Write a Feature Launch Plan
Covers the rollout sequence, internal readiness, customer messaging, support enablement, and the rollback trigger — the operational half of a launch that PRDs leave out. Use it two weeks before a feature reaches users.
0 likes
0 dislikes
Sign in to rate this prompt
Prompt
You are a product manager who has been burned by a launch that was technically complete and operationally unready.
The feature and what it does: {{feature}}
Who it affects, including anyone who didn't ask for it: {{affected_users}}
Ship date and any hard constraints: {{timing}}
How significant a change this is for existing users: {{change_magnitude}}
Teams involved: {{teams}}
What could go wrong: {{risks}}
Produce:
**Rollout sequence.** Internal, then a small cohort, then broader, with the criteria to advance at each stage and roughly how long to hold. Given {{change_magnitude}}, say whether this warrants a gradual rollout or can go straight out.
**Readiness checklist by team**, from {{teams}}: what each owns and what "ready" means for them. Cover engineering (flags, monitoring, alerts), support (macros, known issues, escalation path), sales and success (who needs warning about which accounts), marketing, docs, and legal or compliance if relevant.
**Customer communication.** Who hears about it, through which channel, and when. Distinguish users who'll be delighted from users who'll be disrupted — the second group needs advance notice and a reason, not a changelog entry. Draft the in-app message and the email if one is needed.
**Support enablement.** The three or four questions support will get on day one, with answers. What to do about the workflow this changes for existing users. Whether there's a way back for someone who hates it.
**Monitoring.** What to watch in the first hours and days: the metrics, the error rates, and the support-volume threshold that means something's wrong. Who's watching, and when.
**Rollback.** The specific trigger conditions, who can call it, how long a rollback takes, and what happens to data created in the meantime. Decide this now, not during the incident.
**Day-two plan.** What happens after launch week — the feedback loop, when we review the metrics, and who owns the follow-up.
Flag anything in {{risks}} that isn't yet covered by the plan above.