How Long Does It Take to Form a Habit? What the Data Says
Updated: 1 day ago
Search "how long does it take to form a habit" and the first answer you will see is 21 days. That figure traces back to a 1960 self-help book, not to a study. The research that has actually measured habit formation tells a different story, and the gap between the myth and the data is large enough to change how you plan a new habit.
We pulled the numbers from the two most useful sources and did the arithmetic the headlines skip.
What the studies actually measured
The first is the 2010 study by Phillippa Lally and colleagues at University College London, published in the European Journal of Social Psychology. Ninety-six volunteers each chose one eating, drinking, or activity behavior to do daily in the same context, such as "after breakfast", for 12 weeks. Every day they rated how automatic the behavior felt. Of the 96, 82 provided enough data, and 39 produced automaticity curves that fit the researchers' model well enough to estimate a plateau. For those 39, the time to reach 95 percent of their personal plateau had a median of 66 days and a range of 18 to 254 days.
The second is a 2024 systematic review by Singh and colleagues in the journal Healthcare, which pooled 20 intervention studies with 2,601 participants. Only four of those studies reported a time to habit formation. Across them, the medians ran from 59 to 66 days, the means from 106 to 154 days, and individual participants ranged from 4 to 335 days.
The numbers by habit type
Lally's paper broke the 39 modeled participants down by the type of behavior they chose. A detailed reading of the paper by The Behavioral Scientist lists the median and the middle 50 percent (the interquartile range) for each group. Our method was simple: divide each median by 21 to get the multiple of the "rule", and compare each group's lower quartile to 21 to see how many people could possibly have been done in three weeks.
Drinking habits (15 people): median 59 days, middle half 39 to 75 days. That is 2.8 times the 21-day rule.
Eating habits (10 people): median 65 days, middle half 35 to 106 days. That is 3.1 times the rule.
Exercise habits (13 people): median 91 days, middle half 44 to 118 days. That is 4.3 times the rule.
Everyone (39 people): median 66 days, 3.1 times the rule.
The lower quartile for every category (35, 39, and 44 days) sits above 21. In plain terms, at least three out of four people in every group were still building the habit when the 21-day rule said they were finished. The single fastest participant in the whole study took 18 days.
Exercise stands out twice. Its median is the slowest, and its spread is the widest: the middle half of exercisers spans 74 days, compared with 36 days for drinking habits. Starting a workout habit is both slower and less predictable than starting a glass of water with lunch.
The same numbers on a calendar
Abstract day counts are easy to shrug off. Here is what they look like if you start a habit on January 1.
January 21: the 21-day rule declares you done. In the data, at least three-quarters of every group was not.
February 28: day 59, the median for drinking habits, and the low end of the median range in the 2024 review.
March 7: day 66, the median for the whole 2010 study.
April 1: day 91, the median for exercise. The original study ran only 84 days, so more than half of the exercise group had not plateaued when it ended.
April 28: day 118, the upper quartile for exercise. One in four exercise habits took longer than this.
September 11: day 254, the slowest estimate in the 2010 study.
December 1: day 335, the slowest individual in the 2024 review.
Why the average is so much longer than the median
The 2024 review reports medians of 59 to 66 days but means of 106 to 154 days. Dividing one by the other, the mean is 1.8 to 2.3 times the median. That ratio only happens when the distribution has a long right tail: most people cluster around two months, and a minority take far longer, pulling the average up.
Two practical conclusions follow. If you want a planning number, use the median and budget about ten weeks. If you are slower than that, you are not broken. You are in the tail that every one of these studies found.
What this means for your streak
The data points to three rules.
Plan for ten weeks, not three. A streak that feels hard at day 25 is behaving exactly as the research predicts. The habit is not failing. It is roughly a third of the way to automatic.
Missing one day does not reset you. Lally's abstract states that missing a single opportunity did not materially affect the habit-formation process. This is why a tracker should let you log a skip honestly and keep going, instead of zeroing a counter. In Habit Textapp you reply "not today", the seven-day strip shows one gap, and the next day's check-in arrives as usual.
Keep the context constant. Every participant in the 2010 study tied the behavior to the same daily cue. A text that lands at the same time every day is one of the simplest cues there is. If 7pm is wrong for you, asking it to check in at 7am moves it.
Try it by text
Habit Textapp is a daily check-in for one habit. Text Hi to (332) 301-4574, tell it the habit you want to build, and each day it asks one question. Reply yes or no, watch the streak grow, and ask for an invite code to invite a friend who sees whether you followed through.
It works on any phone that can text, with no app to install and no account to create. Your first 20 messages are free. Details are at textapp.co/apps/habit.
Comments