The 21-Day Habit Myth: What Maltz Said and What Research Found

The idea that it takes 21 days to form a habit traces back to Maxwell Maltz, the plastic surgeon who wrote Psycho-Cybernetics in 1960. Maltz never said that. He wrote that an old mental image usually needs a minimum of about 21 days to fade and a new one to set, based on what he noticed in his patients. Habit research has since measured how long new behaviours take to become automatic, and the answer is longer and far more variable: a median of 66 days in the best-known study.

This page traces where the number came from, how it changed as it travelled, and what research has found about how long change takes. It explains the idea and the evidence. It doesn’t set out a habit plan.

Where does the 21-day habit idea come from?

The source is the preface of Psycho-Cybernetics. Maltz asked readers not to judge his methods too quickly. Keep practising for at least 21 days, he wrote, before deciding whether anything had changed.

He gave three everyday observations as his reasons. Surgery patients took about three weeks to get used to a new face. Amputees’ phantom limb sensations lasted about three weeks. People needed about three weeks in a new house before it felt like home. From observations like these he concluded that an old mental image needs a minimum of about 21 days to dissolve and a new one to set.

The number then runs through the book as a program length. He asks readers to reread a chapter several times a week for the first 21 days, and in his chapter on happiness he suggests acting out new habits of thought for 21 days to see what changes.

Two things stand out. First, Maltz was talking about the self-image, the picture you hold of the kind of person you are, not about habits in general. Second, he called it a minimum. He also wrote that building a self-image goes on for life.

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

Not in those words, and not with that meaning. His claim was narrower and more modest: give a new self-picture at least three weeks before you expect to notice it.

He also offered no study behind the figure. It came from a surgeon’s observations, reported the way a clinician reports what he sees in his practice. That’s a fair basis for a hunch and a weak basis for a rule. One of his own examples works against it. Elsewhere in the same book he mentions that phantom limb pain can persist for years after an amputation.

How did “a minimum of 21 days” become “21 days to form a habit”?

Through retelling. Each time the line passed through another speaker, seminar or article, it lost a qualifier. “A minimum of about 21 days” became “21 days”. “An old mental image” became “a habit”. The tentative observation became a fact.

The number also kept travelling into new fields, and picked up new claims on the way. Teppo Holmqvist’s Practical Hypnosis (2017), a manual written by a former professional hypnotist, uses 21 days throughout. It appears as the time needed to rehearse a new belief, as the time it takes to make any behaviour instinctive, and as a ceiling. Holmqvist presents it as a settled finding of neuroscience and warns that rehearsing longer than 21 days makes behaviour rigid and harder to change. He cites no study for either part. A sample session at the back of the book even tells the listener that their mind believes it has already rehearsed 21 times. By then the number had become a suggestion as well as a claim.

So a surgeon’s observation from 1960 turned up nearly sixty years later in a hypnosis manual as a brain-science result with an upper limit attached. Nobody along the way seems to have measured anything. The hypnosis guide covers the rest of that manual’s claims.

How long does it actually take to form a habit?

In 2010 Phillippa Lally, Cornelia van Jaarsveld, Henry Potts and Jane Wardle at University College London published the study most people now cite. Volunteers each chose one new healthy behaviour, such as drinking a glass of water with lunch, eating a piece of fruit or going for a short run, and did it once a day in the same situation. Every day they rated how automatic the behaviour felt for 12 weeks.

The results, published in the European Journal of Social Psychology, looked nothing like 21 days:

• Automaticity rose quickly at first and then levelled off. Each repetition added less than the one before.
• Among the people whose results fitted that curve, the median time to reach 95 per cent of their plateau was 66 days.
• Individual times ranged from 18 to 254 days. Several of the longer times were projected from the curve, beyond the 12 weeks people were tracked.
• Exercise habits were slower to become automatic than eating or drinking habits.
• Missing a single day didn’t measurably set anyone back.

Later work points the same way. A 2024 systematic review led by Ben Singh at the University of South Australia pooled about twenty studies of health habits and found typical times of around two months, with some people reaching automaticity within a week and others taking most of a year.

The honest answer to “how long does it take?” is “it depends, and usually longer than three weeks”. “66 days” is no more of a law than 21. It describes the middle of a wide spread.

Why do some habits take longer than others?

The research identifies a few reasons, and none of them has to do with willpower.

Complexity. A glass of water at lunch is one small act. A morning run involves several decisions, some discomfort and some planning. The more steps a behaviour has, the longer it takes to run on its own.

A stable cue. Wendy Wood and David Neal’s work on habits (2007, Psychological Review) describes a habit as a response the situation triggers before you think about it. Behaviours tied to the same place, time or preceding action become automatic faster than behaviours done whenever there’s a spare moment. Benjamin Gardner, Lally and Wardle made the same point for doctors giving health advice in 2012.

Disruption. Bas Verplanken and colleagues found that big life changes, such as moving house, loosen old habits and give a window for new ones. Maltz’s new-house example fits that research, though for a different reason than he gave.

Individual difference. People differ a lot in how fast any behaviour becomes automatic, for reasons studies haven’t fully explained.

