Anyone who has ever Googled “how to form a habit” has seen the number 21. Three weeks. One month - and you’re a different person. The problem is that this number didn’t come from a study. It came from a plastic surgeon’s book.

Where did these 21 days come from?

In 1960, Maxwell Maltz’s book *Psychocybernetics* was published. The author - a plastic surgeon - observed that patients needed about three weeks to get used to their new face after surgery or to the absence of an amputated limb. It was simply an observation, not even an experiment. Just a doctor’s note about adapting to changes in appearance. So the figure was picked up by Tony Robbins, Brian Tracy, and the entire self-help industry from the 1970s to the 2000s. By the time actual data became available, “21 days” had become so deeply ingrained in the public consciousness that the data was simply ignored.

If you’ve ever said to yourself, “It’s been three weeks - why isn’t this working automatically yet?” - you were comparing yourself to a number with no empirical basis.

What the actual research showed

In 2010, Philippa Lelli and her colleagues at University College London published a study that remains the most cited on this topic to this day. Ninety-six volunteers chose one daily action - something related to food, drink, or physical activity - and tied it to a consistent context. For example: drinking a glass of water after breakfast. For 84 days, the participants recorded daily whether they had performed the action and how automatic it felt.

The median result was 66 days before the behavior became automatic.

But the most important thing about this study isn’t the number 66. It’s the range: from 18 to 254 days.

Some people established the habit in two and a half weeks. Others needed more than eight months. And these were ordinary people under normal conditions, not some extreme cases.

Lelli herself admitted in later interviews that it annoys her when “66 days” is quoted out of context. Her study didn’t reveal a universal timeframe - it showed that there is no universal timeframe.

The most useful detail that’s almost never mentioned

There’s one more point in Lelli’s data, and in my opinion, it’s worth more than all the others combined.

A single missed day doesn’t derail the process. The automaticity curve isn’t significantly affected by it. What’s harmful are consecutive misses - when one day is followed by a second, then a third.

Now think about how many times you’ve given up on something after just one slip-up. You missed your workout on Tuesday: “Well, that’s it - I’ve slipped up. I’ll start again on Monday.” Monday never came.

All this logic of streaks, chains, and “don’t break the streak” - which is built into every habit tracker - comes not from research but from game design. And it regularly undermines people’s motivation at a time when no real harm has yet been done.

If you miss a day, just do it the next one. It’s not “starting over.” It’s simply continuing.

Why Small Steps Really Work

Now, about the idea itself.

In the study, simple actions with a clear contextual cue - something like “drink water after breakfast” - became automated the fastest. Complex behaviors took much longer to become automatic.

Hence the practical value of small steps. It’s not that five minutes of exercise provide physical benefits - on their own, they don’t do much. The point is that you’ll actually manage to do those five minutes on a bad day, when you have neither the energy nor the time for an hour-long workout. And it’s precisely those bad days that determine the fate of a habit.

A formula that works better than simply “do less”:

Anchor + action + place. Not “read more,” but “after I brush my teeth at night - 10 pages in bed.” Not “work out,” but “right after my morning coffee - 10 minutes of exercises in my room.” Context does half the work for you: over time, the situation itself starts triggering the action.

And about promises—let’s be honest

Here I have to dampen the motivational mood a bit.

Phrases like “do 30 push-ups a day and you’ll see results in a month” or “post every day and you’ll gain followers in two months” are what discourage people most often.

Not because the advice is bad. But because it sets a specific deadline for a process that has no deadline. After a month, a person looks in the mirror, doesn’t see what was promised, and concludes that it didn’t work. Even though, in reality, it worked exactly as it was supposed to.

A realistic timeframe based on Lelli’s data: 60–90 days for the action to stop requiring willpower. And be prepared for the fact that it might take you twice as long - not because of a lack of willpower, but because of the variation shown by the study.

What to Take Away From This

Choose one action. Not three, not a ten-item New Year’s resolution list.

Make it small enough to do even on your worst day of the week.

Tie it to something you already do every day.

Forget about 21 days. Aim for two to three months, and don’t panic if it takes longer.

And most importantly: if you miss a day, just do it the next one.

This last point is worth more than all the motivational literature put together. Because habits rarely die from a single slip-up. They die from the conclusion a person draws after that slip-up.


Source: Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W., & Wardle, J. (2010). How are habits formed: Modelling habit formation in the real world. European Journal of Social Psychology, 40(6), 998–1009.