Call Center Metrics

Balanced metrics scorecard: five domains with key KPIs and targets.

Call center metrics (also contact center KPIs) are the quantitative measures used to evaluate the performance, efficiency, and effectiveness of contact center operations. These metrics form the measurement layer of workforce management, providing the data that drives forecasting, scheduling, real-time management, and performance management decisions.

Selecting, defining, and balancing the right metrics is among the most consequential decisions in contact center leadership. Poorly chosen KPIs create perverse incentives — agents optimize for what is measured, not what matters. This page provides a comprehensive reference for the metrics that matter, how they relate, and how to structure them into a measurement system that drives the right outcomes.

For the organizational framework governing which metrics belong at which level, see WFM KPI Hierarchy and Reporting Cadence. For a quick-reference glossary of all terms, see Contact Center Metrics Glossary.

Metric Taxonomy

Contact center metrics are organized into five domains that map to the CX/Cost/EX triad. No single domain should be measured in isolation — a balanced measurement system draws from all five.

Accessibility Metrics

Accessibility metrics measure how easy it is for customers to reach a live agent (or resolve via self-service). They are the most visible indicators of staffing adequacy.

Metric	Definition	Industry Benchmark	Full Article
Service Level	Percentage of contacts answered within a target threshold (e.g., 80% in 20 seconds)	80/20 traditional; leading centers target 80/15 or 90/15	Service Level
Average Speed of Answer (ASA)	Mean wait time for contacts answered by an agent	20–30 seconds	Average Speed of Answer (ASA)
Abandonment Rate	Percentage of contacts where the customer disconnects before agent answer	3–5% acceptable; <3% best-in-class	Abandonment

Service level, ASA, and abandonment rate are mathematically linked — they are three views of the same underlying queue dynamics governed by Erlang C and arrival randomness. Setting a target for one constrains the other two.

Efficiency Metrics

Efficiency metrics quantify how productively agent time is used. They are the primary levers for cost management but must be balanced against quality and employee experience.

Metric	Definition	Industry Benchmark	Full Article
AHT	Mean duration of talk time + hold time + after-call work per contact	6–10 minutes depending on complexity	Average Handle Time
Occupancy	Percentage of logged-in time agents spend handling contacts or in ready state waiting	75–85% optimal; >85% risks burnout	Occupancy
Contacts per Agent per Hour	Throughput rate measuring agent productivity	Varies by channel and complexity	Performance Management
Transfer Rate	Percentage of contacts transferred to another agent or queue	<10% target; >15% signals routing or training issues	Skill-Based Routing (related)
Schedule Adherence	Percentage of time agents are in the correct scheduled state at the correct time	90% typical target; 85–95% acceptable range	Adherence and Conformance
Schedule Conformance	Whether agents work total scheduled hours regardless of when	≥95% target	Adherence and Conformance

The distinction between adherence and conformance matters: an agent who takes lunch 30 minutes late but works the full shift has perfect conformance but poor adherence. Both are needed — adherence drives interval-level staffing accuracy, conformance drives cost control.

Quality Metrics

Quality metrics assess whether interactions meet standards for accuracy, professionalism, and regulatory compliance. They are the hardest to measure at scale and the most susceptible to evaluator bias.

Metric	Definition	Industry Benchmark	Full Article
Quality Score	Composite score from QA evaluation of sampled interactions	80–85% average; calibration variance <5% across evaluators	Quality Management
Compliance Rate	Percentage of interactions meeting regulatory and policy requirements	95%+ required in regulated industries	Quality Management
Calibration Variance	Consistency of quality scoring across evaluators	<5 percentage points between evaluators	Quality Management
Coaching Completion	Percentage of scheduled coaching sessions actually delivered	>90% target	Coaching and Agent Development
FCR	Percentage of contacts resolved without follow-up contact	70–79% good; ≥80% world-class	First Contact Resolution

FCR sits at the intersection of quality and customer experience. SQM Group research shows that each 1% improvement in FCR correlates with a 1% improvement in customer satisfaction and a 1% reduction in operating costs.

Financial Metrics

Financial metrics connect operational performance to business outcomes. They are the language executives speak and the ultimate justification for WFM investment.

Metric	Definition	Industry Benchmark	Full Article
Cost per Contact	Total operational cost divided by contacts handled	Median $13.50 assisted; $1.84 self-service (Gartner)	Workforce Cost Modeling
Cost per FTE	Total employment cost per full-time equivalent agent	Varies widely by geography and channel	Workforce Cost Modeling
Shrinkage	Percentage of paid time agents are unavailable for customer contacts	30–35% typical (includes PTO, training, meetings, breaks)	Shrinkage
Overtime Percentage	Hours worked beyond scheduled as a percentage of total hours	<5% target; >10% signals chronic understaffing	Employee Scheduling (related)
Attrition Rate	Annual agent turnover rate	30–45% industry average; <25% best-in-class	Attrition and Retention

The 7:1 cost ratio between assisted and self-service contacts explains the economic pressure toward automation and AI containment. However, deflecting contacts that customers want human help with damages satisfaction and lifetime value.

