In this guide
Choose a question the measure can answerDefine the event and periodUse a hypothetical count carefullyTurn the review into a concrete improvementA small team can learn from its records by defining what each count means. Submitted requests, confirmed completions, reopened cases and unresolved questions are different events. A useful review explains its period and denominator rather than presenting a bare percentage as a promise.
Choose a question the measure can answer
For example, count how many cases are waiting for a provider answer at the review time. That describes a workload state. It does not measure service uptime, catalogue quality or every customer’s experience. Keep the measure tied to the operational question.
Define the event and period
State when the review starts and ends, what qualifies as a case and how duplicates are handled. If an order completes but a separate display question remains, decide which state each measure records. Do not change the definition halfway through the comparison without explaining it.
| Measure | Needed definition |
|---|---|
| Pending cases | Which unresolved state and snapshot time? |
| Confirmed actions | What evidence counts as completion? |
| Repeated contacts | Same case or new issue? |
| Reopened cases | Which closure and reopening events? |
Use a hypothetical count carefully
If a made-up review has 20 cases and 5 are awaiting information, that is 25 percent of those 20 cases under that definition. It is not a real Dream statistic. Label the example and preserve the numerator and denominator so the calculation can be checked.
Turn the review into a concrete improvement
A repeated missing order reference might justify clearer intake instructions. A handover gap might justify a better next-action field. Avoid broad claims such as guaranteed faster support unless actual evidence establishes the result. Keep customer identities and credentials out of aggregate public examples.
Questions about this guide
Can a case-completion percentage be advertised as service uptime?
No. Those measure different things. A support case state does not establish continuous service availability. Use a measure only for the question its definition and evidence actually support, and avoid rebranding an operational count as a broader performance claim.