Why this matters
Action: Monitor crash-free sessions and W1 retention; share daily until stable.
Table of Contents
Templates you can reuse
Slack update template (short)
Insight: [Cohort/segment] shows [pattern], impacting [KPI] by [magnitude]. Likely due to [cause]. Action: [next step]. Expect [metric change] by [date].
One-pager template (long)
- Headline insight
- Metric snapshot (before/after, pp and %)
- Cohorts covered (acq month/week, segment)
- Drivers and evidence (flags, releases, campaigns)
- Decision + owner + timeline
- Risks/assumptions + plan to validate
- Next check-in date
Meeting script (3 minutes)
- Say the headline first.
- Quantify the change (pp and %).
- Show the likely cause (1–2 facts).
- State decision and expected outcome.
- Invite 1 question; park deeper dives.
Tailor your message by audience
- Executives: one-line impact on business metric; clear decision ask.
- PM/Eng: cohorts, flags, dependencies, and specific experiment plan.
- Marketing/CRM: segments, timing, and message hooks to test.
- CS/Support: who is affected, how to triage, scripts to use.
If you have only 30 seconds
State the one-sentence template and the single decision you need.
If you have 2 minutes
Add magnitude, driver evidence, and the validation plan.
If you have 5 minutes
Include a minimal table of cohorts and the before/after numbers.
Visuals that help (even without charts)
When you can’t show charts, use compact text tables:
Cohort W1 W2 W4 Jan 40% 32% 22% Mar* 44% 35% 26% (*post-onboarding)
Always include both pp and % change: +4pp (22% → 26%, +18%).
Exercises
Do the exercise below and compare with the checklist.
- Exercise ex1: Write a concise Slack update using the provided cohort data.
Self-check checklist
- Starts with a one-sentence insight
- Specifies cohorts, KPI, and time window
- Includes magnitude (pp and %)
- Provides a plausible driver with evidence
- States a clear action and owner
- Defines expectation and check-in date
Common mistakes and how to self-check
- Burying the lead: Put the insight first. If your first sentence is context, rewrite.
- Only relative % or only pp: Include both to avoid misreads.
- Over-claiming causality: Use language like “likely due to” unless you ran a controlled test. Add your validation plan.
- Missing decision: Every update should propose an action or ask for one.
- Too many numbers: Keep 1–2 metrics; move rest to details.
Practical projects
- Retention Deep Dive Brief: Take three monthly cohorts, write a one-pager with insight, driver, and experiment plan.
- Lifecycle Messaging Plan: From cohort drop-off at Day 3, propose two messages and a success metric.
- Channel Quality Review: Compare two acquisition channels’ cohorts; recommend budget shift with expected ROI proxy (retention or activation).
Learning path
- Read this subskill and complete Exercise ex1.
- Run a real cohort slice from your product or a sample dataset.
- Share a short update with a peer and apply their feedback.
- Take the quick test to confirm understanding.
Mini challenge
Pick one metric and write the one-sentence insight for a recent cohort change. Timebox to 3 minutes. Then extend it to a 2-minute version.
Optional scenario to use
Feb cohort activation 28% → Apr cohort 33% after tooltip launch; biggest lift on web new users.
Write the 30s and 2m versions.
Quick Test
The quick test below is available to everyone. If you log in, your progress will be saved.
Next steps
- Turn one of the Practical projects into a shareable brief.
- Align with a PM/Marketing partner on the next experiment.
- Schedule a follow-up to revisit the expectation you set.