Stress-test a product sunset before you announce it

Stress-test a product sunset before customers, support, and finance live with a kill decision you cannot quietly rewind.

Sunsets fail when migration paths are vague, when revenue you hoped was dying is still paying the bills, when contracts block the timeline, and when the announcement lands as abandonment instead of a clear path forward. A tidy deprecate-by date is not evidence.

What to put on the table

One sentence for what dies, when it dies, who is migrated where, what is refunded or credited, and the success metric after sunset. Attach usage by segment, revenue at risk, contractual notice periods, support load, and the replacement product's real readiness. If product, sales, and support disagree on who still depends on the feature, stop and reconcile first.

Name the decision you will make if the stress test finds nothing new, and the delay criteria if migration tooling or contract notice is not ready.

Attack surfaces

Customer harm: which segment loses a workflow with no real substitute.
Revenue leakage: logos that churn or expand down because of the kill.
Contract and notice: timelines you promised in writing.
Migration capacity: engineering and support seats versus claimed volume.
Morale and rumor: the internal story that makes key people leave.

Optional counsel seat if regulated customers or data retention rules bind the timeline. Optional angry-customer seat if the kill removes a benefit people already paid for.

How to run it

Feed Pingpong the sunset memo and exhibits. Early passes steelman the kill. Later passes attack from customer, finance, support, and skeptic seats. End with a pass that turns surviving objections into phased dates, credits, or a narrower sunset. Delete invented migration completion rates.

Force a day-after narrative: what support hears, which enterprise accounts ask for side letters, and what competitors publish. If those stories are stronger than your migration plan, fix the package before announce.

Separate sunset of net-new signups from sunset of existing usage. If you stop selling while still supporting, say the two dates and the staffing plan for each. Ask finance and support seats to attack blended timelines until both dates have owners and kill criteria.