Support Reporting
Part of Support reporting technology
Checking the denominator behind support service metrics
Audit support percentages by identifying eligible cases, time anchors, exclusions and whether the unit is a ticket or an event.
Before using a support percentage, identify what the fraction counts below the line. Write the numerator and denominator in words. Check which records were eligible, which event placed them in the period and whether the unit is a ticket, conversation, reply or target instance. The same percentage can describe different customer populations.
Reconstruct the fraction
For each rate, record the qualifying event, eligible population, unit counted, time anchor, filters and treatment of records without a measured outcome. “Good ratings divided by rated tickets” answers a different question from “rated tickets divided by surveyed tickets”. These are different possible definitions; check what the report actually counts.
A proposed report note might read:
First-response target met among replied-to tickets: qualifying tickets with a first reply within target ÷ qualifying tickets with a first reply in the selected period. Show eligible tickets still awaiting a first reply separately.
This is a suggested definition, not a platform default. Label the vendor measure separately if its rules differ.
Different interpretations of support satisfaction percentage
- Good ratings divided by rated tickets
- Measures quality of feedback from those who responded
- Rated tickets divided by surveyed tickets
- Measures survey response rate among sent surveys
- Good ratings divided by all tickets
- Measures overall satisfaction across all customer interactions
Look for population changes
| Reported result | Denominator question |
|---|---|
| First-response time or compliance | Are unanswered contacts excluded because no qualifying first reply exists yet? |
| Resolution time or compliance | Are open cases excluded until resolution, and which completion event is used? |
| Satisfaction percentage | Is the base all tickets, surveyed tickets or tickets with a rating? |
| SLA achievement | Is each ticket counted once, or can it contribute several target instances? |
Freshdesk Analytics documents that only tickets assigned to an Agent or a Group are counted in its calculations. It also says that when tickets are merged, only the primary ticket is counted for created or resolved categories, not the secondary tickets. Replies and public notes are considered responses in most cases.
Zendesk describes SLAs as policies that specify and measure response and resolution times, and says its SLA reporting dashboards provide information for each SLA metric. Help Scout’s Email report measures team interactions with email messages, and its reports offer metric definitions and conversation drilldowns.
Check the report’s definition and records rather than assuming that similarly named measures use the same population.
Time anchors can change the denominator even when the filter says “this week”. Creation-date, first-reply-date and resolution-date cohorts contain different cases.
Intercom notes that metrics using the Conversation created at time anchor also include conversations with no customer message, such as outbound or proactive conversations that never received a reply. Record the anchor when sharing a chart or comparing periods.
Key denominator considerations in support reporting
- Unanswered contacts excluded?
- Check if no first reply = exclusion from first-response metrics
- Open cases included in resolution time?
- Only resolved cases typically count; check completion event
- Survey coverage impact on satisfaction
- Satisfaction % depends on base: all tickets, surveyed, or rated
- SLA targets per ticket
- Tickets may contribute to multiple SLA instances; confirm counting method
Check the figure against records
Use a small set of cases around the reporting boundary: an unanswered case, an open case, a reopened case, a surveyed case without a rating and a ticket with several subsequent replies. For each, mark whether it belongs in the numerator, denominator, both or neither. Compare that decision with a report drill-down or export if the proposed plan provides one.
For example, nine successes among ten measured cases gives 90% of measured cases. If thirty contacts were eligible and twenty have no measured outcome, the same nine successes represent 30% of eligible contacts with a known success.
Do not present that second figure as an SLA compliance rate without a rule for unmeasured contacts. Show the unmeasured count so readers can judge coverage.
If the tool cannot expose the records behind a figure, document that limit. Before treating a change in the rate as improved service, inspect changes in filters, survey coverage, automation and case mix.
