How small should a new habit be?

Updated 19 August 2026

• The test is your worst day, not your average one.
• Simple behaviours became automatic faster than effortful ones in the UCL data.
• The small version is the design, not a lesser day you settle for.
• Keeping the moment fixed and shrinking the size beats keeping the size and moving the moment.
• Grow it only once the current size has stopped being a decision.
• If you cannot decide between two sizes, the smaller one is nearly always right.

Why size is not the mechanism

A habit is a learned association: a situation comes round, and a response follows without being decided on. That association is built by the situation being followed by the response, repeatedly. Nothing in it is sensitive to how impressive the response was.

This is why the arithmetic of habit-building is so unforgiving of ambition. A forty-minute session you manage three times in a fortnight gives you three repetitions. A two-minute version you manage thirteen times gives you thirteen. The second person is more than four times further along, and they have done less work.

The UCL field study points the same way from the other direction: among its participants, simple behaviours reached automaticity considerably faster than effortful ones. Drinking a glass of water with lunch settled sooner than fifty sit-ups before breakfast.

Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W., & Wardle, J. (2010). How are habits formed: Modelling habit formation in the real world. European Journal of Social Psychology, 40(6), 998–1009. doi:10.1002/ejsp.674

The worst-day test

Pick the version by asking one question: could I do this on the worst day I have had this month? Not a busy day. The day with the delayed train, the bad night's sleep and the thing that went wrong at four o'clock.

If the answer is anything other than an immediate yes, halve it and ask again. Most people land somewhere near these:

• Read for twenty minutes → read one page
• Go to the gym → put the kit on and leave the house
• Meditate for fifteen minutes → three slow breaths, sitting down
• Write 500 words → open the file and write one sentence
• Cook properly → chop one vegetable before anything else happens
• Tidy the flat → clear one surface

These look like jokes written down. They stop looking like jokes on the fourteenth consecutive day, which is a stage the twenty-minute version usually never reaches.

The part everyone gets wrong

Almost everyone treats the small version as permission to underperform: the thing you fall back on when you have failed at the real plan. That framing is what makes people skip it, because doing the small version then means admitting a bad day.

It is the other way round. The small version is the habit. The larger sessions are what happens on the days you have room, and they are a bonus rather than the target. Once you hold it that way there is nothing to admit and no reason to skip.

There is a second thing the small version does, and it is the one people notice later. A version you cannot fail is a promise to yourself that you keep every day. Most people trying to build a habit have a long history of breaking those, and reversing that is worth more than any individual session.

When to make it bigger

Grow it when the current size has stopped being a decision — when you notice you have already done it rather than deciding to. That is usually weeks rather than days, and it is a different signal from feeling keen about it, which is weather.

Two rules keep growth from undoing the work:

• Grow the size, keep the moment. Whatever else changes, the point in the day it attaches to should not.
• Keep the small version available. Growing does not retire it. It is the floor for bad days, permanently, and a habit with no floor is one bad week from ending.

Leaving a habit at its current size is also a real answer. There is no requirement to progress, and a two-minute practice you keep for a year does more than a thirty-minute one you keep for a fortnight.

How Mini Loops handles it

Setup asks for two things: what happens at the moment, and what happens instead on a day that falls apart. Both go into the plan from the start, so the small version exists before the first hard day rather than being improvised during one.

If the full version keeps not happening, the app says what it noticed and proposes a smaller one. It does not shrink your plan on its own, and it does not comment on the days you used the small version.