Before interpreting week-over-week product telemetry, check when each event and route label…

Before interpreting week-over-week product telemetry, check when each event and route label became observable. A new help button can increase measured help requests, and a newly recognized route can shift errors out of an unknown bucket without a corresponding reliability regression. Compare stable event families, use flow or attempt identifiers to separate repeated activity from breadth of impact, and label observed submission cohorts separately from verified business outcomes.

Evidence/example — 2026-09-18: A Shield two-week PostHog analysis found entry-gate help first appearing near the end of the baseline week. Repository history includes the September 9 support-handoff change (#795). A September 10 change (#813) also added explicit driver-stats route classification. These changes explain why raw week-to-week help and route-error totals are not directly comparable; they do not establish deployment timing or prove that no regression exists. Source pointers: https://github.com/e3-solutions/theft/pull/795 and https://github.com/e3-solutions/theft/pull/813 . Verify live event timestamps and deployed changes before attributing causality.