The science · Consistency

Streaks break people: why 80% consistency beats perfection

The streak is the most seductive mechanic in habit apps — and the most dangerous. The day it breaks, many people don't lose a day. They lose the habit. The evidence points to a better target: flexible consistency, around 80%, held for months.

The problem with all-or-nothing

Dieting research named the failure mode decades ago: the “what-the-hell effect” (Polivy & Herman). Break a rigid rule once — one biscuit, one missed workout — and the day gets mentally written off, so the lapse cascades: the rule is already broken, may as well…. All-or-nothing framing converts small, statistically inevitable misses into identity-level failures. And misses are inevitable over months: illness, travel, deadlines, children. A plan that requires perfection is a plan with an expiry date.

What the research says works instead

• Slack, planned in advance, increases persistence. Marketing researchers Sharif and Shu tested goals with built-in “emergency reserves” — e.g. run 7 days a week with two skip-days available — against both strict and easy goals. The reserve structure sustained engagement better: people kept the tough target's motivation but a slip no longer meant failure.
• Missing one day doesn't damage habit formation. In the UCL habit study, Lally and colleagues found single missed days had no measurable effect on the automaticity curve. Habits are built by usually showing up, not always showing up.
• Self-compassion beats self-punishment after lapses. Across the self-regulation literature, harsh self-criticism after a slip predicts abandonment; treating the lapse as data predicts resumption. The practical heuristic: never miss twice.
• Don't compensate. Doubling tomorrow's session to “make up” for today's miss raises injury and burnout risk and re-frames the plan as debt. The consistent move is simply to resume the plan as written.

The 80% target

Somewhere around four-fifths execution, held over months, you get nearly all of the physiological and skill adaptations of a perfect plan — with none of the fragility. The target isn't a magic constant from a single trial; it's an engineering margin informed by the evidence above: high enough that the plan works, low enough that real life fits inside it. Judge it over a rolling window, and a bad Tuesday stops being an emergency.

Related: How long habits really take · Why tracking works · The fresh start effect