Technique · Hooked
Reminder cadence that steps back as habits form
Size reminders to the behavior's natural cadence, base them on events the user cares about, and send fewer as self-started use grows. Reminders bridge the weeks or months before an internal trigger forms; they shouldn't become the reason people return.
Plain label: reminders that taperAlso known as: notification cadence, lifecycle messaging, event-based notifications, frequency caps, quiet hours, social notifications
How it works
Internal triggers take weeks to months of frequent use to form, so owned triggers have to carry users across that gap. Messages about events that matter to the user get acted on; broadcast volume teaches people to ignore or switch off the channel, which removes the bridge.
Evidence warning
Never inflate social counts in notifications ('3 friends posted') to raise opens.
What the research says
Choosing reminder copy and timing with a bandit algorithm, without raising volume, lifted daily actives about 0.5% and new-user retention about 2% (peer-reviewed, company authors).
Yancey & Settles, KDD (CASE-M01 Duolingo), 2020
How to build it
- Mobile
- Cap pushes per week, send on events the user cares about (a reply, a result, a due item), and lower the cap for users whose self-started sessions rise.
- Email and SMS
- Replace a fixed daily blast with a digest sized to the user's activity, with quiet hours honored.
- B2B
- Daily digest only for active projects, weekly otherwise; stop once the user opens the tool on their own most days.
What can go wrong
- Generic 'we miss you' messages.
- Raising volume to hit a daily-active goal.
- No quiet hours.
Pairs well with
- Measure the self-started share with the trigger inventory.
- Test timing against a holdout before scaling.
In practice
Where it shows up
Duolingo · 2020
RoundDuolingo: smarter reminders, not more reminders
The bandit algorithm lifted total DAU 0.5% and new-user retention 2% over a strong baseline, without sending more notifications (Yancey & Settles, KDD 2020; company authors, peer-reviewed).
M4 ×ReturnCase grade for the number: A. Evidence grade for the mechanism link: B.
Limits
What we don’t claim
- The Hook Model is a practical design tool, not a validated or complete theory of behavior.Hooked, Introduction
- This is a practitioner heuristic (grade D): test it in your own product before relying on it.
- Habit metrics need the longer habit horizon (weeks to months) before they can be read.
Sources
Hooked · Nir Eyal
Ch.2
Source IDs: HK-TRG12HK-TRG13
Book content is paraphrased, never quoted at length. Last verified October 24, 2026.
Your product
Find the stage your product is stuck in
Gamify It reads your product, picks the one move that matters at its stage, and sizes the test to your traffic. Connecting is free; a Pro subscription starts it.