You've probably seen the number on a poster or in a headline. The 66 comes from one study of volunteers who chose their own behavior, and it travels without its fine print. This article walks through what the researchers did, what the 66 stands for, and what a log of your own days can show that an average cannot.
What did the 66 day study measure?
The study followed 96 volunteers for 84 days and asked each of them one question a day: how automatic does your new behavior feel? It was run by Phillippa Lally and colleagues and published in 2010 in the European Journal of Social Psychology.
Each volunteer picked one health behavior, such as eating fruit or going for a walk, and tied it to a daily cue, such as after breakfast. They reported each day whether they did it and rated how automatic it felt. As one review of the study puts it, habit strength "showed an asymptotic increase, with an initial acceleration that slowed to a plateau after an average of 66 days."
| What | Detail |
|---|---|
| People recruited | 96 volunteers, mean age 27 |
| Length | 84 days |
| Behavior | One self-chosen health behavior, tied to a daily cue |
| Measure | Daily self-rated automaticity |
| Headline result | Plateau after 66 days on average |
In plain terms: early repeats produced the biggest jumps in automaticity, then the gains got smaller until the curve went flat.
Why is 66 days not an answer for you?
Because it is an average of a wide spread, and the top of that spread was never observed. Individual times ran from 18 days to a predicted 254 days, according to the British Psychological Society's digest. The study stopped at day 84, so anything beyond that is a projection from the fitted curve, not a day someone was seen reaching.
A few more limits sit behind the headline:
- A secondary account reports that 82 of the 96 gave enough data, and that 39 of those fit the curve well. The 66 days rests on that smaller group.
- About half of the participants had not reached a habit-level score by day 84.
- Automaticity was self-reported, and people chose their own behaviors.
- Even with identical repetitions, a separate review notes that people differed in how quickly habit strength rose.
So 66 days describes the group the researchers studied. It can't say what your own pattern will look like.
There is also a question of what "formed" means. The researchers did not watch for a day when a switch flipped. They fitted a curve to each person's daily ratings and read off the point where it flattened out. A flat curve means the behavior stopped getting more automatic, which isn't the same as feeling effortless. That is why a person could be well past day 66 and still feel the effort, or reach the plateau at 18 days and barely notice.
Do simple habits form faster than hard ones?
In this study, yes, though the evidence is thin. Automaticity peaked sooner for simple actions, such as drinking water, than for elaborate ones, such as doing 50 sit-ups.
The BPS digest says exercise behaviors took about 1.5 times as long to plateau as eating or drinking behaviors. A secondary account puts the exercise median at 91 days, against 65 for eating and 59 for drinking, and adds that the difference was not statistically significant and that the study had limited power. Treat it as a hint, not a result.
What does a missed day do to a habit?
In the study, one missed day did little. The PubMed Central review quotes the finding directly: "Missing the occasional opportunity to perform the behaviour did not seriously impair the habit formation process." Automaticity gains picked up again after a single miss.
Many misses were a different story. The BPS digest reports that lots of missed repeats, even spread out over time, lowered the highest automaticity people reached. If you are weighing what a slip means for a count, see why streaks work and restarting a streak after a slip.
Where did the 21 day idea come from?
Not from habit data. The same review says the 21 day figure is a myth that traces to anecdotal observations of plastic surgery patients in 1960. The 66 day study is the better known measurement, and even it carries the caveats above.
Tracking it yourself
An average can't tell you when a behavior got easy for you. A log can, because the same cue and the same entry every day make the shape of your own calendar visible.
Pick one behavior and one cue, log it daily, and add a quick note on how automatic it felt. Keep the cue the same each day, because the study tied every behavior to one. Sudden changes in the cue muddy the picture, and so does switching to a different behavior halfway through. After a few weeks, look at where the gaps fall and whether the entries got easier to make. The Nut Tracker keeps a streak, a calendar and stats on your phone, with no account, and you can export or delete the data whenever you want. More on how to read those views is in the blog, and the home page has the app.
Bottom line
The 66 days is the average from one study of 96 volunteers who picked their own behaviors, and the spread ran from 18 days to a projection of 254. One missed day did not break the pattern in that study. Your own log says more about you than the average does.
Sources
- PubMed Central: Making health habitual, the psychology of habit-formation and general practice
- PubMed Central: A review and analysis of the use of 'habit' in understanding, predicting and influencing health-related behaviour
- British Psychological Society: How to form a habit
- Wikipedia: Habit
This is general information, not medical advice. The disclaimer says more.

