Making outcomes useful.
WHAT A REPORT MUST ANSWER
Did it pass What failed Why it failed Is this new
WHY THAT LAST QUESTION
A known failure and a new one warrant different responses.
WHAT EVERY FAILURE SHOULD INCLUDE
The test name, describing the behaviour The assertion that failed, with actual and expected A screenshot, for interface tests Logs from the period The environment and version
WHY ACTUAL AND EXPECTED SPECIFICALLY
Without both, diagnosis starts from nothing.
WHAT SCREENSHOTS SHOULD CAPTURE
The moment of failure.
WHAT VIDEO ADDS
The sequence leading to it, which frequently explains more.
WHAT TO RETAIN
Artefacts long enough to investigate.
WHAT TRENDS TO SHOW
Pass rate over time Duration over time Tests failing repeatedly
WHY DURATION TRENDS MATTER
Growth is gradual and unnoticed until the suite is too slow.
WHAT TO SURFACE PROMINENTLY
New failures, distinguished from known ones.
WHAT TO AVOID
Reports requiring interpretation Reports nobody reads Failures reported without context
WHAT TO INTEGRATE WITH
Wherever the team already looks.
WHY
A report in a place nobody visits is not a report.
