How Many Days Does It Take to Form a Habit? (Hint: Not 21)

It takes a median of 66 days to form a habit, not 21. That number comes from the best-known study, where a new daily behavior took anywhere from 18 to 254 days to become automatic, depending on the person and the habit.

So if you're three weeks into a new habit and it still feels like work, nothing is wrong with you. Below: how long it takes to build a habit, where the 21-day myth comes from, why your habit may go faster or slower, what a missed day does, and a 66-day plan with a tracker you can print.

How long does it take to form a habit?

It takes a median of 66 days to form a habit, with a range from 18 to 254 days. Those numbers come from a 2010 study led by Phillippa Lally at University College London, which followed 96 volunteers for 12 weeks. A median means half of the people it was calculated for got there faster and half took longer.

Newer research hasn't moved the number much. A 2024 systematic review and meta-analysis by Ben Singh and colleagues covered 20 studies with 2,601 participants. Only four of them reported how long a habit took: medians of 59 to 66 days, means of 106 to 154 days and individual times from 4 to 335 days. We take those numbers apart further down.

So how many days does it take to form a new habit? Plan on about two months for a simple daily habit, and longer for something complex like going to the gym.

Where the 21-day myth comes from

The idea that it takes 21 days to form a habit traces back to a plastic surgeon, not to a habit study. In his 1960 book Psycho-Cybernetics, Maxwell Maltz wrote that it takes "a minimum of about 21 days" for an old mental image to fade and a new one to settle in.

He was describing how long his patients needed to get used to a new face after surgery, and how long a phantom limb seemed to linger after an amputation. Here's how that turned into a rule:

1. 1960: a surgeon's observation. Maltz ran no study, and he wasn't writing about habits like flossing or running.
2. The retelling. As the idea got repeated, the words "a minimum of about" fell away. A floor became a finish line: "it takes 21 days to make a habit."
3. The challenge era. Our reading: three weeks is short, sounds doable and fits neatly on a calendar, a natural length for a challenge.

To be fair, 21 days isn't wrong for every habit: the fastest person in Lally's study needed 18. It just isn't a rule. Neither is the "21/90 rule" (21 days to build a habit, 90 days to make it a lifestyle), a popular rule of thumb with no study behind it.

The myth has a cost. If you expect to be done in three weeks and it's still an effort on day 22, it's easy to decide you failed and stop, while the habit was still climbing toward its plateau.

What the 66-day study actually found

The study behind "66 days to form a habit" found that a daily behavior, repeated in the same context, gets steadily more automatic until it levels off.

The setup: Lally and her colleagues asked 96 volunteers to pick one new eating, drinking or activity behavior. Each person tied it to a daily cue, such as "after breakfast", and tried to do it every day for 84 days.

What "automatic" means

The researchers measured automaticity. Every day, participants rated statements about their behavior, along the lines of "I do it without thinking."

A habit counted as formed when a person's score reached 95 percent of their own plateau. So "formed" means about as automatic as this behavior was going to get for this person, not effortless forever.

The habit formation curve

Automaticity didn't rise in a straight line. In a 2012 summary for family doctors, Benjamin Gardner, Phillippa Lally and Jane Wardle describe the pattern:

Three details from the paper that rarely survive the retelling:

• The 66 comes from 39 people. Of the 96 volunteers, 82 gave enough data to analyze, and the curve fit 39 of them well. The median is theirs.
• 254 days is a projection. The study ran for 84 days, so the longest times are estimates from each person's curve, not something anyone watched happen.
• Exercise looked slower. Exercise habits took about one and a half times as long to level off as eating or drinking habits. But each group was small, so treat it as a hint, not a rule.

