How Long Does It Take to Form a Habit? (It Isn't 21 Days)

Updated 2026-10-20

It takes longer than 21 days. The most cited study, by Phillippa Lally and colleagues at University College London, found that a new daily behavior took 66 days on average to feel automatic. The individual range was huge, from 18 days to an estimated 254, so 66 is a midpoint and not a deadline.

That is good news in a practical way. If day 10 still feels like hard work, nothing is wrong with you. You are at the normal place in a process that takes most people a couple of months. The rest of this guide covers where "21 days" came from, what the research did and did not show, what makes a habit faster or slower, and what to do while it still feels hard.

Where the "21 days" idea came from

The 21-day rule is usually traced to Maxwell Maltz, a plastic surgeon who wrote the 1960 self-help book Psycho-Cybernetics. Maltz noticed that his patients seemed to need about 21 days to get used to a new face, and he suggested the same minimum for changing how you see yourself. USC psychologist Wendy Wood has explained that this was a claim about self-image, as reported by NBC News. Over decades of retelling, "about 21 days at minimum" hardened into "21 days to build any habit."

Nobody measured habits in that story. It was an observation in a book, repeated until it sounded like science. That is why it keeps failing in real life: you do something for three weeks, it still takes effort, and you decide you are the problem.

What the 66-day study actually found

The study is Lally, van Jaarsveld, Potts and Wardle, "How are habits formed: Modelling habit formation in the real world," published in the European Journal of Social Psychology (volume 40, 2010; first published online in 2009). You can find it through its DOI. The publisher's site sometimes blocks automated visitors, but the record is registered and the journal page loads in a normal browser.

Here is what the researchers did, as summarized by the British Psychological Society's Research Digest:

96 volunteers each picked one new healthy behavior, such as eating a piece of fruit with lunch or going for a run, and tied it to a daily cue.
They did it every day for 84 days and rated each day how automatic it felt. That feeling of "I just do this without thinking" is what psychologists mean by a habit.
The researchers fitted a curve to each person's ratings to see when it leveled off.

The results:

Average of 66 days to reach the plateau, among people whose data fit the model well.
A range of 18 to 254 days. The 254 figure is a statistical estimate beyond the 84 days of the study, not something anyone was observed doing.
About half of the volunteers had not reached their plateau by the end of the 84 days.
Automaticity rose fast at first and then flattened. The early repetitions gave the biggest gains.

What the study did not show

Be honest about the limits, because they matter if you plan around the number.

The volunteers were mostly university students and staff, and the sample was small. The study has been influential but a large replication has not been widely reported yet.
Automaticity was self-reported. People rated their own feelings.
The behaviors were simple, healthy and chosen by the volunteers. A habit you are forced into, or one that involves a lot of steps, may behave differently.
It tells you about averages. It cannot tell you your number in advance.

A 2024 review in the journal Healthcare by Singh and colleagues pooled 20 studies on health-behavior habits. Only four reported time to form a habit. Those gave medians of about 59 to 66 days, means of 106 to 154 days and individual times from 4 to 335 days, and most of the included studies had a high risk of bias. You can read it in PubMed Central. The takeaway is the same as the original: roughly two months is typical, and a wide spread around it is normal.

Is it 21 days or 66?

Plan for about two months and treat 21 days as the point where things start to get easier, not the finish line. 21 days has no research behind it as a habit rule. 66 days is one study's average, so use it as a planning figure and expect your own to land anywhere from weeks to many months.

Number	Where it comes from	How to use it
21 days	A 1960 self-help book about self-image	Ignore it as a rule
66 days	Average in one 2009 to 2010 UCL study	A sensible planning figure
18 to 254 days	Range of individuals in that study	Expect your own to differ
About 59 to 66 days (median)	Four studies in a 2024 review	Confirms "around two months"

What counts as "formed"?

In the research, a habit is formed when the behavior is triggered mostly by a cue and needs little effort or deliberation. Gardner, Lally and Wardle describe this in the British Journal of General Practice: repeating a simple action in a consistent context strengthens the link between the cue and the action, so it relies less on motivation (Making health habitual, 2012).

A few signs you are getting there:

You start the habit before you have decided to.
Skipping it feels slightly odd.
You do it on tired or busy days without a negotiation.
You are more likely to forget the reasons than the action.

"Formed" does not mean "can never be broken." A holiday, illness or a new schedule can loosen a habit, and you rebuild the cue and carry on. It also does not mean the habit is easy forever. It means the effort has dropped to a level you can sustain.

What makes a habit take longer or shorter

The study and later work point to a handful of factors. None is a guarantee, but they are the levers you can control.

Size and complexity

Simple actions become automatic sooner. In the Lally study, new exercise habits took about one and a half times as long to level off as new eating or drinking habits. So a glass of water with breakfast will likely settle faster than a 30-minute workout, and a workout faster than "write every morning, plan the day and meditate."

If you are building something big, start with the smallest version that still counts. Two minutes of stretching is a real habit. You can lengthen it later once the cue is solid. The how to build good habits guide goes deeper on making habits tiny.

A steady cue

Habits attach to a context: a time, a place or something you just did. The more consistent the context, the faster the link forms. Pairing a new habit with an existing one ("after I pour my coffee, I will take my vitamins") gives you a cue that already happens every day. The habit stacking examples guide has lots of them by time of day.

Daily repetition in the same place

Gardner and colleagues note that varying the behavior or setting may reduce boredom but works against automaticity. Early on, boring and repeatable beats interesting and different.

