How long does a habit really take? The 21-day myth

The Short Version

A new habit does not reliably form in 21 days. Mel Robbins and Andrew Huberman both point to the same study, led by Phillippa Lally, in which the time for a simple behavior to become automatic ranged from 18 to 254 days. Robbins gives the average as about 66 days [1][2]. Robbins calls the 21-day idea "malarkey" [1]. Huberman still uses 21 days, but as a training block followed by a 21-day test, not as a promise that a habit will be automatic by then [2]. Both agree that the timeline depends on the person, the behavior and the systems around it. They also agree that missing a day doesn't erase your progress [1][2]. The practical point is to judge a habit by whether it is getting easier, not by the calendar, and not to quit because day 21 has passed and it still takes effort [1][3].

Table of contents

What They Said

Mel Robbins: "that is malarkey"

In her habits episode, Robbins lists three mistakes almost everyone makes when trying to change. The second is believing a habit takes 21 days: "The mistake number two is this notion that it takes 21 days for a new habit. And that is malarkey. This is deeply personal" [1].

Her evidence is the study "How Are Habits Formed," published in the European Journal of Social Psychology and led by Phillippa Lally of University College London [1]. She explains that participants tried to change a simple behavior for 12 weeks, such as drinking a bottle of water with lunch, walking after dinner, or running for 15 minutes a day [1]. The result: "It took between 18 and 254 days" for the behavior to become automatic, and "the average is basically 66 days" [1].

Robbins presents the spread as good news, not bad: "it means that every one of us is a little bit different and the amount of resistance that you feel when you try a new behavior change is going to be dependent on you. It's going to be dependent upon the thing that you pick. It's going to be dependent upon the systems that you have in place. So all of this is in your control" [1]. Her definition of a formed habit is practical. It is the point at which "your resistance to doing it is gone because you have programmed your mind, body and nervous system to learn this new pattern" [1]. The goal is a behavior as automatic as brushing your teeth or making coffee. Nobody stands at the coffee maker debating whether they can push the button [1].

She uses her own life to show the range. Her best-known tool doesn't feel automatic even now: "even the lady that invented the 5-second rule, that is still not a habit. I still do it most days but I still feel resistance and it's been 14 years" [1]. Other changes settled quickly: drinking more water and walking the dogs every day took "more like 18 days to lose the resistance" [1]. She describes herself as a slow learner in some areas: "Something that it might take me, Mel Robbins, the stubborn learner, 200 days of pushing through this, that's kind of empowering" [1].

The second finding she stresses is that "missing a day does not materially affect the habit formation process" [1]. You don't go back to zero [1]. Later in the episode she pushes this further in her own words: "if you miss two or three days, if you miss a week, if you miss a month, whatever, it's not going to impact all the progress that you've made" [1]. Readers should note that the summary of the study describes missing a day, or even two or three. The week-and-month version is Robbins's own broader encouragement [1].

Her advice for a slow habit is curiosity, not self-blame: "You can expect that to be shorter or longer depending upon your resistance. You can also start to get really curious. Like if it's taking longer, don't beat yourself up. There's nothing wrong with you" [1]. And she ends by moving attention from the finish line back to the start: "It doesn't matter if it takes you two hundred and fifty four days to make that a reality because you will never get to that automatic two hundred and fifty four day habit unless you commit to day one" [1]. Her imagined listener's response sums up the mood she wants: "It might take me less than 21 days. Might take me more. I can hack this" [1].

Why some habits take longer, according to Robbins

Robbins's explanation for a slow habit is usually a weak reward. She breaks every habit into a trigger, a behavior and a reward [1]. The reward causes the brain to release dopamine, which encodes the habit loop. Once the loop is set, the brain skips over the effort and jumps to the treat when the trigger appears, which is why the resistance fades [1]. She ties this directly to the 18-to-254-day spread: "This is why it's not a one size fits all. This is why in the biggest study around habits and how long it takes to form habits, it's between 18 and 254 days" [1].

Her own case: in 2019, while hosting a daytime talk show, she had to be in a 5:30 a.m. exercise class. She had triggers (the alarm, her clothes on the floor) and a class she'd signed up for the night before, but feeling good after exercise has never been enough of a reward for her [1]. The habit stuck once she found a coffee shop selling homemade breakfast burritos. The next morning the burrito was the first thing she thought of when the alarm rang [1]. "My husband does not need a breakfast burrito to get to the gym. Mel Robbins does," she says. "I need something immediately following the exercise because that's what makes the habit stick" [1]. Citing habit scientist Dr. Wendy Wood, she says the reward should come, or you should be moving toward it, within about two minutes of the behavior [1]. If a habit still feels hard, she says, "I guarantee you the resistance is still there because you haven't come up yet with a reward that's made it worth it" [1]. Guilt doesn't count: "That's not a reward. That's a punishment" [1].