Were Maltz’s examples right?

Partly. The new-house feeling is real enough as experience, and research on habit disruption gives it a sound basis. The new-face example is hard to check. Maltz reported it from his own practice, and he also wrote that some patients went on feeling inadequate long after a successful operation.

The phantom-limb example is the weakest. Studies of people who have had a limb amputated find that most still have phantom sensations, and many have phantom pain, years after surgery. A 2000 survey of people with lower-limb amputations led by Dawn Ehde found phantom sensations in roughly four out of five. Three weeks is not a typical timeline.

Does it matter if the number is wrong?

It matters most in the way people use it. As a trial period, three weeks is a sensible length. It’s long enough to get past the first enthusiasm and short enough to commit to. Many good programs, including the guided work on this site, use 21- or 30-day blocks.

Trouble starts when day 21 becomes a finish line. If someone believes a habit should be automatic by then and it still takes effort, the natural conclusion is “this doesn’t work for me” or “I lack discipline”. Lally’s data suggest the more accurate conclusion: they’re probably about a third of the way there. Belief in a deadline can turn a normal stage of change into a reason to quit.

Maltz’s own framing protects against that. A minimum is a reason to keep going, not a promise that you’ll be finished.

What does the research say?

Claim	Who made it	What research shows
An old mental image needs a minimum of about 21 days to fade	Maltz, 1960	Untested. Based on clinical observation, and he called it a minimum
It takes 21 days to form a habit	Later retellings of Maltz	Not supported. Median of 66 days, range 18 to 254 (Lally et al. 2010)
21 days makes any behaviour instinctive, and longer makes it rigid	Holmqvist	Not supported. Automaticity rises and then levels off; nothing suggests longer practice removes choice
Phantom limb sensations fade in about 21 days	Maltz	Not supported. Most amputees report them years later (Ehde et al. 2000)
Missing a day breaks a new habit	Popular advice	Not supported. A single missed day made no measurable difference (Lally et al. 2010)
A consistent cue and setting speed habit formation	Habit researchers	Supported (Wood and Neal 2007; Gardner, Lally and Wardle 2012)
Major life changes open a window for new habits	Habit researchers	Supported (Verplanken and colleagues)

Who should take care with habit timelines?

Some behaviours shouldn’t be approached as a 21-day or 66-day project at all.

Dependence on alcohol, drugs, gambling or gaming. These involve more than habit, and stopping some substances abruptly can be medically dangerous. Talk to a doctor or addiction service before changing them.

Eating. Strict eating challenges can feed disordered eating. Anyone with a history of an eating disorder should make changes with professional support.

Compulsions. Behaviours driven by anxiety or obsessive thoughts aren’t bad habits, and pushing against them alone can make them worse. A licensed therapist trained in these conditions is the right place to start.

Low mood. If you can’t keep up a new routine because you feel flat, exhausted or hopeless, that may be depression, not a discipline problem. Talk to a doctor.

What does this mean for changing who you are?

Under the myth sits Maltz’s real argument, and it holds up better than his number. People act in line with the picture they hold of themselves. Change the picture, and new behaviour starts to feel like “me” instead of a performance. That process runs on repetition, and it runs on its own schedule. You can read more about it in the identity pillar and in the pages on the success mechanism and imagination and the nervous system. The pages on mental rehearsal research and the habit of happiness fill in more of Maltz’s picture.

The useful question isn’t “how many days?” It’s “what will I do on the day it still feels hard?” Building a new self-image over the months it takes, with someone checking your progress, is the work we do in coaching.

Frequently asked questions

Does it really take 21 days to form a habit?

No study supports a fixed 21 days. In the best-known study, by Phillippa Lally and colleagues in 2010, new daily habits took a median of 66 days to become automatic, and individual times ranged from 18 to 254 days.

Where did the 21-day habit idea come from?

It traces back to Maxwell Maltz's 1960 book Psycho-Cybernetics. Maltz, a plastic surgeon, wrote that an old mental image takes a minimum of about 21 days to fade, based on what he saw in patients. Retellings shortened that into '21 days to form a habit'.

What is the 66-day rule?

It comes from the median result in Lally's 2010 study at University College London. It's the middle value across widely different people and behaviours, so it's no more a rule than 21 days. Simple habits got there faster and exercise habits took longer.

Does missing a day ruin a new habit?

Not in Lally's study. Missing a single opportunity didn't measurably set people back. What mattered was keeping the behaviour going in the same situation over time.

Why do some habits take longer to form than others?

Complex behaviours, such as exercise, take longer than simple ones, such as drinking a glass of water. A stable cue and setting speed things up, and people differ widely in how fast any behaviour becomes automatic.

Are 21-day challenges useless, then?

Not as a way to start. Three weeks is a reasonable trial period. The problem comes when people treat day 21 as a finish line, stop, and conclude they failed when the behaviour still takes effort.

This article is educational and is not medical, psychological or financial advice. Talk to a qualified professional before making major changes, especially if you live with a mental health condition. Read the full disclaimer.