Call Centre KPIs: AHT, FCR, CSAT and Service Level
The core call centre KPIs explained with formulas and worked examples: AHT, FCR, CSAT, service level and abandonment, plus how to keep them in balance.
· 4 min read · Summarix team
The core call centre KPIs are average handle time (AHT), first call resolution (FCR), customer satisfaction (CSAT), service level and abandonment rate. Together they tell you whether callers get through, whether their problem gets solved, and what it costs you. This guide gives the formula for each, a worked example and the traps to avoid, whether you call it a call centre or a call center.
Average handle time (AHT)
AHT = (talk time + hold time + after-call work) ÷ calls handled
Say an agent handled 50 calls in a day with 250 minutes of talk time, 25 minutes of hold and 75 minutes of after-call work. AHT = 350 ÷ 50 = 7 minutes.
AHT drives staffing: longer calls mean more agents for the same volume. But pushing AHT down on its own is dangerous. Agents rush, callers phone back, and your real workload goes up. Cut the parts of AHT that add nothing to the caller, such as long holds while agents search for information or after-call note-taking that AI call summaries can replace.
First call resolution (FCR)
FCR = calls resolved on first contact ÷ total calls needing resolution × 100
If 800 of 1,000 issues were solved on the first call, FCR is 80%. The hard part is measuring ‘resolved’. Common methods:
- Repeat-call method: no call from the same number about the same issue within, say, 7 days.
- Survey method: ask the customer whether their issue was resolved.
- Call outcome method: classify each call’s outcome from the transcript (resolved, escalated, pending).
Customer satisfaction (CSAT)
CSAT = satisfied responses (for example 4 or 5 on a 5-point scale) ÷ total responses × 100
If 340 of 400 survey responses score 4 or 5, CSAT is 85%. Watch your response rate: if only a small, self-selected group answers, CSAT can look better or worse than reality. Caller sentiment from transcripts can fill the gap, since it covers every call rather than only those who answer a survey. See call sentiment analysis for how.
Service level and speed of answer
Service level = calls answered within target time ÷ calls offered × 100
A common way of expressing a target is ‘X% of calls answered within Y seconds’, for example 80/20: 80% within 20 seconds. The right target depends on your business and what callers expect. Say 2,000 calls were offered and 1,560 were answered within 20 seconds: service level is 78%.
Average speed of answer (ASA) is the related measure: total wait time for answered calls ÷ answered calls.
Abandonment rate
Abandonment rate = calls abandoned in queue ÷ calls offered × 100
If 140 of 2,000 callers hung up before reaching an agent, abandonment is 7%. Many centres exclude very short abandons (a few seconds) since those are often misdials. Decide your rule and apply it consistently.
Occupancy: the hidden pressure metric
Occupancy = time agents spend handling calls (talk, hold and after-call work) ÷ time logged in and available × 100
If agents are logged in for 420 minutes and spend 360 of them on calls, occupancy is about 86%. Running occupancy very high for long periods leaves no breathing space between difficult calls, which tends to show up later in absenteeism, staff turnover and falling quality. It is worth watching alongside service level, because hitting service level by pushing occupancy ever higher is not sustainable.
Putting them together
| KPI | Tells you | Risk if chased alone |
|---|---|---|
| AHT | Cost and capacity per call | Rushed calls, lower FCR |
| FCR | Whether problems get solved | Longer calls as agents over-handle |
| CSAT | How customers feel afterwards | Skewed by low response rates |
| Service level | Whether callers get through quickly | Overstaffing costs |
| Abandonment | Callers you lost before speaking | Hides callers who never phone again |
Review them as a set, weekly, by team. A good pattern is stable AHT, rising FCR, steady CSAT and service level on target. If AHT drops while repeat calls rise, you have traded quality for speed. For wider support metrics such as ticket backlog and email response time, see customer support KPIs.
Summarix turns connected or uploaded calls into transcripts, outcomes, sentiment and coaching scores, and builds call-trend reports across weeks and teams, which gives you outcome and sentiment data to sit alongside the queue statistics from your phone system.
Turn your spreadsheet into an executive report in under a minute.
Free plan: 5 AI reports a month, no card needed.
Conclusion
Use AHT for capacity, FCR for effectiveness, CSAT and sentiment for experience, and service level and abandonment for access. Define each one precisely, measure the same way every week, and never optimise one at the expense of the others. Then use call analytics for coaching to move them.
Frequently asked questions
What is a good average handle time?
There is no single good AHT: it depends on call complexity. Set a baseline per call type and focus on removing waste such as long holds and manual wrap-up.
How do you calculate first call resolution?
Divide issues resolved on the first contact by total issues and multiply by 100. The usual way to spot unresolved issues is a repeat call about the same issue within a set window, such as 7 days.
What does an 80/20 service level mean?
It means 80% of calls are answered within 20 seconds. It is a common way of expressing a target, but choose the target that fits your callers.
Which call centre KPI matters most?
Many teams treat first call resolution as the most important because it affects cost, workload and satisfaction at once. It should still be read alongside AHT and CSAT.