How many habits you start at once

Each new habit needs its own cue and its own attention while it is fragile. Starting one to three gives each a fair chance. Starting eight tends to mean all eight wobble at once. This is a practical point, not a finding from the 66-day study, but it is a common way people stall.

Motivation and personal factors

The wide range between individuals means some people simply take longer. Stress, sleep, your schedule and how much you care about the habit all play a part. Comparing your day 30 to someone else's day 30 tells you very little.

Does missing a day reset the clock?

No. One of the more reassuring results in the Lally study is that a single missed day made little difference to the later gains in automaticity. The researchers did note that missing many days, even spread out, lowered the level that people reached. So the pattern that matters is how often you show up over weeks, not whether the chain is perfect.

A practical rule many people find useful is to never miss twice in a row. That is a rule of thumb, not a research result, but it fits the finding: one slip is small, and a run of slips is the real risk. If you want more on restarting without drama, the missed a day guide covers it, and you can see how long a streak has run with the habit streak calculator.

How long for different habits?

Nobody has a table for every habit, and you should be suspicious of anyone who does. What the research supports is a direction, not exact numbers.

Small, easy habits such as drinking water, taking vitamins or flossing tend to settle sooner.
Habits with more steps or more effort, such as exercise, tend to take longer. In the Lally study they took about one and a half times as long.
Reading and waking early sit in between and depend on you. A page a night is tiny. Getting up an hour earlier changes your whole evening, so it is closer to a big habit. Shrink it ("get up 10 minutes earlier this week") and it behaves like a small one.

The study also does not tell you how long it takes to break a habit. The 66 days applied to building a new behavior, so don't assume the same number for stopping one. Gardner and colleagues suggest that new, small behaviors are an easier place to start than trying to remove existing ones.

What to do on day 10 when it still feels hard

Day 10 is where many people quit, and it is exactly where the research says you would expect effort to still be high. Automaticity climbs fastest early on but you won't notice it yet. A few things help.

Lower the bar, not the frequency. If you only have the energy for the two-minute version, do that. Showing up keeps the cue alive.
Check the cue. If you keep forgetting, the cue is the problem, not you. Move the habit to something you never skip, like brushing your teeth or making coffee.
Make it visible. Seeing a row of ticks gives you evidence of progress that your feelings won't. A paper grid works as well as an app. The printable habit tracker gives you a monthly grid for up to eight habits.
Plan the slip. Decide now what you will do after a missed day. "Next time my cue happens, I do it" is enough.
Pick a check-in date. Put a reminder on day 30 and day 66 to look back at how it feels. You are judging a trend over weeks.
Don't add a second habit yet. Wait until this one feels lighter.

How Atomic Streak can help

Atomic Streak is a habit tracker for Android and iPhone made by Winkle, the same team that makes this site. It supports the things the research says matter, with some honest limits.

A cue and a reward for each habit. You write when you will do it and why it is worth it, such as "After I eat breakfast, I will take my vitamins." That is the cue-based approach described above.
Daily or chosen weekdays. Repeat a habit every day, or only on the days you pick, with a reminder time for each one.
60 ready-made habits across 11 categories, so you can start small without inventing something.
Streak calendars by week, month and year, plus your current and longest streaks, so one missed day doesn't erase the evidence of the rest.
Statistics and streak badges from 1 up to 150 days, which give you milestones to reach on the way to the 66-day mark and beyond.
No account needed. It is free to download. The free version tracks up to 3 active habits, which suits the advice to start small, and some features need Pro.

What it doesn't do: it doesn't measure how automatic a habit feels, and it can't tell you when a habit is "formed." The 66-day average is a guide, not something the app calculates for you. You will judge that yourself, using the signs above. If you are comparing options, the best habit tracker apps guide covers what to look for, or you can start from the Atomic Streak home page.

What the research suggests you do next

Pick one small habit, ideally one that takes under two minutes.
Attach it to a cue you hit every day.
Do it daily for the first weeks, in the same place if you can.
Track it somewhere visible.
Check back at day 30 and day 66, and keep going if it is not automatic yet.

If a habit relates to a health condition, medication or an eating or exercise change, check with your doctor or another qualified professional before changing your routine. This guide is general information and not medical advice.

Questions

Is it 21 days or 66?

Neither is a rule, but 66 has real data behind it and 21 does not. The 66 figure is the average from one UCL study, with individuals ranging from 18 to 254 days. Plan for about two months.

Does it take 66 days to break a habit?

Not necessarily. The study looked at forming new behaviors, not breaking old ones, so the number does not carry over. Many people find it easier to add a replacement behavior in the same cue than to remove one.

Does missing one day reset the clock?

No. In the Lally study, one missed day did not meaningfully slow progress. Missing many days, even scattered, lowered how automatic the habit got. Get back to it at the next chance. The missed a day guide goes through what to do.

How long does a small habit like drinking water take compared with exercise?

Simpler habits usually settle faster. In the study, exercise habits took about one and a half times as long as eating or drinking ones to level off. Your own numbers will vary, so use that as a direction, not a timetable.

Why do some habits take 250 days?

The 254-day figure was a statistical estimate for the slowest individuals, not an observed result, but the spread is real. Complex behaviors, unstable routines, a weak cue and many missed days can all stretch it out. A slow habit is not a failed one.

What if I'm past 66 days and it still takes effort?

That is normal, since about half of the volunteers had not leveled off by day 84. Check the cue, shrink the habit and keep the schedule steady. If it feels like a chore for other reasons, the habit may be the wrong size or the wrong time of day. The how to build good habits guide has a five-step fix.

The Atomic Streak makers are not affiliated with James Clear or his book Atomic Habits. This site and app are independent, and the book is mentioned only because people often look for both.