Case: nightly data refresh reversed a 90-day decay on 3,200 pricing pages

Case: nightly data refresh reversed a 90-day decay on 3,200 pricing pages

Pricing pages were built once, never updated. Over 90 days they lost 41% of clicks as stale prices drove pogo-sticking. Recovery SOP:

— Step 1 (analyst): correlated decay against last-modified date. Gate: pages stale >30 days with rising bounce = decay candidate.
— Step 2 (data): wired a nightly job pulling live prices from the source API, updating only changed fields.
— Step 3 (dev): exposed a real dateModified in schema, updated only when a field actually changed. Gate: no fake timestamp bumps.
— Step 4 (analyst): tracked clicks and dwell over 8 weeks.

Outcome: clicks recovered to 96% of the pre-decay peak. Bounce rate on the set fell 12 points.

Guardrail: bump dateModified only on real content change, or you train Google to distrust it.

Ship gate: don't publish until all boxes are checked.