She also takes the pressure of willpower off the reader. Citing Dr. Angela Duckworth of the University of Pennsylvania, she says successful, disciplined-looking people "don't have any more discipline than you and me"; they have better habits and systems [1]. Her husband Chris, whose routines make her feel "married to a human robot," fills his water bottle the night before and leaves it on the kitchen island [1].

Andrew Huberman: same study, different emphasis

Huberman cites the same Lally research, a 2010 study, and adds details. The behavior studied was walking after dinner, which he notes helps clear glucose from the blood [2]. "In order to form that habit It took some people 18 days and other people 254 days" [2]. He explains how the researchers defined success: "They were doing it about 85% of the time And they also reported not having to spend that much mental effort" [2]. He expects listeners to be surprised and reassures them: "people are highly variable and if you can't form one habit easily it doesn't mean that you can't form other habits easily" [2].

His explanation for the variation is a concept he coined, limbic friction. He is open that it "is not a term that you're going to find In the formal neurobiological literature or even psychological literature" but a term he uses to pull together several findings [2]. It describes the strain needed to overcome one of two states: being too anxious to calm down and do something, or "feeling too tired or lazy or not motivated" [2]. Both relate to the autonomic nervous system, which works like a seesaw between alertness and calm. Limbic friction measures "how much activation energy you need in order to engage in a particular behavior" [2]. High-energy habits are harder when you're tired, and calming habits like meditation are harder when you're keyed up [2].

He measures how strong a habit is in two ways. The first is context dependence: whether you do the habit the same way whatever the setting. He brushes his teeth at the same time relative to waking whether at home or traveling, so that habit is strong. The second is how much limbic friction it takes [2]. The goal is automaticity, which he defines this way: "Automaticity is fancy language for the neural circuits can perform it automatically" [2].

Huberman's 21-day system

Huberman then offers a 21-day program, used for a long time by someone he knows well and their kids [2]. You choose six new habits to do every day for 21 days, expecting to complete only four or five a day. The real goal is the habit of performing habits, and the structure allows for things like resistance training that shouldn't be done daily [2]. There is no punishment for missing a day, and you shouldn't try to make up for the slip by doing extra the next day [2].

He breaks the 21 days into smaller units: chunking this 21 days into two day bins can be very very useful... thinking okay I can do this for a day and if I can do it for a day I can probably do it for two days and then resetting [2].

The key feature, which separates his system from the 21-day myth, is the second block: "After 21 days You stop engaging in this 21 day deliberate four to five things per day type schedule and you simply Going to autopilot you ask yourself how many of those particular habits that I was deliberately trying to learn in the previous 21 days Are automatically incorporated into my schedule" [2]. In that second period you add nothing new. He says this fills a gap: in the habit literature he searched, there wasn't much on "What to do once you form to habit and how to evaluate whether or not that habit Is likely to persist long into the future" [2].

He expects the results to be mixed: "What you may find is that you kept up two of those things very consistently throughout the 21 days And perhaps there was one of them that you did sporadically and that there were Three others that frankly you didn't manage to execute" [2]. Someone who gets all six he jokingly calls one of "these mutants," and that person should then see whether they can keep all six going for another 21 days [2]. The cycle can repeat, 21 days to build and 21 days to test, but "it's only once you've achieved all those six habits as reflexive that you would move forward" [2]. He describes the first block as "the stimulus period in which the habits get wired in" and the following periods as "that kind of thermometer or the test bed" [2].

He is careful not to oversell the number: "There's nothing neuroscientifically unique about it except for the fact That it's not just 21 days broken up into two day chunks" [2], and "there's nothing holy about the system" [2]. What he values is its features: "the lack of compensation for missed days," a high-intensity 21 days followed by "a kind of a competition test with yourself," and the fact that "habit slips missing of particular habits and not doing all six is kind of built into the system" [2]. His warning against overload is central: "many people are trying to cram so many new behaviors into their nervous system That they don't stand a chance of learning all those behaviors" [2].

He links this to advice he got from Berkeley neurologist Bob Knight when he started graduate school: "you don't want to do Anything or engage in a routine in any way that you can't keep up consistently for at least five and ideally six days per week" [2]. Knight added that you might need to update that every four or five years [2]. Huberman says this fits his view of habit slips: if you miss, "the next day you just get right back on on the on the horse" [2].

Mel Robbins on the long middle: "It's the day that you return"

In her six-step goals episode, Robbins adds the attitude that goes with a long habit timeline. She tells listeners to accept "that it's going to be kind of tedious and boring for a long time and I'm not going to see results for a long time" [3]. She is still in the slow phase herself: working toward a pull-up, she has spent two weeks just learning to hang from the bar [3].

