Habits Don’t Form on a Fixed Timetable
The number varies enormously by person and by behaviour. Expecting a deadline for automaticity is itself one of the more reliable ways to abandon something.
You start a new habit in the first week of the month, and somewhere along the way you read that habits take twenty-one days to form, or sixty-six, or ninety, depending on the article. You count forward and circle a date. The appeal is easy to understand, because a number turns an open-ended undertaking into something with an end, and an ending is much easier to commit to than an indefinite stretch of effort.
The difficulty is that no such number holds for everyone. The variation around any average is enormous, and carrying a date in your head turns out to be one of the more reliable ways to give up shortly after reaching it.
What the research found about how long a habit takes
The best-known study found that the time a habit took to become automatic varied hugely from person to person and behaviour to behaviour. Phillippa Lally and colleagues followed people adopting a new daily behaviour over twelve weeks, modelling how automatic it became as the days went on.
The median was 66 days, and that figure has travelled widely while the rest of the result has mostly stayed at home. The estimates ranged from 18 days to 254. Some behaviours became automatic in under three weeks, others looked set to take the best part of a year, and for some the curve had not levelled off by the end of the study at all, so the longer figures were projections.
With a spread that wide, the median tells you very little about your own case. It summarises a broad distribution of people, and using it as a personal target is a bit like buying trousers by the average height.
One further finding deserves more attention than it gets. Missing a single opportunity to perform the behaviour did not materially affect how the habit developed, so a lapse did not send anyone back to the start of the curve.
What decides whether a habit forms quickly or slowly
Where you land in that range seems to depend more on how the habit is set up than on how much willpower you bring to it. Two features of the set-up stand out.
How complex the behaviour is. A simple action attached to an existing routine tends to become automatic fairly quickly. Something that needs preparation, travel or several steps takes much longer, because every step is another point at which the sequence can break before it has had time to settle.
How reliable the cue is. Wood and Neal describe habits as triggered by context cues and running relatively independently of current goals. For someone building a habit, the useful part is the first half of that: the cue is what eventually carries the behaviour, so a habit tied to something that happens every single day has more chances to bed in than one tied to something that happens on most days.
Both are design choices, which makes them encouraging. If a habit is forming slowly, the productive thing to check is whether the behaviour is too complicated or the cue too unreliable. How badly you want it is rarely the thing that needs fixing.
Why holding a deadline in mind makes quitting more likely
The bigger problem with a fixed timetable is what having a date does to you when it arrives, even beyond the fact that the number is wrong.
If day 66 is the target, day 67 brings a question that never needed asking, which is whether the thing has worked. Automaticity arrives gradually and never announces itself, so on any particular day the honest answer is usually “not really”. The behaviour still takes a little deliberate effort, and against a deadline that promised it would not, that effort reads as failure.
The number creates a moment of evaluation that will probably go badly, and it places that moment exactly where carrying on is the only thing that matters. People give up at that point who would otherwise have continued indefinitely, simply because they had a date in their head.
Why a habit that still takes effort is still working
There is a more basic objection to the whole idea of a timetable, and it is worth stating plainly. What you wanted was for the behaviour to happen, and whether it also comes to feel easy is a separate matter. Only the first of those is under your control.
A practice that still needs a small deliberate push after four months is doing its job. It is producing whatever it was meant to produce, and the fact that it has not become invisible tells you something about how complex the behaviour is. It says nothing about your commitment. Judging a habit by how it feels, instead of by whether it happens, is how a perfectly good practice gets thrown away as a failure.
How to design a habit that has no finish line
Once you accept there is no finish line, the design changes, and the three changes that matter most all involve taking something away.
Make it small enough to do badly. Aim for under a minute for something like a daily note, or two minutes for the first step of a larger task. Anything that needs energy will only happen on days when you had some, and those are the days you need the habit least.
Hang it on something that already happens. You cannot know when the behaviour will start running by itself, so for the weeks or months before that, something else has to prompt it. Gollwitzer’s research on implementation intentions found that if-then plans fixing when, where and how you will act improve follow-through compared with holding the goal alone. An if-then plan tied to a fixed moment, such as the kettle boiling or the laptop closing, does the prompting during the long stretch before automaticity arrives, whenever that turns out to be.
Leave the counting out. Keep no tally of consecutive days. A count turns a missed day from a plain fact into a loss, and with formation times running anywhere from 18 to 254 days, you will miss plenty of days long before anything feels automatic. Counting makes the ordinary course of forming a habit feel like repeated failure.
That third point follows most directly from the evidence. If forming a habit can take months and varies so widely, a system that penalises gaps will go on penalising you for the whole of that time, which is why returning after a gap has to cost nothing.
What to keep an eye on instead of the day count
With the date set aside, two simple observations are worth making in its place.
Whether it happened today. Record a yes or a no and nothing more. Over a fortnight that shows you the pattern without asking you to pass judgement on yourself.
Whether the cue is still there. When a practice lapses, the cause is usually that its trigger has changed: a routine has ended, a desk has moved, or the evenings have got later. Spotting that takes a few seconds and so does the fix, but you will miss it entirely if you are watching a day count.
Neither of these has a completion date attached, and that is deliberate. A practice with no arrival point cannot be failed on a technicality, and adherence is the only variable that ever mattered.
The Do One Thing diary is five printed pages with one line a day, no rollover and nothing to count. It carries no day numbers and never declares a habit formed, because that moment does not arrive on a schedule, and waiting for it is often how people stop.
Nobody knows how long your habit will take to form, including the researchers who measured other people’s. Whether it happens today is the part within your reach, and it is enough to be going on with.
