Two things a day, done where they belong. This is the one that makes the other three actually happen.
is the median time for a behaviour to become automatic. But the real range across people was 18 to 254 days — so if it still feels like effort at week six, nothing is wrong with you.
Lally, van Jaarsveld, Potts & Wardle (2010), European Journal of Social Psychology — 96 people tracked daily for 12 weeks.First name. It stays on this phone — nothing gets sent anywhere.
Four things came out of the research that change how you should run this. Two of them contradict what most people are told.
In the study that produced the 66-day figure, missing a single opportunity did not materially affect habit formation. One miss is noise. What breaks it is missing twice, then telling yourself the run is over.
Lally et al. (2010) — which is why this tool has no streak to lose.Habits form through repetition in a consistent context — the cue is the mechanism, not the willpower. And cues tied to something you already do (“after I make coffee”) hold up better than cues tied to a time (“at 7am”), because you notice them without having to monitor for them.
Gardner & Lally on cue-dependent automaticity; Pimm et al. on routine versus time-based cues.Simple behaviours become automatic considerably faster than complex ones, and complex ones vary far more between people. A version of the thing that takes five minutes will stick where the full version won't.
Lally et al. (2010); reviews of behavioural complexity and contextual stability.This is the one nobody says. Sessions experienced as unpleasant fail to produce lasting automaticity — repetition alone isn't enough. Grinding through something you hate every day builds resentment, not a habit.
Weyland et al. (2020) on affect and habit strength in exercise.What this doesn't mean. Not that it should be easy, or that you skip when you don't feel like it. It means if you've built something you actively dread, change the thing — not your willpower.
Two is the limit. More than that and neither sticks. For each one you need three things: what it is, what already happens right before it, and the smallest version that still counts on a bad day.
On the minimum version: it isn't cheating. It's what keeps the cue intact on the days the full thing isn't happening — and the cue is the part that's actually building.