Habit Formation Timeline Calculator
Pick your habit's effort level and how often you'll do it. You get an estimated number of days until it runs on autopilot, the date that lands on, and an honest range either side of it.
Habit Formation Timeline Calculator
Your habit
Optional, and display only. It never changes the maths.
How much effort does one rep take?
10 to 30 minutes, or a small amount of prep. Read 10 pages, meditate, journal, skincare, follow a diet.
How many days a week will you do it?
Pick at least one day a week.
Start date
Day 1 is your start date. Already started? Put the original date in and the calculator tells you which day you are on.
Individual results in the study spanned 18 to 254 days. Yours is an estimate, not a countdown.
Where a 30-day run lands
Day 30 lands once you pick a start date.
Thirty days is the steep part of the curve, which is why it feels like it works. Run it on the 30-day habit challenge tracker.
Milestones on the curve
| Milestone | Day | Date |
|---|
Want the calendar dates for a fixed-length run instead? Use the 30-day challenge end date calculator.
Sixty-six days is a long way past the point most people stop counting. LockIn30 keeps you on one habit and one tick a day. It holds the streak and stacks your progress photos into a timelapse, so the stretch between Day 30 and Day 66 is something you can see rather than just endure.
Run one habit in LockIn30Everything is worked out in your browser using date-only arithmetic, so daylight saving changes and time zones will not add or drop a day. Nothing is stored or sent anywhere.
What the research actually says about how long a habit takes
The number on this page comes from Lally, van Jaarsveld, Potts and Wardle (2010) in the European Journal of Social Psychology. Ninety-six volunteers each picked one eating, drinking, or activity behaviour, performed it daily in a consistent context, and rated its automaticity every day for 84 days using the Self-Report Habit Index.
The funnel matters, so here it is in full: 96 people recruited, 82 supplied usable data, the curve model fitted 62 of them, and only 39 fitted well. Those 39 are where the headline comes from. Their median time to reach 95% of the automaticity plateau was 66 days, the middle half sat between 39 and 102 days, and the two extremes were 18 days and 254 days.
A 2024 systematic review backs the ballpark up. Singh, Murphy, Maher and Smith, publishing in Healthcare, pooled 20 studies and 2,601 participants and reported medians of 59 to 66 days, means of 106 to 154 days, and individual times from 4 to 335 days. Fifteen years apart, the two datasets land on about two months and disagree wildly about any individual person.
Where the 21-day number came from (and why 66 replaced it)
Maxwell Maltz was a plastic surgeon, not a habit researcher. In Psycho-Cybernetics (1960) he wrote that "it usually requires a minimum of about 21 days to effect any perceptible change in a mental image", describing how long patients took to stop seeing their old face in the mirror, how long an amputee kept feeling a phantom limb, and how long a new house took to feel like home.
None of that is a claim about behaviour. Sixty years of self-help repetition sanded "a minimum of about 21 days for a perceptible change" down to "21 days to form a habit". The 21/90 rule that followed has no study behind either number.
Be fair to the feeling, though. The automaticity curve is steep early, so three weeks in, a daily moderate habit is already around 59% of the way to its plateau on this model. Things genuinely do get easier by Day 21. It is a checkpoint that got promoted to a finish line.
How this calculator works, constant by constant
Lally fitted an asymptotic curve to each participant, y = a minus b times e to the power of minus cx (Mitscherlich's law of diminishing returns), and defined time to automaticity as the point at 95% of the asymptote. Here is every number this page uses and where it comes from.
| Constant | Value | Source |
|---|---|---|
| Simple base | 59 reps | Median for the drinking behaviour group (N=15) |
| Moderate base | 66 reps | Overall median across all 39 good-fit participants |
| Demanding base | 91 reps | Median for the exercise behaviour group (N=13) |
| Simple band | 39 to 75 | Q1:Q3 for the drinking group |
| Moderate band | 39 to 102 | Q1:Q3 for all 39 good-fit participants |
| Demanding band | 44 to 118 | Q1:Q3 for the exercise group |
| Full range | 18 to 254 | Min and max across the 39. Context only, never scaled |
| Curve constant k | 2.8416 | Derived from the paper's own time-to-95% equation |
The curve constant is the only derived figure, so here is the derivation. Lally's time to 95% of asymptote is minus the natural log of a over 20b, divided by c. Substituting the paper's median parameters a = 35 and b = 30 gives k = ln(600/35) = 2.8416. Progress after r repetitions is then 1 minus e to the power of minus 2.8416 r over R, where R is the base repetition count for your effort tier. The calculator carries the full precision of that constant rather than the rounded 2.84, because at the 30-day mark a moderate habit sits on 72.5% and the rounding of that one figure decides whether you see 72% or 73%.
