One thousand senior
collaborators, in a zip file.

You type

What you get back

One structured report. Six sections. No preamble. No pleasantries. Every Next action has an owner and a timeframe.

# 📋 Task summary
Diagnose a 12-point onboarding conversion drop (34% → 22%) inside one week and return a recovery plan.

# 👥 Experts consulted
- Growth analytics lead — funnel attribution, cohort isolation
- Product UX researcher — step-level friction, session replay
- Release engineer — deploy diff, flag rollouts, 3rd-party regressions

# 🔍 Key findings
- A 12-point drop in seven days is not drift — it points to a discrete change.
- Three plausible buckets: recent deploy/flag, 3rd-party regression, acquisition-mix shift.
- Step-level attribution is missing — any fix without it is a guess.
- Session replay + funnel breakdown for the affected cohort is the fastest path to signal.

# ⚖️ Tradeoffs surfaced
- Rollback now restores conversion but hides root cause vs. hold 24h for clean diagnosis.
- Mix-shift is easier to confirm; product regression is more likely given the speed.

# ✅ Recommendations (in priority order)
1. Segment the drop by step, device, OS, locale, source.
2. Pull the deploy log for 48h pre-drop; diff anything touching sign-up.
3. Check 3rd-party status pages in the same window.
4. If a deploy correlates, hide behind a feature flag — don't hard-revert.
5. Ship a step-level conversion canary alert.

# 🎯 Next actions
- [ ] Growth analytics lead — funnel delta table by source/OS/locale — by end of today
- [ ] Release engineer — 48h deploy diff — by end of today
- [ ] Product UX researcher — 20 session replays from affected cohort — by tomorrow morning
- [ ] TBD — ship step-level conversion canary alert — by end of this week