The essentials
See what actually happened.
Read findings in context, understand replay limits, and distinguish clear results from missing coverage.
Check the scope first
Before drawing a conclusion, check the commit, flow, browser, and viewport recorded by the run. Evidence only supports the targets that were actually executed.
A successful run is not a claim about every possible journey through the application. Pay attention to skipped steps, unavailable artifacts, and any limits recorded with the result.
Follow the finding back
An aberration should describe an observable problem and point to evidence. Follow the steps before and after the finding to understand its effect on the user’s task.
A screenshot captures a moment. A timeline or recording can explain how the page reached that state. Console and network evidence can add context, but do not replace the visible experience.
A replay is a recording
DOM replay reconstructs a historical browser execution. It is not the live application. The flow is the recipe that can be run again; the recording is what happened during one execution.
Changing the viewer’s size does not prove that the application worked at another viewport. Verified responsive behavior requires an execution at that size. Each browser and viewport needs its own evidence.
Missing coverage stays visible
A setup failure, interrupted flow, or unavailable recording is a limit on the review. It cannot establish that the affected experience is clear of problems.
Resolve the blocker and rerun the relevant flow before relying on it. Keep the difference between a verified observation and an untested assumption explicit.