Storytime
The first week of a new habit is easy in a way that is actively misleading. You are interested. The thing is new. You notice yourself doing it, and the noticing feels like progress. Then, somewhere in the second week, the interest quietly runs out and you find that nothing has been built underneath it yet.
That gap is the whole story, and it has three parts. Novelty carries you at the start. Automaticity, the point at which the behaviour runs without a decision, does not arrive for weeks. In between there is a stretch where nothing is carrying you, and you have to do the thing on purpose, every single time, for no reward you can feel. Most people stop there. Then they explain it to themselves as a character flaw, which is the only part of this that is actually wrong.
Week one is not evidence
The best public numbers on how quickly new efforts fall away come from research on New Year's resolutions. Norcross and colleagues followed a group of resolvers and found that 77% were still going after one week, 55% after one month, 43% after three months and 40% after six (figures as summarised by Oscarsson, Carlbring, Andersson and Rozental in PLoS ONE, 2020). Look at the shape rather than the numbers. The single steepest fall in that curve happens between the first week and the first month. After that it flattens out. The people who are still going at one month are mostly still going at six.
Two honest caveats. Those outcomes are self-reported, and the study does not pin the drop to a particular day, so nobody can tell you that week two specifically is the cliff. What the data supports is narrower and more useful: the first weeks shed far more people than the months that follow, and week one tells you almost nothing about whether you will still be doing this in March.
This matters because week one feels like evidence. It feels like proof that you are the sort of person who does this now. It is not. It is the report from the one week in which the habit had help.
What is actually happening in the gap
Starting a behaviour and keeping it are not the same task, and they do not run on the same machinery. Rothman, Sheeran and Wood made the argument directly, writing about dietary change in the Annals of Behavioral Medicine in 2009: action control has to be split along two axes at once, reflective versus automatic and initiation versus maintenance, because what gets a behaviour started and what keeps it going have different determinants and need different strategies. Kwasnicka, Dombrowski, White and Sniehotta reached a similar place from the other direction, reviewing more than 100 behaviour-change theories in Health Psychology Review in 2016 and grouping what sustains a behaviour into five areas, one of which is habit. Keeping is its own problem.
Automatic control takes time to build. In the study most often quoted on this, Lally, van Jaarsveld, Potts and Wardle followed 96 volunteers who each picked one daily eating, drinking or activity behaviour and repeated it in the same context for 12 weeks (European Journal of Social Psychology, 2010). Automaticity rose along a curve that climbed quickly at first and then flattened, reaching a plateau after a median of 66 days, with a range across individuals of 18 to 254 days.
Put those two findings next to each other and the gap has an obvious shape. Your deliberate motivation is at its highest on day one and falls from there. Your automatic control starts at zero and takes something like two months to arrive. In the second week the first curve has dropped and the second has barely started, and there is a stretch of days where you are the only thing holding the habit up.
You will see this stretch called the valley of disappointment, after James Clear's Atomic Habits (2018). That is a description, not a study, but it is a good one, and having a name for a thing makes it much easier to sit through.
Three things that bridge it
You cannot make the gap shorter by wanting it more. You can make it easier to cross, and there are only really three levers.
Make the habit smaller than the gap. The size of a habit decides how much motivation it needs, and in the second week you have very little to spend. Cut the thing until it survives a day when you are tired, late and not interested, because those are the days that decide this. Two pages, not a chapter. One set, not a session. More on how far to take this in make it ridiculously small and the two-minute rule.
Give it a cue that fires without you. A habit attached to an intention ("I'll read more this month") has nothing to trigger it. A habit attached to an event that already happens ("after I put the kettle on, I read two pages") borrows a trigger that is already reliable. Wood and Neal set out the case for habits as context-cued responses in Psychological Review in 2007, and the if-then version of this has the strongest evidence of anything in this article: Gollwitzer and Sheeran's 2006 meta-analysis of 94 independent tests found a medium to large effect (d = 0.65) on goal attainment for simply writing the plan in that form. Where you do it counts too, which is the subject of why where you do it matters.
Have a reason to show up that is not the outcome. Results are the wrong fuel for the second week because there are none yet. Something that is true on the day is better: that you said you would, or that this is the kind of person you are choosing to be. Oscarsson and colleagues' own experiment (PLoS ONE, 2020, 1,066 participants followed for a year) found that goals framed as something to start beat goals framed as something to stop, which is the same idea in a different coat. Identity-based habits takes this further.
If you want a record that treats a missed day as a missed day, Seeded does that. Set a habit as a plain daily check while you are crossing the gap, and add a count or a duration later, once the showing up has stopped being the hard part. It is free for three habits on iPhone.
What not to do in week two
Do not add more habits. The urge to fix a wobbling habit by starting a second one is strong and completely backwards. Every new behaviour needs its own share of the attention you are already short of.
Do not raise the bar. Week one often goes well enough that the target creeps up. Two pages becomes a chapter, a walk becomes a run. You have just made the habit more expensive at the exact moment your budget collapsed.
Do not turn it into a test of character. The most damaging thing about the second week is not the missed day, it is the conclusion people draw from it. The habit has not failed. The scaffolding has come down before the concrete set, which is a scheduling problem, not a moral one.
Do not wait to feel like it. In the gap, you will not. The feeling arrives after the behaviour has been repeated enough times to stop costing anything, not before.
What a good second week actually looks like
Unremarkable. Small. A bit boring. You do the thing, it takes two minutes, you do not think about it much afterwards, and there is nothing to report to anybody.
That is not a sign it is too easy. It is what the middle of habit formation looks like from the inside. The interesting part was week one and the interesting part will be month three; the bit in between is supposed to be dull, for the same reason that the boring routine is the point.
If you do miss a day, and most people will, the thing that decides the habit is what happens on the next one, not the gap itself. That is a whole subject of its own: see what to do the day after you break a streak and, if the number in your tracker is doing more harm than good, why streaks backfire.
And if you want the honest timescale to hold in your head while you cross the gap, it is in how long habits actually take: a median of 66 days and a range of 18 to 254. Week two is not the end of the process. It is about a fifth of the way in.
Seeded is a habit tracker for iPhone that does not shout. It is free for three habits, and in week two the cap is a feature rather than a limit, because the fix for a wobbling habit is fewer of them, not more. Get it on the App Store.
Bitesized
- The steepest fall in new habits happens in the first weeks, not the first months. Of the resolvers Norcross and colleagues followed, 77% were still going at one week and 55% at one month, and the curve flattens after that.
- Starting and maintaining a behaviour run on different processes: deliberate motivation at the start, automatic context cues later (Rothman, Sheeran and Wood, 2009).
- Automaticity takes a median of 66 days to plateau, with a range of 18 to 254 (Lally and colleagues, 2010), so there is a stretch in the middle where motivation has dropped and automaticity has not arrived.
- Three levers close that gap: a smaller habit, a cue attached to something that already happens, and a reason to show up that is not the result.
- Writing the plan in if-then form has the strongest evidence of the three (d = 0.65 across 94 tests, Gollwitzer and Sheeran, 2006).
- A good second week is boring. Adding habits, raising the bar or treating a miss as a verdict are the three ways people make it worse.