Start hereThe book’s sum says that getting 1% better every day for a year makes you about thirty-seven times better. In the main study of habit formation, automaticity rose along a curve that levels off. So which of the book’s ideas hold up?
About the book
Atomic Habits is a practical guide to building good habits and breaking bad ones, by the writer James Clear. It was first published in the US on 16 October 2018 by Avery, an imprint of Penguin Random House, and in the UK two days later by Random House Business.
It runs to 320 pages. Its publisher describes it as a No. 1 New York Times bestseller translated into more than 60 languages.
It is written for general readers, and it mixes short stories, findings from psychology and a step-by-step framework. Clear publishes several of its chapters as articles on his own website, each labelled as an excerpt from a named chapter, so you can read parts of it free.
Small changes add up
The excerpt Clear labels as chapter 1 argues from compounding. A change too small to notice today can make a large difference once it is repeated for months, in either direction.
He calls habits “the compound interest of self-improvement”. He illustrates it with a sum: get 1% better every day for a year and you end up about thirty-seven times better.
Does it hold up? The arithmetic is right: 1.01 multiplied by itself 365 times is about 37.8. But no study we found measures anyone improving by a fixed 1% a day. In a 2010 study of habit formation led by Phillippa Lally, automaticity rose along a curve that levels off.
The excerpt opens with the story of British Cycling, whose new performance director arrived in 2003. The excerpt says the team won 60 percent of the gold medals available at the 2008 Olympics. Clear’s own footnote corrects that figure: British riders won 8 of the 14 road and track gold medals.
His line that they had won only one gold medal since 1908 does not match Olympedia’s results, which list British track-cycling golds in 1920, 1992 and 2000.
Team Sky, the professional road team whose riders won the Tours de France in the story, was later accused in a UK parliamentary committee report of March 2018 of using medicines allowed under anti-doping rules to enhance performance, which the team rejected.
And a winning team cannot show which of its many changes made the difference.
Try it: judge a small change by whether you keep repeating it, not by how fast the results grow.
Systems matter more than goals
Goals say where you want to go; the system is the set of daily routines that gets you there. Clear writes that goals “are good for setting a direction, but systems are best for making progress”.
His book page calls this one of the book’s core ideas: “You do not rise to the level of your goals. You fall to the level of your systems.”
Does it hold up? It is a way of framing effort rather than a claim a study could settle. The parts of a system that research has tested, cues, a steady setting and if-then plans, come up below.
Try it: write down the daily routine behind one goal, and review the routine each week rather than the goal.
Change who you think you are
Clear describes three layers of change: your outcomes, your process and your identity. He argues that lasting habits start from the inside, with the kind of person you want to be.
Each small action, he says, is “a vote for the type of person you wish to become”. In a note, he says Simon Sinek’s Golden Circle framework is similar.
Does it hold up? We found no study of Clear’s three layers. The nearest research is a set of voting experiments. In 2011, wording survey questions as “being a voter” rather than “voting” raised turnout in two US elections.
A larger field experiment in 2016 found no significant difference, and the two teams disagreed in print about why. Two further large replications by the second team, published in 2018 and 2020, also found no effect. The evidence so far leans against the original result, though the replications all come from one team.
Every habit runs through four steps
Clear’s own summary of the book describes a habit as four stages: a cue, a craving, a response and a reward. From these Clear builds his Four Laws of Behaviour Change.
To build a habit, make it obvious, attractive, easy and satisfying. To break one, turn each law around: make it invisible, unattractive, difficult and unsatisfying.
Does it hold up? Cues and setting are the heart of habit research. A 2016 review of habit research describes people repeating the same behaviours in recurring contexts. The volunteers in Lally’s study carried out their behaviour daily in the same context, such as “after breakfast”.
The Four Laws themselves are Clear’s own framework, and we found no study that tests all four as a set.
Try it: keep the time, the place and the action just before the habit the same each day. Put the cue where you will see it, such as the book on the pillow or the running shoes by the door.
| Make it obvious | To break a habit: make it invisible. |
| Make it attractive | To break a habit: make it unattractive. |
| Make it easy | To break a habit: make it difficult. |
| Make it satisfying | To break a habit: make it unsatisfying. |
The Four Laws are Clear’s own framework; we found no study that tests all four as a set.
Tie a new habit to an old one
Habit stacking means attaching a new behaviour to something you already do every day, in the form “After [current habit], I will [new habit]”. Clear calls it a special form of an implementation intention (an if-then plan), and credits the method to BJ Fogg’s Tiny Habits program.