Planning Metrics

Planning metrics evaluate the accuracy of the WFM process itself — forecasting, scheduling, and capacity planning. They are leading indicators: degradation in planning metrics precedes degradation in all other categories.

Metric	Definition	Industry Benchmark	Full Article
Forecast Accuracy	Deviation between predicted and actual contact volume or AHT	±5% at daily level; ±10% at interval level	Forecast Accuracy Metrics
Net Staffing	Difference between scheduled and required staff per interval	Zero variance is ideal; ±1 FTE tolerance typical	Scheduling Methods
Schedule Efficiency	How well generated schedules match staffing requirements curve	>90% coverage of required positions	Schedule Generation

Key Metrics in Depth

Service Level

Service level is the most widely used accessibility metric in contact centers and the primary input to Erlang C-based staffing calculations. The traditional 80/20 standard (80% of contacts answered within 20 seconds) dates to the early days of ACD technology and has no scientific basis — it persists because it represents a reasonable balance between customer tolerance and staffing cost.

Service level follows a non-linear cost curve: improving from 80% to 90% in 20 seconds requires disproportionately more staff than improving from 70% to 80%. This non-linearity means the cost of each incremental percentage point increases as the target rises — a critical consideration in cost modeling. See Service Level for the full mathematical treatment.

Average Handle Time

AHT = Talk Time + Hold Time + After-Call Work (ACW). It is simultaneously the most measured and most misunderstood metric in contact centers. The industry average ranges from 6 to 10 minutes depending on contact type and complexity.

Why AHT targets are dangerous: Agents under AHT pressure will shorten conversations, skip discovery questions, and avoid warm transfers — all of which reduce FCR and generate repeat contacts. Research consistently shows that reducing AHT through agent pressure increases total contact volume and total cost. The correct approach is to reduce AHT through process improvement (better tools, simpler policies, knowledge management) while holding agents accountable for quality, not speed. See Average Handle Time for optimization strategies.

Occupancy

Occupancy measures the percentage of available time agents spend handling contacts versus waiting for the next contact. It is driven primarily by contact volume and staffing levels — individual agents have almost no control over it.

The 85% threshold is well-documented as a critical inflection point. Centers operating above 85% occupancy see measurable increases in agent fatigue, error rates, and attrition. One major telecom reported a 12-point CSAT decline when occupancy increased from 78% to 88%. COPC benchmarking data shows centers operating between 75–85% occupancy consistently outperform those above 90% in both customer satisfaction and agent retention.

Occupancy is inversely related to service level at any given staffing level — this is one of the fundamental tradeoffs in contact center management. See Occupancy for the mathematical relationship.

First Contact Resolution

FCR is widely regarded as the single most impactful metric in contact centers because it simultaneously improves customer satisfaction, reduces cost, and reduces agent workload. The industry average is approximately 71%, with significant variation by sector: retail averages 78%, insurance 76%, and technical support 65%.

The FCR–AHT tension is real but often overstated. Yes, resolving complex issues on the first contact may require longer handle times. But the repeat contact avoided by achieving first-contact resolution would have consumed far more total handle time than the incremental minutes spent on the first interaction.

Metric Relationships and Tradeoffs

Contact center metrics exist in a system of interdependencies. Optimizing any single metric in isolation will degrade others. Understanding these relationships is essential to sound management.

The Service Level–Cost Tradeoff

Improving service level requires more staff, which increases cost per contact and decreases occupancy. The relationship is governed by Erlang C mathematics:

• Moving from 80/20 to 80/15 requires roughly 5–8% more staff
• Moving from 80/20 to 90/20 requires roughly 15–20% more staff
• The cost curve steepens at higher service levels — each incremental improvement costs more

This is the most fundamental tradeoff in WFM and should be an explicit business decision, not a default.

The AHT–FCR Tension

Reducing AHT through agent pressure almost always reduces FCR, because agents:

• Rush through discovery and miss root causes
• Provide partial answers that generate callbacks
• Transfer rather than resolve
• Skip proper documentation, causing next-agent repeat work

The net effect: lower AHT per contact but higher total handle time across the customer journey. SQM data indicates each 1% FCR improvement saves 1% in operating costs — a far more efficient lever than AHT reduction.

