Books and philosophies

Does it take 21 days to form a habit? Where Maltz's number came from

No. The 21-day figure comes from plastic surgeon Maxwell Maltz, who wrote in Psycho-Cybernetics (1960) that it seemed to take a minimum of about 21 days for people to adjust to a changed self-image. He was describing patients getting used to a new face, not measuring how long a new behavior takes to become automatic. When researchers did measure that, the median was about 66 days, with very wide variation between people and behaviors. Treat day 21 as a first review point, not a finish line.

What Maltz actually claimed

Maxwell Maltz was a plastic surgeon. In the preface to Psycho-Cybernetics he described a pattern he said he saw in his practice: after facial surgery, patients seemed to need about three weeks to get used to their new appearance, and people who lost a limb often reported sensing it for a similar period. From observations like these he concluded that it “requires a minimum of about 21 days” for an old mental image to dissolve and a new one to settle.

Three details matter, and most retellings drop all three:

  1. It was a minimum, not a deadline. Maltz said “a minimum of about,” which leaves room for much longer.
  2. It was about self-image, not behavior. He was describing how long it takes to accept a changed picture of yourself, not how long it takes for flossing or running to feel automatic.
  3. It was clinical observation, not a study. There was no measurement, control group or count of patients behind the number.

Over the following decades, self-help writing compressed this into “it takes 21 days to form a habit.” That sentence is not what Maltz wrote, and it is not what later research found. If you want the rest of his method, our Psycho-Cybernetics practice guide covers the mental-movie exercise he is better known for.

What researchers found when they measured it

In 2010, Phillippa Lally and colleagues at University College London asked 96 volunteers to pick one daily eating, drinking or exercise behavior tied to a cue (for example, “drink a glass of water after breakfast”) and to rate each day how automatic it felt, for 84 days (Lally and colleagues, 2010).

The headline result is the 66-day median. The details are more useful:

  • A clear curve could be fitted for only part of the group. Among those with a good fit, the median time to reach 95% of their own automaticity plateau was 66 days, with a range of 18 to 254 days. The upper estimates are projections, because the study stopped at day 84.
  • Automaticity rose fastest early on and then leveled off. The first few weeks felt like the biggest change, even though the plateau came much later.
  • Missing one day did not significantly affect the process. People who were very inconsistent did not form habits.
  • Exercise behaviors tended to take longer than eating or drinking ones.

A 2024 systematic review pooled 20 studies with 2,601 participants and concluded that health habits can begin forming within about two months, with large individual differences. Only four studies reported formation time directly; their medians ran from 59 to 66 days, and individual estimates ranged from 4 to 335 days (Singh and colleagues, 2024). The authors rated many of the included studies at high risk of bias, so these are descriptive figures, not a rule.

Two clocks, not one

Here is the practical synthesis. Maltz and Lally were measuring different things, and you can use both.

QuestionWhat it measuresUseful time frameWhat to check
”Does the new version of me feel strange?”Adjustment to a changed self-picture (Maltz)A first look around day 21Whether the new routine still feels like pretending
”Do I do it without deciding?”Behavioral automaticity (Lally)Often two to three months, sometimes far longerWhether the cue alone starts the behavior
”Am I still doing it?”ConsistencyEvery weekHow many planned days actually happened

This explains a common experience. Around week three, a new routine often stops feeling foreign, and people assume the habit is done. Then a busy week arrives and it disappears, because it had become familiar but not yet automatic. Familiar and automatic are different milestones.

A 12-week automaticity tracker

Researchers often measure automaticity with four short statements from the Self-Report Behavioural Automaticity Index, which asks whether a behavior is something you do automatically, without having to consciously remember, without thinking, and before you realize you are doing it (Gardner and colleagues, 2012). You can borrow that idea as a weekly self-check. This is a personal tracking aid, not a validated scoring of your own data.

Once a week, rate each statement from 1 (not at all) to 5 (completely) for your one behavior:

  • I did it without having to remember to do it.
  • I did it without much thought.
  • I started before I had decided to.
  • It would feel odd to skip it.
WeekDays done / plannedFour ratings (1–5)Total (4–20)One note
1/
3 (Maltz check)/Does it still feel like pretending?
6/
9/
12 (Lally’s study length)/Keep, adjust or drop?

Worked example (illustrative, not a real participant). Someone plans a five-minute evening wind-down after brushing their teeth. Week 1: 4 of 7 days, total 7. Week 3: 6 of 7 days, total 12, and the note says “no longer feels like a performance.” Week 6: a trip breaks the routine, 3 of 7 days, total 10. Week 9: back to 6 of 7, total 15. Week 12: total 17.

Reading this record: the week-3 check passed Maltz’s test (it felt normal) but not Lally’s (a trip still knocked it out). The dip at week 6 is the kind of disruption that tempts people to restart a “21-day challenge” from zero. The tracker shows there was no need. One rough week did not erase the gain, which matches Lally’s finding about missed days.

What to change when the curve stalls

  • Low days done: the cue is weak. Tie the behavior to something that already happens every day.
  • Days done are high but ratings stay low: the behavior varies too much. Do it the same way, in the same place, after the same cue.
  • Ratings drop after a disruption: expected. Resume at the next cue rather than restarting a count. Our guide to self-talk after missing practice days has wording for that moment.
  • It feels foreign after a month: this is closer to Maltz’s question. Rehearse the new self-picture as well as the behavior.

Rehearse the self-image part with Manai

Maltz’s own method was imaginative rehearsal: picturing yourself acting from the new self-image before doing it. That fits the “does it feel like me?” clock. You can create a short rehearsal in Manai that describes your cue, the behavior and how you want to carry yourself, then replay it daily alongside the real practice. After each week-three or week-six review, revise the script to match what actually happened.

Original illustrative request, not a tested example:

Make a four-minute rehearsal for my evening routine. The cue is turning off the bathroom light after brushing my teeth. Walk me through sitting on the edge of the bed, five slow breaths, and writing one line about the day. Speak to me as someone for whom this is normal, without saying I have already changed. Leave a quiet pause after each step. No mention of 21 days.

A recording does not form the habit for you, and no audio rewires your brain on a schedule. It gives you a consistent way to rehearse the self-picture while repetition does the slower work of automaticity.

Frequently asked questions

Is the 21-day habit rule a myth?

As a rule for how long habits take, yes. Maltz described a minimum of about 21 days for people to adjust to a changed self-image, based on what he observed in surgical patients. Measured habit studies found a median around 59 to 66 days, with individual times from a few weeks to many months.

Did Maxwell Maltz say it takes 21 days to form a habit?

Not in those words. He wrote that it seemed to require a minimum of about 21 days for an old mental image to dissolve and a new one to form. The habit version came later, as his idea was repeated and simplified in self-help writing.

How long does it really take to form a habit?

There is no single number. In Lally’s 2010 study the median was 66 days among participants whose data fitted well, ranging from 18 to 254 days. Simpler behaviors, a stable cue and consistent repetition tend to help; complex behaviors like exercise tended to take longer.

Does missing a day reset a habit?

No. Lally’s study found that missing one opportunity did not significantly affect habit formation. What prevented habits was being very inconsistent over time. Resume at the next cue instead of restarting your count.

Is 21 days still useful for anything?

Yes, as a first checkpoint. Around three weeks is a reasonable time to ask whether the routine still feels foreign and whether the cue is working. Just do not treat it as the point where the habit is finished.

Create your personal practice with Manai →

A note on wellbeing

Manai offers educational wellness practices. It is not medical care, diagnosis, or treatment. If a practice feels uncomfortable, stop. Seek qualified support for health concerns or urgent distress.