What newer research adds

The timing numbers in Singh's 2024 review come from different studies of different habits, so they aren't one pooled estimate. The 66 is Lally's study, counted again, and one of the four studies only asked people to estimate their own time. The other figures come from two small studies:

• A 2021 trial. In a trial led by Jan Keller, 192 adults linked a small nutrition habit to a daily routine or to a set time for 84 days. Of the 117 who followed the protocol, 27 (23 percent) met the researchers' bar for a formed habit. Those 27 took a median of 59 days, with modeled times from 4 to 335 days (anything past day 84 is a projection).
• A 2017 study. In a study led by Marion Fournier, 48 students did a daily stretch for 90 days, after waking or before bed. The curves projected automaticity at about 106 days for the morning group and about 154 days for the evening group: the review's two means.

The review also looked at what helps: morning practice and self-chosen habits each came out stronger in a single study, so treat both as leads, not rules.

Why some habits take 18 days and others 254

How long it takes to develop a habit depends mostly on what the habit is, how often you repeat it and how steady its cue is.

In a 2015 study of 111 new gym members, Navin Kaushal and Ryan Rhodes found that people needed at least 4 workouts a week for 6 weeks to establish an exercise habit. It's one 12-week study, based on surveys.

Time of day is less clear. Fournier's morning group was projected to get there sooner, but that's one small study, and Caltech's summary of the 2023 gym study says time of day had no effect on gym habit formation.

Here's how that adds up.

None of these numbers is a promise for your habit. Five things move your own timeline:

Complexity. In the Kaushal and Rhodes study, simpler workouts predicted stronger habits. What to do: start with the simplest version of your habit.

Frequency. A daily habit collects repetitions faster than a weekly one. What to do: make it daily if you can, even if the daily version is tiny.

A stable context. Same place, same trigger, same days: the trigger is the cue, the first step of the habit loop. What to do: tie your habit to something you already do every day.

Enjoyment. Kaushal and Rhodes also found that people who felt good about their workouts built stronger habits. What to do: pick a version you like, and give yourself a check mark right after.

You. How long it takes to create a habit differs enormously from person to person: from 18 to 254 days in Lally's study. What to do: compare this week with your own last week, not with the median.

Repetitions, not days: what actually builds a habit

There is no fixed number of repetitions that builds a habit. What the research supports is repeating the same behavior after the same cue, many times. The calendar is just how we count it.

In Lally's study people aimed for once a day, so 66 days means at most about 66 repetitions in the same context, and fewer for anyone who skipped days. The study counted days, not repetitions.

In the 2023 hospital data, handwashing became predictable after about 9 to 10 shifts, which was roughly 220 chances to wash: a couple of weeks on the calendar.

A habit you do three times a week collects repetitions less than half as fast as a daily one, so expect it to need more calendar time. No study gives a clean conversion.

Does missing a day reset your progress?

No. In Lally's study, missing one opportunity to do the behavior did not materially affect the habit formation process. Gardner, Lally and Wardle add that the gains in automaticity soon resumed after a single miss.

Long gaps look different. In the 2023 gym data, the longer it had been since a person's last visit, the less likely they were to go on a given day. That held for 76 percent of gym-goers. It's a pattern, not proof that a gap causes quitting.

One practical rule covers both findings: never miss twice. It's a rule from James Clear's Atomic Habits, one of the techniques on our Atomic Habits cheat sheet. One miss is noise. The day after is the one that counts.

Why does the stretch after week 3 feel like the place to quit? No study pins down a quitting window; this is our reasoning. The novelty is gone, the fast early gains have slowed, and the habit doesn't feel automatic yet. A streak helps here as a way to see your repetitions pile up, not as an exam you fail with one empty box.

How to tell when a habit has formed

A habit has formed when the situation starts it for you: you do it without deciding, remembering or pushing. You don't have to wait for a day number to find out.

Researchers use a four-question scale for this, the Self-Report Behavioural Automaticity Index, published in 2012 by Benjamin Gardner, Charles Abraham, Phillippa Lally and Gert-Jan de Bruijn. Here are its four ideas in our own words, as an "Is it a habit yet?" check:

• It happens on autopilot.
• I don't need a reminder to do it.
• I do it without thinking it through.
• I've often started before I notice I'm doing it.