The Occupancy–Attrition Spiral

High occupancy → burnout → attrition → understaffing → higher occupancy on remaining staff. This is a positive feedback loop that can rapidly destabilize operations. The intervention point is before occupancy exceeds 85% sustained. See Occupancy and Attrition and Retention.

The Forecast Accuracy Cascade

Forecast accuracy is the upstream determinant of every other metric. A 10% volume forecast error cascades through:

1. Incorrect staff requirements
2. Over- or under-scheduling
3. Service level miss (understaffed) or cost overrun (overstaffed)
4. Reactive overtime or VTO (voluntary time off) actions
5. Agent dissatisfaction with schedule instability

This is why WFM professionals treat forecast accuracy as the most important planning metric. See Forecast Accuracy Metrics.

Other Key Relationships

• Shrinkage → Scheduling → Service Level: Underestimating shrinkage produces insufficient scheduled staff, degrading service level.
• Adherence → Interval staffing → Service Level: Even perfect scheduling fails if agents don't follow the schedule.
• Quality Score → FCR: Well-trained agents who follow quality standards resolve more issues on first contact.

For the mathematical framework for balancing competing metrics, see Multi-Objective Optimization in Contact Center.

Building a Balanced Scorecard

A balanced scorecard prevents the tunnel vision that results from tracking metrics in isolation. COPC Inc. and ICMI both advocate scorecard approaches as the foundation of contact center performance management.

Scorecard Design Principles

1. Balance across domains: Include metrics from accessibility, efficiency, quality, financial, and employee experience categories. No single domain should dominate.
2. Weight customer outcomes heavily: Best practice is to weight customer outcome metrics (FCR, CSAT, quality) at 40–50% of the total scorecard.
3. Limit total metrics: Agent scorecards should contain 5–7 metrics maximum. More than that dilutes focus and makes it impossible to optimize.
4. Separate controllable from system metrics: Agents can control quality, adherence, and AHT (somewhat). They cannot control occupancy, service level, or abandonment rate. Do not hold individuals accountable for metrics they cannot influence.
5. Include leading and lagging indicators: Adherence (leading) predicts service level (lagging). Quality score (leading) predicts FCR and CSAT (lagging).

Example Balanced Scorecard Weighting

Domain	Metric	Weight
Customer Outcome	FCR	25%
Customer Outcome	CSAT	15%
Quality	Quality Score	20%
Efficiency	AHT (range-based, not minimum)	15%
Efficiency	Schedule Adherence	15%
Development	Coaching Completion	10%

Note that AHT is measured as a range (e.g., 5:30–7:30), not a minimum. This prevents the race-to-the-bottom behavior that a simple AHT target creates.

Metric Hierarchy: Agent to Enterprise

Metrics must be reported at the appropriate organizational level. Not every metric belongs on every dashboard. See WFM KPI Hierarchy and Reporting Cadence for the full framework.

Agent Level

Individual contributors should see metrics they can directly influence:

• Schedule adherence and conformance
• AHT (with context, not as a target)
• Quality score
• FCR (where measurable at agent level)
• Customer satisfaction (per-agent CSAT)

Team/Supervisor Level

Supervisors manage groups and need aggregate views plus variance:

• Team-level service level and abandonment
• Team AHT average and distribution
• Team adherence percentage
• Quality calibration results
• Agent-level exception reports (outliers)

Site/Center Level

Operations managers focus on capacity and cost:

• Service level, ASA, abandonment (aggregate)
• Occupancy and utilization
• Shrinkage (actual vs. planned)
• Forecast accuracy
• Cost per contact and cost per FTE
• Attrition rate and pipeline

Enterprise Level

Executives need strategic indicators:

• Customer satisfaction (CSAT, NPS, CES) trends
• Cost per contact trends and benchmarks
• FCR trends
• Attrition trends and cost of turnover
• Channel mix and migration rates
• AI containment and automation ROI

Reporting Cadence

Different metrics require different reporting frequencies. The general principle: the closer to real-time operations, the higher the frequency.

Cadence	Metrics	Purpose
Real-time (interval/15-min)	Service level, ASA, abandonment, queue depth, agents available, adherence	Intraday management and immediate intervention
Daily	AHT, contacts handled, occupancy, adherence summary, forecast accuracy (daily)	Supervisor coaching and next-day planning
Weekly	FCR, quality scores, schedule efficiency, shrinkage actuals, transfer rate	Performance trending and operational adjustments
Monthly	CSAT/NPS/CES, cost per contact, attrition rate, forecast accuracy (monthly), overtime %	Executive reporting and strategic decisions
Quarterly	Cost per FTE, capacity plan variance, technology ROI, balanced scorecard results	Business reviews and budget planning

Real-time metrics should be displayed on wallboards and supervisor dashboards. The danger is overreacting to real-time data — a single bad interval is noise; a sustained trend requires action. See Real-Time Operations for decision frameworks.

