can perceive without a meeting with yourself: a doorway, a kettle, a calendar alert, the end of a call, putting a child to bed. Internal states ("if I feel inspired") are weak ifs. They reimport the problem you were trying to leave. The then should be an action you can complete, not an outcome. "Then I am healthier" is not an action. "Then I put on shoes" is.

BJ Fogg's Tiny Habits method is almost an if-then with the then shrunk. After I brush my teeth, I floss one tooth. After I pour coffee, I open the notebook to a blank page. Fogg cares about ability because an if-then that points at a 45-minute ordeal will lose to the sofa. Gollwitzer's structure plus Fogg's sizing is more robust than either slogan alone.

What the sentence does not promise

It does not mint a habit on Tuesday. Lally and colleagues (2010, European Journal of Social Psychology) found a median of about 66 days to an automaticity plateau, with a range from 18 to 254. An if-then is how you get the repetitions that the curve needs. It is not a substitute for the curve. Charles Duhigg's cue-routine-reward loop is a popular sketch of the same sequence. Duhigg popularised the loop; he did not trial if-then wording against goal wording. Keep the credit with Gollwitzer for the sentence, with Wood for the context, with Lally for the time.

Cues that already exist

The best if is already in your day. Inventing a new time slot ("if it is 5 a.m.") adds a second behaviour: waking, which you may not have. Piggybacking on a behaviour that already runs is cheaper. That is why stacking and implementation intentions get confused. Stacking is a lifestyle-blog name for using an existing cue. The mechanism is still if-then.

Unstable days need portable ifs

If your context jumps (travel, shifts, a new baby), a place-based if will fail. Use an action-based one: if I wash my hands, then I take the supplement. The wash happens in more buildings than your kitchen does. Wood would still prefer a stable context when you can get one. When you cannot, the hinge should attach to something that travels.

Webb and Sheeran's gap widens when the plan is left as "I should, in general." It narrows when the critical situation is identified. That is the whole trick. It is not mystical. It is a reduction in the number of decisions required at the moment of action.

Too many if-thens become wallpaper

A list of twelve if-thens is a second job. Attention is limited. One or two sentences, practised until they are boring, beat a manifesto. Lally's participants worked on one self-chosen behaviour. That constraint is underrated. If-then is a scalpel. People use it as a spray.

Failure modes of if-then plans

The plan is too vague. The action is too large. The cue never happens. The cue happens in a room that already owns a competing response (the sofa owns "sit and scroll"). Competing responses are Wood's territory: the old habit has a head start in that context. You may need to change the room, not only the sentence.

Weak plan	Why it fails	Stronger rewrite
If I have time, then I will run	"Time" is not a perceptible cue	If I hang my bag, then I put on shoes
If I am motivated, then I journal	Motivation is the scarce resource	If I pour tea, then I write one line
If it is morning, then I do a full workout	The then is sized for a perfect day	If the kettle switches off, then I do two squats
If I remember, then I take the tablet	Remembering is not a cue	If I rinse the mug, then I take the tablet

Duhigg's reward box can still help initiation: the then can include a small, immediate consequence. Do not wait for the reward to do the cue's job. The cue's job is retrieval. Gollwitzer assigned that job to the if.

Write the sentence where you will see it for a week, then hide it. If you still need the poster after a month, the if is probably wrong, or the then is still a speech.

Habit AI's one-sentence parse is a practical if-then: the habit has to fit in a line you can actually run, which is Gollwitzer's constraint wearing a text field.

Read next

Automaticity versus willpower explains why the sentence beats grit. The habit loop: cue, routine, reward is the cartoon this sentence makes specific. the iPhone tracker is where that sentence lives on a phone.