SYSE.35:4.3 - Define observations and three outcomes
Select observations that can reveal the consequential failures of the changed user task. Include correctness, delay and missing results as the task requires. Infrastructure indicators can help explain effects but do not automatically define user success.
Set the evaluation interval, observation delay and adequate-evidence conditions before judging the candidate. Use an appropriate statistical or domain Method when a quantitative inference requires it. A handful of successful attempts cannot establish a rare-failure probability merely because a tool returns a percentage.
| Outcome | Meaning and dependent action |
|---|---|
| Proceed basis | The stated evidence conditions and acceptance comparison are met for this population and interval; the authorized next exposure may be considered. |
| Stop basis | A stated unacceptable effect or violated operating condition has been observed; apply the qualified restriction or recovery. |
| Inconclusive | Required observations, attribution or discrimination are insufficient; do not convert absence of a failure signal into a pass. |
An inconclusive result can lead to a longer permitted observation, a better measurement, a different representative exercise or abandonment of this exposure strategy. Each changes an identified missing condition; repeatedly waiting on irrelevant traffic does not.