Home · The science
Evidence, then design

Built on how habits actually form

Every product claims science. Here is the specific research GrowthArc is built on, and - more useful - exactly where each finding shows up in the product. If a feature has no evidence behind it, it didn’t ship.

No card required to start. Cancel anytime.

The mapping

Finding → feature

Research on the left, the exact product mechanism on the right.

The findingThe researchWhere it lives in GrowthArc
Habits take ~66 days to automate, with wide varianceLally et al., 2010 - European Journal of Social PsychologyThe milestone ladder celebrates day 3, 7, weekly to 66, then monthly - matched to the real curve, not a 21-day myth.
Implementation intentions (“when X, I will Y”) roughly double follow-throughGollwitzer & Sheeran, 2006 meta-analysisHabits carry a time and trigger, and nudges name the moment: “after lunch” beats “today”.
Missing once doesn’t break formation - quitting after a miss doesLally et al., 2010Streak repair and “the tiny version counts” nudges: a two-minute version protects the identity, not the number.
Loss aversion: protecting a streak motivates more than starting oneKahneman & Tversky’s prospect theory, appliedDay 6-of-7 nudges (“keep tonight and the week closes”) and the near-milestone star that burns brightest on the watch.
Celebration consolidates habit identityFogg, Tiny Habits research programmeThe payoff: completing a milestone closes the arc on screen and wrist - a felt reward at the moment of action.
Accountability to a person beats accountability to an appATD/ASTD accountability studies traditionNudges wear your face, check-ins flow to your dashboard, and the AI escalates to the human - the person stays the point.
Self-monitoring alone improves outcomesMichie et al., behaviour change technique taxonomyOne-tap logging, readings on schedule, photos, and charts both of you can read.
Passive self-monitoring lowers the cost of trackingMichie et al., behaviour change technique taxonomy (self-monitoring of outcomes)Coming: sleep, HRV and steps from wearables arrive as readings without a tap - status and pricing.

The honest bit

What we don’t claim

No software makes change happen. Coaching does - the evidence for human accountability is the strongest in this entire stack, which is why GrowthArc is coach-first rather than another self-help app with streaks.

What the platform does is remove the places where change usually dies: the forgotten Tuesday, the unfelt milestone, the slip nobody saw until the no-show.

  • A tool for coaches, not a replacement for them
  • Mechanisms, not miracle claims
  • Your method carries; the platform carries it further
Where change dies
×
Tuesday, 9pm
Nobody there → nudge in your voice
×
Day 7, unnoticed
No felt win → the arc closes on the wrist
×
The quiet fortnight
Seen too late → health score surfaces it day 3

Questions

On the evidence

Is the “66 days” number real?
It’s the average automaticity plateau in Lally et al. (2010); individual habits ranged from 18 to 254 days. That variance is why milestones continue monthly after 66 rather than declaring victory.
Do streaks backfire?
They can, when a broken streak reads as failure. That’s why the product treats a miss as repairable and nudges the tiny version instead of shaming the zero.
Can I see the research yourselves cite?
Every claim above names its source; all are published, peer-reviewed work searchable by author and year.

Put the evidence to work

Fourteen days free. Your method, carried by mechanisms that have data behind them.

Start free · 14 days

No card to start. Your first client can be in tonight.