How Long Does It Take to Form a Habit?

Ask how long it takes to form a habit and you will usually be told 21 days. That number is confidently repeated, widely printed, and not supported by any study of habit formation. The real answer is less tidy, considerably longer, and far more useful once you understand where it comes from.

The short answer

The most cited empirical estimate comes from a 2010 study by Phillippa Lally, Cornelia van Jaarsveld, Henry Potts and Jane Wardle, published in the European Journal of Social Psychology under the title How are habits formed: Modelling habit formation in the real world.

Participants chose a new eating, drinking or activity behaviour to perform once a day in the same context, then reported daily for twelve weeks on how automatic it felt. Among the participants whose data could be fitted to a curve, the modelled time to reach a plateau of automaticity had:

• a median of about 66 days, and
• an individual range from 18 to 254 days.

So: roughly two months for a typical simple behaviour, with a spread wide enough that the median tells you very little about your own case.

What that study actually measured

The details matter, because the 66-day figure is quoted far more often than it is understood.

It measured self-reported automaticity, not adherence. Participants rated statements about whether the behaviour felt automatic — something they did without thinking, without having to remember. The study was tracking how effortless the behaviour became, not whether people were still doing it a year later.

It used simple, once-daily behaviours in a fixed context. Things like drinking a glass of water with lunch, or walking for fifteen minutes before dinner. The design deliberately held the cue stable.

Complexity changed the timeline. Behaviours that required more effort took longer to become automatic than trivially easy ones. That is one of the more actionable findings in the paper, and it argues directly for starting smaller than you intend to finish.

A single missed day did not derail the process. Participants who missed one opportunity showed no meaningful disruption to the overall trajectory. This is probably the single most practically useful result in the study, and it is the one least often repeated.

What that study does not say

Being precise here is worth the effort, because the 66-day number gets stretched a long way past what it can support.

It is not a deadline. Nothing switches on at day 66. Automaticity accumulates gradually and then flattens; the curve has no cliff edge.

It is not a guarantee. Some participants never reached a plateau within the study window at all. The published range covers those whose data fitted the model.

The longer estimates are extrapolations. The study ran for twelve weeks, or about 84 days. Any modelled estimate beyond that — including the upper end of the range — is the fitted curve projected forward, not an observation of someone at day 254.

It was a small study of self-chosen everyday behaviours. It does not tell you how long it takes to form a demanding habit like a daily hour of study, and it makes no clinical claims about changing entrenched behaviour.

Treat 66 days as a well-founded order of magnitude — months, not weeks — rather than as a precise constant.

Where the 21-day myth came from

The 21-day figure traces back to Psycho-Cybernetics, a 1960 book by Maxwell Maltz, a plastic surgeon. Maltz observed that his patients seemed to need a minimum of about 21 days to get used to a change in their appearance, and that people who had lost a limb reported a phantom sensation for roughly the same period. He wrote that it appears to require a minimum of about 21 days for an old mental image to dissolve and a new one to form.

That is a clinical observation about psychological adjustment to a changed body. It was not an experiment, it did not concern habit formation, and Maltz described it as a minimum rather than a typical duration.

The book sold in enormous numbers. Over the following decades, “a minimum of about 21 days for a mental image to change” was compressed, repeated and eventually hardened into “it takes 21 days to form a habit”. The qualifier disappeared, the subject matter changed, and the claim acquired a scientific authority it never had.

Believing it has a real cost. If you expect the behaviour to feel automatic after three weeks and it does not, the natural conclusion is that something is wrong with you rather than with the expectation. That is a common point at which people quit — three or four weeks in, which is exactly when the research suggests you should still be finding it effortful.

Why the range is so wide

Between 18 and 254 days is not measurement noise. Several things genuinely move the number.

Behaviour complexity. Adding a glass of water to a meal you already eat is one step. Getting to a gym involves a change of clothes, travel, and a series of decisions, each of which needs its own repetition.

Cue stability. A behaviour anchored to something that happens at the same point in every day builds an association quickly. A behaviour that floats — sometimes morning, sometimes evening, sometimes not at all — is being relearned each time in a new context.

How often the opportunity comes round. Automaticity is built by repetitions, not by calendar days. A daily habit gets seven learning events a week; a three-times-a-week habit gets three. In elapsed time, the daily version will get there sooner.

How pleasant it is. A behaviour with an immediate reward is easier to repeat than one whose entire payoff sits months in the future, which is why flossing is harder to establish than a morning coffee.

Individual variation. Even holding all of the above constant, people simply differ. The median exists to describe a group; it makes no promises about you.

What “formed” even means

Part of the confusion is that “formed” is doing a lot of unexamined work in the question.

If it means I no longer have to decide, that is automaticity, and the two-month scale is a reasonable expectation for a simple daily behaviour.

If it means I will keep doing this indefinitely, no study can give you a date. Long-term maintenance depends on whether the context that supports the habit survives — a house move, a new job or a new baby can disrupt a behaviour that felt entirely settled, precisely because habits are tied to the situation in which they were learned.

If it means it feels easy, expect that to arrive gradually and unevenly. It tends to be noticed in retrospect rather than on a particular morning.

What to do with an answer this vague

The number is not actionable. The shape of the curve is.

Plan in months. Commit to a timescale of eight to twelve weeks before you evaluate whether an approach is working. Three weeks is too early to conclude anything.

Reduce the behaviour so it can be daily. Frequency of repetition is the variable you control most directly, and it is the one that shortens the timeline. There is more on that in how to build a habit.

Do not count down to a finish line. Waiting for day 66 sets you up for disappointment on day 67. Watch the trend instead.

Expect week three to be the low point. The novelty is gone and the automaticity has not arrived. That is the ordinary shape of the process, not evidence of failure.

Track the process, not the countdown

Because automaticity accumulates gradually, a measure that reflects accumulated repetitions is more informative than any calendar target — and much more informative than a consecutive-days counter that resets to zero after one bad evening.

This is the reasoning behind the habit strength score in the free browser-based habit tracker. It combines total completions with your consistency over the last four weeks, and maps the result onto a formation stage from Starting through to Mastering. The repetition term is scaled so that the higher stages sit in the range the Lally study associates with high automaticity, which keeps the feedback roughly in step with the evidence rather than with a three-week fantasy.

It also means one missed day barely moves the number, which is faithful to the research. A missed fortnight does move it, which is also faithful to the research. The related argument for regularity over effort is in why consistency beats intensity.

A realistic timeline

• Days 1–7. Entirely deliberate. You will need to remember every single time.
• Days 8–21. Novelty fades and the first misses appear. Returning the next day matters more than anything else you do this fortnight.
• Days 22–60. The grind. Nothing feels different, which is why an external record is worth more here than your own sense of progress.
• Around day 60 and beyond. For a simple daily behaviour, the effort of starting begins to drop noticeably. For a more complex one, expect this later — possibly much later.

If you want reminders at a specific time to carry you through the first fortnight, that is one of the things the Ring Habit iPhone app adds on top of the browser tracker. Either way, the timescale to plan for is months.