Clear’s online excerpt on habit stacking opens with a striking figure: adults have 41% fewer neurons than newborns. The 2007 study behind it compared 8 newborn and 8 adult brains. It found 41% fewer neurons in adults in one region of the thalamus, not across the whole brain.
Does it hold up? If-then plans are well studied, though the effect is small. A 2006 meta-analysis of 94 tests found a medium-to-large effect on reaching goals. A 2024 update of 642 tests found them effective, with larger effects when plans were in if-then form, people were highly motivated, and the plan was rehearsed.
The same 2024 paper found strong signs of publication bias. Corrected estimates ranged from d = .35 with one method to d = .15 with a more conservative one, and the authors call the effect robust but small. We found no review that tests stacking onto an existing habit on its own.
Try it: write the plan in full, with one fixed cue: “After I pour my morning coffee, I write one line in a notebook.” Then rehearse it.
Make the first step tiny
The Two-Minute Rule says a new habit should take less than two minutes to start: “read before bed each night” becomes “read one page”. The point is to make starting easy and let the rest follow.
The same idea runs through the Third Law: make a good habit easy, and turn it around to make a bad one difficult.
Try it: shrink tomorrow’s version of the habit until it takes less than two minutes.
Track it, and get back on quickly
Clear suggests a habit tracker: in its simplest form, a calendar on which you cross off each day you stick with the routine. When a day is missed, his rule is simple: never miss twice. One slip is normal; the aim is to stop it becoming a new pattern.
Does it hold up? The main study supports the first half. In Lally’s study, missing one opportunity to perform the behaviour did not materially affect habit formation. We found no study that tests missing twice as a threshold.
How long it takes. In Lally’s study, 96 volunteers repeated one daily behaviour for 12 weeks. The time participants took to reach near-full automaticity ranged from 18 to 254 days. A 2024 review reports the median as 66 days; Clear’s own article calls 66 days an average.
The studies point to about two months as a midpoint, with wide variation from person to person. They measured habit by how automatic the behaviour felt to the person, not by a date on the calendar.
Try it: expect weeks to months, and treat a missed day as a blip rather than a restart.
What holds up
What the sources show for each idea
The practical core holds up best: cues, a steady setting and if-then plans. The famous numbers are shakier, and the most famous of them, the 21-day rule, did not come from a study of habits.
Where the 21-day rule comes from. Clear himself traces it to the plastic surgeon Maxwell Maltz. In Psycho-Cybernetics (1960), Maltz wrote that it takes “a minimum of about 21 days for an old mental image to dissolve and a new one to jell”.
Clear says Maltz drew it from patients getting used to a change, such as a new face. Clear adds that Maltz noticed it took him about 21 days to form a new habit himself. Clear stresses that Maltz was observing, not stating a fact, and wrote “a minimum”.
Who it’s for, and who can skip it
It suits anyone who wants a readable, practical system for starting or stopping everyday behaviours and likes worked examples. Its core techniques (cues, if-then plans, a steady setting) rest on real research.
Readers who already know the research on habits and if-then planning will find much of it familiar. Its examples include stories such as the cycling team, so its numbers are best read as illustrations. It is a general self-help book, not a treatment for addiction or compulsive behaviour.
Quick answers
How long does it take to form a habit?
In a 2010 study by Phillippa Lally and colleagues, the time to reach near-full automaticity ranged from 18 to 254 days, with a median of 66. A 2024 review found medians of 59 to 66 days and means of 106 to 154 days across the four studies that measured timing, with individuals ranging from 4 to 335 days.
Is the 21-day habit rule true?
Not as a finding about habits. Clear traces it to Maxwell Maltz’s Psycho-Cybernetics (1960), which said it takes “a minimum of about 21 days for an old mental image to dissolve and a new one to jell”. Clear stresses that Maltz was observing, not stating a fact, and the studies that measured habit formation found much longer and more varied times.
Does getting 1% better every day make you 37 times better?
The arithmetic is right: 1.01 multiplied by itself 365 times is about 37.8. But no study we found measures anyone improving by a fixed 1% a day, and in Lally’s study automaticity rose along a curve that levels off.
What is habit stacking?
Attaching a new behaviour to something you already do every day, in the form “After [current habit], I will [new habit]”. Clear calls it a special form of an if-then plan and credits the method to BJ Fogg’s Tiny Habits program.
What are the Four Laws of Behaviour Change?
To build a habit, make it obvious, attractive, easy and satisfying; to break one, make it invisible, unattractive, difficult and unsatisfying. They are Clear’s own framework, and we found no study that tests all four as a set.
