Skip to content
Back to Architecture Decisions

PRD Review Checklist

A pre-build checklist that catches under-specified PRDs before engineering wastes a sprint.

Why

Engineering sprints are expensive — a two-week sprint spent on an ambiguous PRD is two weeks of rework. A review checklist applied before sprint planning catches the common gaps: missing out-of-scope, unmeasurable success metrics, acceptance criteria that are vibes not tests. The checklist takes 15 minutes and saves, on average, a week of mid-sprint clarification churn. It also creates a shared quality bar so PRD quality doesn't depend on which PM wrote it.

When

Before sprint planning, by a reviewer who did not write the PRD. Takes 15 min. Skip only for changes < 1 day of work.

How

1. Problem statement names a specific user and a measurable pain? (Y/N)
2. Out-of-scope section lists ≥3 things explicitly NOT being built? (Y/N)
3. Every requirement has acceptance criteria in Given/When/Then? (Y/N)
4. Success metrics are measurable with a baseline and target? (Y/N)
5. NFRs (performance/reliability/security) stated for user-facing features? (Y/N)
6. Open questions either resolved or have a decision date? (Y/N)
7. No requirement describes implementation (only user-visible behavior)? (Y/N)
8. PRD links to the validated idea and relevant ADRs? (Y/N)
→ If any N, send back to author with the specific gap before planning

Example

PRD review for "release-notes-generator":
✓ 1. Problem: release managers spend 2h/release (measurable)
✗ 2. Out-of-scope missing → send back, ask for ≥3 explicit exclusions
✓ 3. All 4 requirements have Given/When/Then
✓ 4. Metric: time-to-generate < 30s (baseline: 2h manual)
✗ 5. No NFR for the 200-commit case → ask for p99 latency target
→ 2 gaps found, returned to author. Re-review in 1 day.

Common Problems

  • ⚠The reviewer rubber-stamps because they trust the PM — trust is good, but the checklist exists to catch the gaps trust misses; actually check each item.

Best Practices

  • ✓Rotate the reviewer so the same person doesn't develop blind spots; a fresh reviewer catches gaps a habitual one skims.
  • ✓Track which checklist items fail most across PRDs — if 'out-of-scope' fails 60% of the time, coach all PMs on that section.

Related Resources