Call center KPIs: what to measure — and what your phone system can't tell you

The three KPI families

Call center KPIs measure three different things. Confusing them is why so many dashboards produce reports nobody uses.

• Phone-system KPIs (ACD): AHT, service level, ASA, abandonment, occupancy. Source: your queue system. Tell you how many and how fast.
• Survey KPIs: CSAT, NPS, CES. Source: post-call or periodic surveys. Tell you how the customer feels — for the small % who respond.
• Conversation-content KPIs: QA score, script adherence, topic mix, sentiment, compliance flags. Source: content of every call. Tell you what actually happened and why.

Phone-system KPIs and their limits

Orientation benchmarks (industry averages, source-dependent): AHT ~6 min, service level 80/20 (arbitrary standard), ASA ~28 s, abandonment ~5–8%, occupancy ~85–90%. Every one of these can look "green" while quality drops — a short call can be a bad call.

Survey KPIs

CSAT typical range 75–85%. NPS varies wildly by industry. CES (Customer Effort Score) is often the best predictor of retention. All three suffer from response bias — the loudest respond.

FCR as the bridge

First Contact Resolution (~70–75% typical, 80%+ target) is the metric that connects operational and quality views. Measured correctly, it tells you why customers repeat contact — not just that they do.

Conversation-content KPIs

QA score against your scorecard, script adherence, topic categories, sentiment, compliance violations. Sampled at 2% under manual review; measurable on 100% of calls with auto-QA. The KPIs that finally answer why.

How to build a KPI set

Pick one KPI from each family. Track the trend, not just the number. Set targets that reflect quality (FCR, QA score) not just throughput (AHT, SL). Review the operational and content metrics together — separately they mislead.