Scoring: answer yes or no for one habit. Three or four yeses: you're close to your plateau. Two or fewer: keep the cue steady and keep going. The cutoff is our rule of thumb, not part of the scale.

A 66-day plan to start a habit and make it stick

How long does it take to start a habit? One day. The work is the two months or so after it.

This plan splits those two months into four stages. The day ranges are our suggestion, not something a study tested. For the full how-to, see our guide to habit formation.

1. Days 1 to 7: make it tiny

   Goal: show up every day. Pitfall: starting at full size.

   Begin smaller than you want to. James Clear's Two-Minute Rule says a new habit should take less than two minutes to do, so "read before bed each night" becomes "read one page".

   Pick one time and one place. If you keep putting off day 1, here's how to stop procrastinating.

2. Days 8 to 21: protect the cue

   Goal: the same trigger every day. Pitfall: "sometime today".

   Attach the habit to something you already do, which is habit stacking: "After I pour my coffee, I will read one page." Pick a cue you'll meet every day, and lay out what you need the night before.

3. Days 22 to 45: survive the dip

   Goal: keep the repetitions coming while it still feels like effort. Pitfall: reading "not automatic yet" as "not working".

   Never miss twice, tell one friend what you're doing, and keep your count where you can see it.

4. Days 46 to 66 and beyond: let it run

   Goal: grow it carefully. Pitfall: doubling it overnight.

   Add a little (one page becomes five) and run the four-question check above. If day 66 comes and it isn't automatic yet, you're not behind. You may just be on the longer side of an 18 to 254 day range.

Give day 66 a date. It's 9 weeks and 3 days: start on March 1 and it's May 5. Our days since calculator counts the days from your start date and can count the start day as Day 1.

One way to hold all of this: give the habit a fixed spot in a morning or evening routine in Greatr. You tap Play and run the routine one habit at a time, so the habit before it becomes the cue. Every habit earns XP, your streak counts the days, and the apps you chose stay locked until the routine is done.

How long does it take to change or break a habit?

Nobody has a solid number for how long it takes to change a habit, because breaking habits has been studied far less than building them. Our expectation, not a research finding: give it at least as long as a new habit, so think in months, not weeks.

In a 2014 review of basic laboratory research on learning, Mark Bouton concluded that the usual ways of changing behavior tend to inhibit the old behavior, not erase it, and that the change is often tied to the context where you learned it. So an old habit can resurface when you're back in the old setting.

The practical move is to build a stronger new behavior on the same cue: keep the trigger, swap the action. "After dinner, I scroll" becomes "after dinner, I walk around the block." Then the new behavior needs its own repetitions. Our list of 50 bad habits and what to replace each one with has swaps to borrow.

A 66-day habit tracker you can print (and what happens after day 66)

Print this page or copy the tracker into a notebook. Write your habit as one sentence, then cross off each day you do it. At days 21, 45 and 66, run the four-question check.

My habit: "After I ______, I will ______." Start date: ______. Day 66: ______.

Missed a day? Leave the number open and cross off the next one. You don't start over.

Prefer a screen? Our habit tracker gives each habit a one-tap check for today, its streaks and a year heatmap, and prints monthly and weekly sheets.

What happens after day 66? Probably nothing dramatic. Our suggestion: stop checking boxes once the habit runs on its own, but keep it in its slot, because the routine around it is what protects it in a busy week.

If you'd rather keep it on your iPhone, build the habit into a daily routine in Greatr. Otto, the app's guide, keeps count of every day you show up, and medals like the 30 Day Streak and 66 Habit Days mark the way.

Yesterday is already in my notebook. Today, we add one more day.

Your next step, today: write one sentence, "After I [cue], I will [two-minute version]", do it once and cross off day 1.