She expects lapses: "there are going to be days where you don't do the reps. There are going to be weeks where you don't do the reps. There might even be months that you don't do the reps. That's okay" [3]. What counts is coming back: "It's the day that you return. That defines success because it means you didn't quit" [3]. Her example: if it is day 37, you have not eaten the way you said you would... You can always come back to this formula [3]. She also tells people to throw the timeline out the window. Attitude, actions, reps and persistence are in your control, and how long it takes is not [3]. Quoting James Clear, she says "you will rise or fall to the level of your systems," and if the reps are hard to keep up, remind yourself: "The fact that it's hard means it's working" [3].

Where They Agree, and Where They Don't

Agreement

The range, not the number. Both cite 18 to 254 days from the Lally study, and both treat the spread as the main finding [1][2]. Robbins adds the 66-day average [1], and Huberman adds the definition behind it: about 85 percent follow-through with little mental effort [2].

Variation is normal and not a character flaw. Robbins says the timeline depends on you, the behavior and your systems, and that there's nothing wrong with you if it takes longer [1]. Huberman says if one habit is hard for you, others may still come easily, and he explains the difference with limbic friction [2].

Missed days don't reset the clock. Robbins reports that missing a day doesn't materially affect habit formation [1]. Huberman builds misses into his program and says to get back on the next day [2]. They also agree you shouldn't try to make up for a missed day. Huberman rules out doing extra the next day [2], and Robbins says what matters is simply returning [3].

Systems matter more than willpower. Robbins calls struggling to change a systems problem [1]. Huberman treats habits as neuroplasticity that can be designed with specific tools [2].

Disagreement

Is 21 days useless or useful? This is the headline debate, and it is narrower than it sounds. Robbins rejects 21 days as a measure of when a habit is formed [1]. Huberman doesn't claim that. He uses 21 days as a training block and then tests the result, and he calls the number nothing "holy" [2]. His system actually shares Robbins's skepticism: its whole purpose is to find out, after 21 days, which habits did not become automatic [2]. The real difference is emphasis. Robbins wants people to drop the clock and stop blaming themselves [1][3]. Huberman thinks a fixed, intense block with a scheduled check gives structure the open-ended approach lacks [2].

How forgiving are missed days? Robbins goes furthest, saying even missing a week or a month won't undo the progress you've made [1], and that there may be months without reps [3]. Huberman is more demanding. Bob Knight's rule is not to take on any routine you can't keep up at least five, ideally six, days a week, and the 21-day program expects four or five of six habits every day [2]. These are different kinds of advice. Robbins is protecting people from quitting after a lapse. Huberman is setting a standard of consistency for choosing which routines to take on at all.

One habit or several? Huberman's program asks for six new habits at once, while warning against cramming in too many [2]. Robbins's examples usually work on one behavior at a time, such as water, a morning class, or a drink at the end of the day [1][3]. Robbins doesn't argue against Huberman's approach directly. The contrast is in how their programs are built.

Exact times or phases of the day? In her goals episode, Robbins recommends implementation intentions: a specific time and place, like writing at 9 a.m. in the office [3]. Huberman says scheduling a habit at an exact clock time helps in the short term but not the long term, because the nervous system produces behavior based on state, not clock time. He suggests anchoring habits to phases of the day with predictable biology [2]. Both still agree that a clear, repeatable cue helps.

What kind of reward? Robbins relies on concrete rewards within about two minutes, such as a breakfast burrito, an iced latte, or a selfie texted to a friend [1]. Huberman's version is internal. It uses reward prediction error: expected rewards reinforce behavior, unexpected ones give a bigger dopamine release, and expected rewards that don't arrive drop dopamine below baseline [2]. He recommends looking forward to the whole before-during-after sequence, while being honest that the start is unpleasant, because "you can't lie to yourself" [2]. The two can work together. Robbins's burrito is a form of anticipation, and she herself thought about it throughout the class [1].

How much of life is habit? A smaller difference: Robbins says 44 percent of waking actions are automatic [1], while Huberman cites an estimate that "up to 70% of our waking behavior is made up of habitual behavior" [2]. The sources don't explain the gap, but both make the same point: a large share of daily life runs on habit.

Why It Works

Habits are rewiring, and rewiring varies. Huberman describes habits as learned neuroplasticity: changes in connections between neurons that make some behaviors more likely [2]. Through Hebbian learning, neurons that are active together strengthen their connections, and NMDA receptors help by recruiting more receptors when a neuron gets strong input [2]. Mentally walking through the steps of a new habit once sets these neurons in motion so "the dominoes fall more easily" when you act [2]. Because this process depends on the person's state and the behavior, fixed timelines don't fit.

