
How to Diagnose Bugs in a Feature Flag Rollout
When a rollout bug affects only some accounts, capture the actual flag variation and context, compare controlled paths, and decide when to pause exposure.
Read the guidePractical field notes
Research-backed, practical guides for collecting better customer context, triaging issues calmly, and turning reports into work developers can ship.

When a rollout bug affects only some accounts, capture the actual flag variation and context, compare controlled paths, and decide when to pause exposure.
Read the guide
A customer says the app is slow while uptime remains green. Define the affected task, compare real-user and service timings, find the cohort, and choose a proportionate response.
Read the guide
A practical two-ledger method for prioritizing accessibility bugs by blocked customer tasks, workaround quality, evidence, and WCAG conformance.
Read the guide
Set an update contract for unresolved software bugs, choose cadence from customer impact, and write useful progress updates without inventing a fix date.
Read the guide
A practical investigation method for intermittent customer bugs: document failed attempts, test meaningful differences, collect targeted evidence, and pause with a clear restart condition.
Read the guide
Turn a customer's proposed feature into an evidenced problem brief, test competing explanations, and make a clear review promise without committing to delivery.
Read the guide
Review aging customer issues with distinct retirement reasons, protected cases, evidence-request deadlines, and reopening rules that keep unresolved problems visible.
Read the guide
A practical questioning protocol for turning a live customer voice report into an actionable bug record without making the customer diagnose the problem.
Read the guide
A practical framework for choosing supported browser combinations, handling exceptions, testing accessibility, and routing customer compatibility reports.
Read the guide
A 90-second review method for checking AI-generated type, priority, routing, spam, and duplicate suggestions before they change the queue.
Read the guide
A four-gate declaration test for deciding when one customer report needs incident response, plus a clean escalation handoff and stand-down rule.
Read the guide
A release-to-reply protocol for verifying customer-reported fixes, choosing who to notify, writing a useful update, and handling partial fixes or regressions.
Read the guide