Use this
Fix the cue, keep the setting steady, and expect weeks to months, not 21 days.
What exactly is my cue, and will I see it when it comes?
Quick quiz
Nothing is scored or saved. Recalling it once now is what makes it last.
1. What do the studies point to as a rough midpoint for forming a new habit?
Answer: About two months.Read it again ↑
2. In Lally’s study, what did missing one day do to habit formation?
Answer: It did not materially affect it.Read it again ↑
3. Who does Clear credit for the habit-stacking method?
Answer: BJ Fogg, from his Tiny Habits program.Read it again ↑
How we checked this · Last checked October 2026
- Quotations come from the chapter excerpts Clear publishes on his website, his summary and book pages.
- We also read his article on how long habits take, and the US and UK publishers’ pages.
- The habit-stacking excerpt is labelled chapter 4; his habit-tracker page places habit stacking in chapter 5.
- For the studies we read the abstracts, plus the full texts of the 2024 habit-timing review by Singh and colleagues and the 2024 if-then meta-analysis. The median of 66 days comes from that review, not from the abstract of Lally’s study.
- The two 2016 letters in which the voting-study teams disagreed were not read in full.
- Statements that no study exists (of 1% a day, of “twice” as a threshold, of habit stacking on its own, of the Four Laws as a set, of Clear’s three layers) mean none turned up in our searches of Europe PMC and Crossref in October 2026.
- The 1% figures were computed by us: 1.01 to the power 365 is about 37.8, and 0.99 to the power 365 is about 0.03.
- Clear’s article on how long habits take was re-read in October 2026. It says Maltz drew the 21 days from patients adjusting to changes and from his own new habits, and that Maltz was observing, not stating a fact.
- Lally’s abstract reports 96 volunteers, 82 with enough data, a model fit for 62 and a good fit for 39; it does not say which group the 18–254 day range covers.
Checked October 2026.
Sources
- J. Clear, Atomic Habits (Avery, 2018): chapter excerpts on the author’s website, each labelled with a chapter on the page: “This Coach Improved Every Tiny Thing by 1 Percent” (ch. 1), “Forget About Setting Goals” (ch. 1), “Identity-Based Habits” (ch. 2), “Habit Stacking” (labelled ch. 4; Clear’s habit-tracker page places habit stacking in ch. 5), “The 2-Minute Rule” (ch. 13), “The Habit Tracker” (ch. 16); the author’s summary page and book page; and “How Long Does it Actually Take to Form a New Habit?”.Primary
- Penguin Random House, US product page; Penguin UK, UK product page.Popular
- P. Lally, C. H. M. van Jaarsveld, H. W. W. Potts and J. Wardle, “How are habits formed”, European Journal of Social Psychology 40 (2010).Scholarly
- B. Singh, A. Murphy, C. Maher and A. E. Smith, “Time to form a habit”, Healthcare 12 (2024).Scholarly
- P. M. Gollwitzer and P. Sheeran, “Implementation intentions and goal achievement”, Advances in Experimental Social Psychology 38 (2006); P. Sheeran, O. Listrom and P. M. Gollwitzer, “The when and how of planning”, European Review of Social Psychology 36 (2024).Scholarly
- W. Wood and D. Rünger, “Psychology of habit”, Annual Review of Psychology 67 (2016).Scholarly
- C. J. Bryan, G. M. Walton, T. Rogers and C. S. Dweck, “Motivating voter turnout by invoking the self”, PNAS 108 (2011); A. S. Gerber, G. A. Huber, D. R. Biggers and D. J. Hendry, “A field experiment shows that subtle linguistic cues might not affect voter behavior”, PNAS 113 (2016); and their exchange of letters in PNAS (2016); A. S. Gerber, G. A. Huber and A. H. Fang, “Do subtle linguistic interventions priming a social identity as a voter have outsized effects on voter turnout?”, Political Psychology 39 (2018), and “Voting behavior is unaffected by subtle linguistic cues”, Behavioural Public Policy 7 (2023; online 2020).Scholarly
- M. Abitz et al., “Excess of neurons in the human newborn mediodorsal thalamus compared with that of the adult”, Cerebral Cortex 17 (2007).Scholarly
- Olympedia, cycling track results for 1920, 1992 and 2000.Popular
- House of Commons Digital, Culture, Media and Sport Committee, Combatting doping in sport (HC 366, 5 March 2018); CNN’s report of the team’s response (5 March 2018).Primary