Limbic friction explains the spread. A habit that fights your current state, like hard exercise when you're tired or meditation when you're wired, needs more activation energy and so takes longer to become automatic [2]. That is why Huberman puts the highest-friction habits in the first eight hours after waking, when norepinephrine, epinephrine, dopamine and healthy cortisol are naturally high. Lower-friction habits like journaling or practicing a skill you already have fit the afternoon and evening [2]. The actual rewiring happens during deep sleep, so protecting sleep protects habit formation [2].

The brain brackets finished habits. Huberman describes the dorsal lateral striatum, part of the basal ganglia, becoming active at the start and end of a habit, not during it. This creates a neural imprint that lets the habit happen even after bad sleep or on a hard day [2]. Once a habit is well learned, the pattern is transferred from the hippocampus to the neocortex and no longer needs the bracket. At that point, he says, it's good to move the habit around the day. Being able to do it in any phase shows context independence [2]. This gives a practical test for "formed" that has nothing to do with counting days.

Dopamine locks in the loop. Robbins explains that rewards release dopamine, which encodes the trigger-behavior-reward loop until the brain jumps from trigger to treat and the resistance disappears [1]. Huberman adds that dopamine "is not about feeling good. It's about feeling motivated," and much of it comes during anticipation [2]. A missing or weak reward is one reason a habit might sit near the 254-day end of the range [1].

Breaking habits follows the same logic. Huberman explains long-term depression (unrelated to mood): when neuron A fires and neuron B doesn't within a certain window, their connection weakens [2]. Reminders work only in the short term, according to a 2020 scoping review by Heather Fritz, and self-punishment is rarely applied consistently enough [2]. What works is doing a small positive behavior right after catching yourself in the bad habit, such as drinking a glass of water, doing breath work or spending five minutes on a language, which changes the whole firing sequence [2]. Robbins's version is to replace, not quit: keep the trigger and the reward and swap the behavior. For her end-of-day cocktail, she found that "The reward is the ritual," so she kept the ritual with mocktails [1].

Where It Applies, and Where It Doesn't

The 18-to-254-day range comes from simple health behaviors: water with lunch, a walk after dinner, a 15-minute run [1][2]. It is a reasonable guide for everyday habits like these, but the sources don't claim it predicts timelines for complicated skills or deep behavioral change.

Huberman's 21-day system suits people who like structure and want to take on several changes at once, and it includes a built-in check against overload [2]. It fits less well for someone making one hard change, where Robbins's one-behavior-plus-reward approach may be simpler [1]. Neither source says a habit is "done" at a fixed point. Robbins's own 14-year-old 5 Second Rule still takes effort [1].

Robbins's reassurance that a missed week or month doesn't undo progress is meant to keep people from giving up [1][3]. It shouldn't be read as permission to practice only now and then. Huberman's five-to-six-days-a-week standard is the counterweight [2].

Addiction is a different category. Huberman notes that addiction needs a much fuller approach than habit tools and points listeners to his episode with Dr. Anna Lembke [2]. Some of the supporting practices he mentions, such as supplements for alertness or sleep, should be discussed with a doctor first, as he himself advises [2].

What To Do With It

1. Drop the 21-day deadline as a verdict. Expect anywhere from about 18 to 254 days, with 66 as a rough average, and judge progress by how much less effort the habit takes, not by the date [1][2].

2. Define "formed" the way the researchers did. Aim for doing the habit about 85 percent of the time with little mental effort [2]. A sign it is fully set is when you can do it at any time of day or in any setting [2].

3. If you like structure, run 21 on, 21 off. Pick six habits, expect four or five a day, think in two-day chunks, and don't make up for missed days. Then spend the next 21 days adding nothing and see what stuck [2].

4. Only take on what you can do five or six days a week. Follow Bob Knight's rule, and don't cram in more new behaviors than you can realistically learn [2].

5. Put the hardest habit where your biology helps. Do high-friction habits in the first eight or so hours after waking, and easier ones later in the day [2].

6. Walk through the steps once in your head. Before starting, or when returning after a lapse, think through the exact sequence of the new habit [2].

7. Fix the reward before blaming yourself. If a habit isn't sticking, check the reward first. Pick one you actually want, within about two minutes of the behavior, and never use guilt [1]. Look forward honestly to the whole before, during and after [2].

8. Make the next rep obvious. Set up your environment the night before, like Chris Robbins's water bottle on the kitchen island, because you'll rise or fall to the level of your systems [1][3].

9. When you miss, just come back. Don't restart the count and don't make up for the missed days. Return the next day, because "It's the day that you return" that defines success [2][3].

10. Commit to day one. Whatever your personal timeline turns out to be, you only reach it by starting [1].

Sources

1. Build Amazing Habits: Simple Steps to Break Bad Habits & Make Good Ones Stick · The Mel Robbins Podcast
2. The Science of Making & Breaking Habits · Huberman Lab
3. How to Achieve Any Goal You Have in 6 Simple Steps · The Mel Robbins Podcast