Blog · 8 February 2026
An instrumentation checklist for trial products
The minimum event set we ask clients to verify before any trial-to-paid experiment ships.
Experiments fail quietly when events are missing, duplicated, or named differently across platforms. Before we design a conversion test, we walk through a short instrumentation checklist with the client’s analysts.
Confirm signup, trial start, activation milestones, paywall view, plan selection, payment success, payment failure, and cancellation. Each needs a stable name, a timestamp, and user identity that survives device switches when relevant.
Check that trial length and grace periods are encoded as properties, not inferred from chart filters. If your product offers multiple trial lengths, conversion windows must respect that property or the funnel will lie.
Finally, verify that internal users and test accounts are excluded the same way in every tool. One forgotten filter is enough to invent a fake conversion lift.