STREAKS ARE NOT VANITY METRICS

Streak counters have a reputation problem. They read as gamification, as a number designed to manufacture obligation, as the part of the app you are supposed to be too serious to care about.

The dismissal is half right. The mechanism underneath is well supported. Most implementations of it are careless, and a careless streak is worse than none.

WHY THE MECHANISM WORKS

Two findings do most of the explanatory work here.

The first is habit formation itself. Phillippa Lally and colleagues at UCL tracked participants forming new daily habits and found automaticity built gradually, with a median of around 66 days to plateau and enormous variation between individuals — from a few weeks to the better part of a year. The relevant point is that automaticity is a function of repetition over time, not of motivation on any given day.

The second is implementation intentions. Peter Gollwitzer's work shows that specifying in advance when and where a behaviour will happen substantially raises follow-through compared to holding the same intention in general terms. Consistency is helped far more by a fixed cue than by resolve.

A streak is a crude but genuine instrument for both: it makes repetition visible, and it pushes you toward a stable daily slot rather than an opportunistic one.

WHERE IMPLEMENTATIONS GO WRONG

The standard design is all-or-nothing. Miss one day and a 94-day streak becomes zero.

This is a bad model of the thing it claims to measure. Missing a single day has a negligible effect on habit strength — Lally's data found that a lapse did not meaningfully derail the formation process. Resetting the counter to zero communicates the opposite, and it does so at the worst possible moment.

The failure mode is predictable. Having lost the number, the streak is no longer worth protecting, and the thing people abandon is not the streak — it is the habit the streak was standing in for.

The danger is not that people care too much about the streak. It is that a reset teaches them the past ninety days counted for nothing.

WHAT A STREAK SHOULD ACTUALLY TELL YOU

Read as a signal rather than a score, the number answers a narrow and useful question: has this been stable recently? A long streak means the practice has a reliable slot in your day. A streak that keeps breaking at the same point in the week is telling you something specific about that part of your week — and that is diagnostic information, not a failure.

This is why TockLog reports consistency alongside the daily productivity score. A strong score on an erratic pattern and a moderate score on a steady one are very different situations, and the productivity number alone cannot distinguish them.

USING IT WITHOUT BEING USED BY IT

• Judge the trend over a month, not the current number.
• Treat a single miss as a data point, not a verdict. The research is clear that one lapse does not undo the accumulation.
• Notice where breaks cluster. A streak that dies every Friday is information about Fridays.
• If protecting the number ever leads you to log a session you did not do, stop looking at it. A corrupted log is worse than no streak.

The number is a proxy. When the proxy and the goal disagree, the goal wins — and the goal was never the streak. It was a working week you can actually see.