🎁 Get the FREE AI Skills Starter Guide β€” Subscribe β†’
BytesAgainBytesAgain
πŸ¦€ ClawHub

Retention Phase Intervention Selector

by @quochungto

Use this skill to diagnose which retention phase (initial / medium / long-term) is broken for a user cohort and select the RIGHT type of intervention for tha...

Versionv1.0.0
⚑ When to Use
TriggerAction
- A cohort analysis shows users are churning but you are not sure when or why
- Retention looks stable in aggregate but early cohorts are thinning out
- You are about to invest in retention experiments but have not diagnosed the phase
- A PM asks "should we improve onboarding, add features, or run win-back campaigns?" β€” the answer depends entirely on which phase is broken
- You suspect churn is being hidden by strong acquisition growth
- Long-term cohorts are declining despite high feature velocity
**Prerequisite signals:** Your product has cleared product-market fit (stable retention curve exists for at least one cohort segment). If the retention curve never stabilizes, run `product-market-fit-readiness-gate` first β€” this skill cannot fix a pre-PMF product.
---
πŸ’‘ Examples

Example 1: Mobile app with initial phase broken

Situation: A fitness tracking app shows that 65% of new users are active on Day 1, but only 22% return on Day 7, and only 14% return on Day 14. The curve never plateaus.

Product type: Mobile app. Initial phase = Day 1. Medium phase = Days 2–14. Long-term = Week 3+.

Diagnosis: Initial phase failure. 78% churn within the first week indicates users are not reaching a meaningful experience of core value before disengaging. The curve's failure to plateau confirms there is no retained user segment.

Churn masking check: Total downloads growing 20% month-over-month, which was masking the cohort-level pattern in dashboard reviews.

Interventions prescribed:

  • NUX audit: map from app install to first completed workout. Identify all steps that precede the first success moment. Remove mandatory account creation from before first workout.
  • Day 3 trigger: if user has not completed a second workout by Day 3, send a contextual push notification tied to the user's first workout type.
  • At-risk threshold: flag any user who has not returned by Day 2 for a targeted re-engagement sequence.
  • Run activation-funnel-diagnostic to identify the specific funnel step with highest drop-off.
  • Not prescribed: Habit formation experiments, new feature development, or win-back campaigns β€” all medium- and long-term phase tactics.


    Example 2: SaaS product with medium phase broken

    Situation: A project management SaaS shows 60% of Month 1 users still active in Month 2, but only 28% in Month 4 and 18% in Month 6. The initial drop is acceptable for a SaaS product; the Month 2–4 drop is steep.

    Product type: SaaS. Initial phase = Month 1. Medium phase = Months 2–5. Long-term = Month 6+.

    Diagnosis: Medium phase failure. Users are successfully onboarded (strong Month 1 retention) but are not forming a habit of returning. The tool is not yet their default project management environment.

    Churn masking check: Not detected β€” new sales are flat, so the cohort-level pattern is visible in aggregate.

    Interventions prescribed:

  • Engagement loop audit: map what power users (top 10% by session frequency) do in Months 2–4 that new users do not. Identify the behavioral correlate of medium-term retention.
  • Weekly digest email: summarize team activity from the prior week, sent Monday morning when motivation to plan is high. Trigger is behavioral (inactivity for 5+ days) not calendar-based.
  • Collaboration prompt: prompt users to invite a second team member. Social investment (teammates' data is in the tool) raises switching costs.
  • Promise of future value: send a "what's coming in Q3" product update to Month 2 cohorts. Creating anticipated value gives users a reason to maintain the subscription through a low-engagement period.
  • Not prescribed: NUX redesign (initial phase is healthy), feature bloat audit (feature velocity is not the issue at Month 2–4).


    View on ClawHub
    TERMINAL
    clawhub install bookforge-retention-phase-intervention-selector

    πŸ§ͺ Use this skill with your agent

    Most visitors already have an agent. Pick your environment, install or copy the workflow, then run the smoke-test prompt above.

    πŸ” Can't find the right skill?

    Search 60,000+ AI agent skills β€” free, no login needed.

    Search Skills β†’