The first 3 days run on novelty
On day 1, you're not so much doing as deciding. The decision itself feels good: you have a plan, and you already feel a little like a different person. On days 2 and 3, the action still counts as an event - it stands out, you remember it all day, and you note that you did it.
But novelty is a consumable. It fades precisely because you repeat the action. By the 4th time, a run stops being a run and becomes just Tuesday morning: there's nothing to mark, and no one inside is applauding.
Days 2-3: novelty
Reward from starting something new: lasts while it is new
The second line - automaticity - rises at a different pace. For a cue ("poured coffee") to pull the action along by itself, the link needs dozens of repetitions: studies count in weeks and months, not days.
The slump comes midweek, not at the end
There's a gap between the two lines. Novelty has already dropped, automaticity hasn't risen yet, and in that stretch the action runs only on willpower - the most unreliable fuel of all. Day 4 isn't special: it's simply the first day when the action gives you nothing but itself in return. The number is arbitrary: "exactly day 4" is an observation, not a measured law; it's the gap itself that's reliably described, not the date when it happens.
How easy the action feels day by day. This is a shape, not a measurement: it's hardest in the middle, not at the finish.
This also explains why one missed day hits harder than the miss itself: it feels like the chain is broken and you have to start counting again.
A consecutive-day counter reinforces that feeling. It shows not how many times you did it, but how many times in a row you didn't slip up. Zero after 4 days looks as though those 4 don't count - even though the repetitions haven't gone anywhere.
Missed a day - the chain is broken, so start from zero.
In a study where people repeated the same action for 84 days, missing 1 day did not noticeably affect the growth of automaticity. What a long string of missed days does was barely tested, so it's more honest to be cautious here.
So count not consecutive days, but your hit rate. One failed day 4 decides nothing. A failed day 4 that you declare the end of the attempt decides everything.
Why does a habit often feel harder on day 4?
What's the best way to track progress after missing a day?
Sources
- Lally et al., "How are habits formed: modelling habit formation in the real world," European Journal of Social Psychology, 2010. The observation lasted 84 days and included fewer than 100 participants. The median estimate for automaticity reaching a plateau was 66 days, with a range from 18 to 254 days
- missing 1 day did not noticeably affect the curve
Next in the “Habits” series
3 articles in full on the site, 4 more in the app