Common Metric Mistakes

Using AHT as an Agent Target

The most prevalent and damaging metric mistake. Agents who are penalized for long handle times will terminate calls prematurely, transfer rather than resolve, and skip after-call documentation. The downstream cost in repeat contacts and customer dissatisfaction far exceeds any efficiency gain. Use AHT as a diagnostic tool, not a performance target.

Ignoring Metric Interactions

Celebrating a service level improvement while ignoring that it was achieved by pressuring agents on AHT (which tanked FCR and CSAT) is not a win. Every metric change should be evaluated in the context of what happened to related metrics.

Measuring Occupancy at the Agent Level

Occupancy is a system-level metric driven by volume and staffing decisions. Individual agents cannot meaningfully control their occupancy rate. Holding agents accountable for occupancy creates confusion and resentment.

Setting Arbitrary Targets Without Baseline

"Our service level target is 80/20 because that is the industry standard" is not a business case. Targets should be set based on customer tolerance analysis, cost modeling, and competitive positioning — not borrowed benchmarks.

Confusing Adherence with Conformance

An agent who works 8 hours but takes every break at the wrong time has 100% conformance and poor adherence. Both matter, but for different reasons: adherence drives interval staffing accuracy, conformance drives cost. See Adherence and Conformance.

Over-Measuring

Organizations that track 30+ KPIs on agent scorecards achieve focus on none of them. Cognitive overload causes agents to optimize for whichever metric their supervisor most recently mentioned. Five to seven well-chosen metrics, consistently reinforced, outperform an exhaustive measurement system.

The Evolution Toward AI-Era Metrics

The rise of conversational AI, chatbots, and agentic AI is forcing a fundamental rethink of contact center measurement. Traditional metrics were designed for a world where every interaction involved a human agent. That world is ending.

Emerging AI Metrics

Metric	Definition	Why It Matters
Containment Rate	Percentage of interactions resolved end-to-end without human handoff	Primary measure of automation ROI; 60%+ target for routine inquiries by 2026
AI CSAT Delta	Difference in satisfaction between AI-resolved and human-resolved interactions	Reveals whether automation is actually serving customers or just deflecting them
Escalation Rate	Percentage of AI-started interactions that require human takeover	Inverse of containment; measures AI failure modes
AI-to-Human Handoff Quality	Customer satisfaction with the transition from bot to human	A poor handoff experience is worse than no AI at all
Deflection vs. Resolution	Whether AI "contained" interactions were actually resolved or just closed	High containment + low resolution = worst of both worlds

The Measurement Shift

Traditional metrics assumed human agents handled every contact. AI changes this in several ways:

• AHT loses relevance for AI interactions — bot processing time is measured in milliseconds, not minutes. AHT becomes a human-only metric.
• Service level bifurcates — AI channels provide instant response (effectively 100% service level), making the metric meaningful only for human-handled contacts.
• FCR must span channels — a customer who fails in the chatbot and calls a human should not count as "first contact" when the human resolves it. True FCR is measured across the entire journey.
• New cost model — cost per AI-resolved contact is a fraction of human cost, but the total cost includes AI platform investment, maintenance, and the cost of escalated contacts that AI made worse.

Organizations that simply add AI metrics alongside traditional ones miss the point. The measurement system itself must evolve to treat human and AI interactions as parts of a single customer journey, not separate channels with separate scorecards. See AI Containment Rate and Its Workforce Implications for the full treatment.

See Also

• Workforce Management — Overview of the WFM discipline
• WFM Goals — The CX/Cost/EX triad
• Service Level — Deep dive on the primary accessibility metric
• Average Handle Time — Deep dive on AHT measurement and optimization
• Occupancy — Agent utilization and the burnout threshold
• Adherence and Conformance — Schedule execution metrics
• First Contact Resolution — The highest-impact quality metric
• Shrinkage — Paid non-productive time
• Average Speed of Answer (ASA) — Wait time measurement
• Abandonment — Customer abandonment metrics
• Contact Center Metrics Glossary — Quick-reference definitions
• WFM KPI Hierarchy and Reporting Cadence — Which metrics at which organizational level
• Customer Experience Management — CSAT, NPS, and CES
• Reporting and Analytics Framework — Data architecture for metric delivery
• AI Containment Rate and Its Workforce Implications — AI-era measurement
• Multi-Objective Optimization in Contact Center — Balancing competing metrics
• Contact Center — Operational environment
• WFM Glossary — Definitions of all WFM terms