Answer-time measurement
Call center service level: threshold and denominator
Your office may receive an average answer-time report and still need to know how many callers waited beyond its target. A call center service level measures the share answered within a defined time. Before comparing two answering arrangements, agree on the clock, the denominator and the reporting period. This guide provides a measurement specification, not a recommended universal target or a Revo performance claim.
| Field | Decision to record | Audit evidence |
|---|---|---|
| Population | Which number, queue, hours and caller types are included? | Report filters and coverage schedule |
| Clock | Queue-entry timestamp to connected-answer timestamp | Event definitions and sample records |
| Threshold | Chosen target in seconds; equality rule | Approved operating requirement |
| Denominator | Treatment of abandons, short abandons and flowouts | Configured inclusion options |
| Aggregation | Same interval, timezone and calculation version | Counts alongside the percentage |
01 / Answer-time measurement
Write a calculation that another person can reproduce
Genesys defines service level using conversations answered within a time target divided by conversations included in the calculation. Its queue timer runs from queue entry until answer, and an answer at the target counts as within it. Its configurable inclusion rules address abandoned calls, short abandons and flowouts. Those product definitions are useful evidence, but another platform may use a different clock or population.
For your evaluation, write: qualifying answers within the chosen threshold ÷ eligible interactions under the agreed rules × 100. Keep both counts with the result. If a report starts its timer after a menu but another starts when the number is dialed, the percentages describe different experiences. Record any time before the queue separately rather than silently treating it as zero.
02 / Answer-time measurement
Decide what an unanswered interaction does to the denominator
Create an explicit rule for a caller who disconnects, including a short disconnect, and for a call removed by a workflow before an agent answers. An exclusion changes the percentage even when the underlying calls have not changed. Show excluded counts next to included counts so the office can see how much traffic the result omits.
Use an event ledger with interaction ID, queue entry, answer time if any, exit reason and inclusion decision. Check one record of each outcome against the report configuration. Keep a rule change dated: two months calculated with different exclusions should not appear as a continuous improvement trend without an explanation.
03 / Answer-time measurement
Compare like intervals instead of averaging percentages
A small weekday interval and a busy evening interval should not receive equal weight simply because each has one percentage. For a combined result, add the qualifying-answer counts and eligible-interaction counts, then calculate the ratio. Only combine records using compatible definitions. A period with no eligible calls has no observed service level; show it as not applicable rather than perfect performance.
For a home-service office, split reporting by staffed coverage and after-hours responsibility when those arrangements differ. Choose a target from the actual callback and staffing requirement, then examine the associated workload. A threshold on a dashboard does not create additional available staff, prove appointment completion or establish a contractual commitment.
04 / Answer-time measurement
Use service level alongside the measures it cannot replace
Average speed of answer describes an average; service level describes a fraction within a boundary. Abandonment describes callers who leave before an answer. None establishes whether the office collected the correct address, confirmed an appointment or reached an authorized on-call person. Keep those outcomes in their own records.
Before accepting a vendor report, ask for the event definitions, inclusion settings, period boundaries and counts behind its percentage. Compare a small authorized sample with the call log. Label missing timestamps as a measurement gap. Do not assume an AI answering plan publishes or guarantees the same metric as a staffed queue.
Common questions
Make the next step clear.
Resolve the details that matter to your team and callers.
Is 80 percent in 20 seconds the required target?
This guide does not prescribe that target. Choose an answer-time requirement for your operation and document the population and exclusions before using any percentage.
Can two services with the same average answer time have different service levels?
Yes. An average does not show how many individual waits fall on either side of your threshold. Compare the underlying counts and compatible timestamps.
Does an excluded short abandon disappear from call records?
It should remain visible in the event ledger and excluded count. Exclusion from this calculation does not mean the call never occurred.
Sources and current terms
A useful service-level report includes its threshold, eligible population, exclusion rules and counts, so your office can reproduce the percentage.
Review Revo’s current features and plans for the workflow your business needs.
Compare plans