Two honest caveats. First, the interquartile band (39 to 102 for a moderate habit) is a more useful "range" than the 18 to 254 everyone quotes, because those two figures are two individual people at the edges of a 39-person sample. Second, the Kruskal-Wallis test comparing time-to-asymptote across the three behaviour groups was not significant (p = 0.328), and Lally notes the study was not powered for sub-group analyses. The paper's own authors could not prove exercise habits take longer than water habits, because the sample was too small. Their medians point that way and their discussion argues for it, so that is what this page uses.
Frequency is plain arithmetic, not a hidden multiplier: base repetitions times 7, divided by your days per week. And what is deliberately not modelled: motivation, prior experience, cue strength, identity, age. All of them matter. None of them has a published number you can multiply by, so putting one in would only add fake precision.
What to actually do with the number
The curve is front-loaded. On a moderate daily habit you are a quarter of the way there by Day 7, half by Day 17, three quarters by Day 33, and then you spend another month grinding out the last quarter. That shape should change how you plan.
Pick the shortest version of the habit you will actually repeat, not the impressive one. The effort tier moves the answer by roughly a month: 59 reps against 91. Ten push-ups you do daily beat a gym session you do twice a week.
Anchor it to a fixed daily cue. Consistency of context is what the study was measuring in the first place, which is why the calculator warns you when you drop below four days a week. Stretching 66 repetitions across a 3-day week is arithmetic, not evidence.
Expect the stretch between Day 30 and Day 66 to feel like nothing is happening, because on the curve, not much is. That is also where the missed-day rules start to matter, which the guide on strict vs forgiving habit tracking covers. Pair this estimate with the 30-day challenge end date calculator for the dates, and the 30-day habit challenge tracker for the ticking off. LockIn30 exists for the part after that: one habit, past the point where the counting gets boring.
Frequently Asked Questions
Common questions about habit formation timeline calculator
How long does it really take to build a habit?
The median was 66 days in the Lally et al. (2010) study, but the middle half of participants landed between 39 and 102 days and the full spread ran from 18 to 254. The 2024 Singh meta-analysis of 20 studies and 2,601 people put medians at 59 to 66 days, with individual times from 4 to 335 days. So two separate bodies of evidence agree on the ballpark. Budget about two months, and expect the spread around that to be wider than the number itself.
Is it really 21 days to form a habit?
No, and it never was a habit claim. Maxwell Maltz, a plastic surgeon, wrote in Psycho-Cybernetics (1960) that "it usually requires a minimum of about 21 days to effect any perceptible change in a mental image", an observation about patients adjusting to a new face, an amputated limb, or a new house. Six decades of quoting turned his minimum into a deadline. That said, 21 days is not nothing: on this calculator's curve, a daily moderate habit is about 59% of the way there by Day 21. It feels easier. It just is not finished.
Why is 66 days the number everyone quotes?
It is the median time to reach 95% of the automaticity plateau among the 39 Lally participants whose data fit the curve model. It is worth knowing what it is a median of: 96 people signed up, 82 supplied usable data, the model fitted 62, and only 39 fitted well. Sixty-six days describes those 39 people, not the average person.
Does missing one day reset habit formation?
No. Lally tested this directly and found that after a missed opportunity, automaticity dropped by 0.29 points on the SRHI subscale, a rounding error. Their conclusion was that one missed opportunity did not meaningfully dent the process. Missing a week is a different matter: Armitage (2005), which Lally cites, found week-long lapses did predict worse future performance. One day, no. A fortnight of nothing, yes. The rules you set for a missed day are covered in the guide on strict vs forgiving habit tracking.
Why do some habits take longer to form than others?
Effort per repetition. In Lally's data, drinking habits hit the plateau at a median of 59 days, eating habits at 65, and exercise habits at 91, so the exercise group took roughly one and a half times as long. The authors put that down to complexity: more moving parts, slower automaticity. Their statistical test on the difference was not significant (p = 0.328), so treat it as direction rather than proof.
Is 30 days long enough to build a habit?
Enough to be most of the way, not enough to be done. On the model this page uses, 30 daily repetitions gets you about 76% of the way for a simple habit, 73% for a moderate one, and 61% for a demanding one. That is why 30-day challenges feel like they work: the steep part of the curve is at the start. The last quarter is the slow part, and it is where most people stop counting. The 30-day challenge end date calculator gives you the dates for that first stretch.
What if I only do the habit three times a week?
Then the same 66 repetitions take about 154 calendar days. The calculator does the conversion, but read the warning that comes with it: every participant in the source study practised daily, in a fixed context. A habit with no daily cue may plateau lower rather than simply later. If a daily version of the habit exists, it is the better bet.
How accurate is this calculator?
It is an estimate built from published medians and quartiles, not a prediction about you. The base numbers, the ranges, and the curve constant all come from one 2010 study of 39 successful habit-formers, most of them postgraduate students in their twenties. Every figure it uses is printed on this page so you can check it. Treat the output as a realistic expectation to plan against, not a date on which something switches on.