What if missing one day on a streak doesn't actually matter? Rebuilt my tracker around the 2010 paper everyone mis-cites.
Just shipped imperfectly.cc this week. Free, no signup, no email.
The premise came from an itch I could not scratch with any other habit app. I deleted Streaks, Habitica, Way of Life, Fabulous, and Habit Coach AI in the same month. Not because they were bad apps, but because every one of them used the same loop, and that loop was breaking me. Build a streak. Miss a day. Watch the counter crash to zero. Quietly delete the app a week later.
So I went looking for the actual research, expecting some nuance. What I found was much stranger.
The most cited paper on habit formation is Lally et al. 2010, out of UCL. They followed 96 people forming new habits for 12 weeks. Two findings sit completely outside the worldview of every streak app I have ever used:
-
Habit formation follows an asymptotic curve, not a streak. The average time for a behaviour to feel automatic was 66 days, but the range was 18 to 254 days.
-
Missing a single day had no statistically significant effect on the final outcome. The curve barely moves.
Read that second one again. The thing every streak app spends its visual design screaming about, the broken chain, just does not show up in the data.
So I built around the actual research:
-
A personal Lally curve fitted to your own check-ins. The shape of the curve is the progress, not the count.
-
An estimated day to automaticity that bends when you slip but never crashes to zero.
-
A soft message on a missed day. No fire emoji going dark. No "you broke your chain."
What is interesting from a builder angle: removing the streak counter took the most addictive part of the category out. That was the whole point, but it does mean retention has to come from the curve actually feeling motivating, not from a guilt loop. I do not know yet if that is going to work as a business. I just know I could not keep using a model that the actual paper directly contradicts.
Honest day-2 numbers: ~50 visitors day 1, ~20 onboarding completes, 4 day-2 returners. Tiny but real.
The thing I am genuinely curious about and would love this community's read on: does seeing a curve instead of a streak feel motivating, or does the projected day to automaticity start to feel like a soft deadline the same way a streak does? I cannot tell from the inside.
If you have shipped something deliberately less sticky than the category norm, I would also love to hear how you thought about retention without the usual hooks.
Site: imperfectly.cc
Swapping a high-pressure streak for a resilient curve is a smart way to stop treating users like failing machines and start treating them like learning systems. It turns a missed day from a catastrophic crash into a simple data point, making habit formation feel more like a trend analysis than a high-wire act.
Does the curve make people feel like they are building a trend, or do they still fixate on the gap when they miss a day?