Call Center Productivity Benchmarks — What to Measure Against and What the Numbers Mean
Benchmarking is only useful if you know what the benchmarks actually are, what they mean in context, and what action to take based on where you fall. Most call center managers know their own numbers but have no reliable reference point for whether those numbers are good, average, or a problem.
This post provides the benchmark ranges for the metrics that matter most in call center and BPO operations — and more importantly, explains what to do when your numbers fall outside the range. For deeper coverage of individual metrics, see our call center KPI guide and BPO KPI guide.
How to use benchmarks correctly
Benchmarks are reference ranges, not targets. A benchmark tells you where most call centers of similar size and type operate. It does not tell you where your specific operation should operate, because that depends on your service type, client contracts, channel mix, and cost structure.
Three rules for using benchmarks:
-
Compare against your own segment. A 20-agent inbound support center and a 500-agent outbound sales operation have different benchmark ranges for nearly every metric. Industry-wide averages blend these together and are misleading for both.
-
Trend matters more than position. If your AHT is at the high end of the benchmark range but has been declining steadily for 6 months, you are improving. If it is in the middle of the range but has been climbing for 3 months, you have a developing problem.
-
Metrics interact. You cannot benchmark one metric in isolation. An operation with low AHT but low FCR is rushing calls — the AHT looks good but total handle time (including repeat calls) may be higher than an operation with higher AHT and higher FCR. Always look at related metrics together.
Service delivery benchmarks
These metrics measure whether your operation is meeting customer expectations.
Service level
| Metric | Typical benchmark range | What it means |
|---|---|---|
| Service level (% answered within threshold) | 80/20 to 80/30 | 80% of calls answered within 20–30 seconds |
| Abandonment rate | 3–8% | Percentage of callers who hang up before reaching an agent |
| Average speed of answer (ASA) | 20–40 seconds | Average time callers wait before connection |
If you are below the range (service level under 80/30):
- Check scheduled coverage vs. actual coverage by interval — the problem is usually concentrated in specific time windows, not all day
- Check adherence — agents may be scheduled but not on the phones
- Check whether your absence buffer is adequate for your actual unplanned absence rate
If you are above the range (service level consistently above 90/20):
- You may be overstaffed. Check occupancy — if it is below 70%, agents are idle and you are paying for unused capacity
- Consider whether the excess staffing could be redeployed to reduce overtime on other shifts
First contact resolution (FCR)
| Metric | Typical benchmark range | Notes |
|---|---|---|
| FCR (phone) | 70–75% | Varies significantly by complexity — technical support may be 55–65%, simple inquiries may be 85%+ |
| FCR (email/chat) | 60–70% | Lower due to asynchronous nature and complexity filtering |
If you are below range:
- Analyze repeat contact reasons — the top 3–5 reasons typically account for the majority of callbacks
- Check whether agents have the authority and system access to resolve issues on the first contact, or whether unnecessary escalation policies force callbacks
- Review QA evaluations for whether agents are confirming resolution before ending the interaction
If you are above range:
- Verify your measurement method. Some operations undercount repeat contacts by not tracking same-customer callbacks within 24–72 hours. A high FCR that does not account for same-day callbacks may be overstated.
Agent productivity benchmarks
These metrics measure how efficiently agents handle their workload.
Handle time and utilization
| Metric | Typical benchmark range | Notes |
|---|---|---|
| Average handle time (AHT) | 4–8 minutes | Varies enormously by call type — billing inquiries may be 3 min, technical troubleshooting may be 12 min |
| Occupancy | 75–85% | Percentage of logged-in time spent handling contacts or in after-call work |
| Agent utilization | 85–92% | Percentage of scheduled time the agent is logged in and available (includes occupancy + available/idle time) |
| Calls per agent per hour | 8–15 | Inverse of AHT — directly depends on call complexity |
| After-call work (ACW) | 30–90 seconds | Time spent on documentation after the call ends |
| Schedule adherence | 90–95% | Percentage of time the agent is in the correct state (on call, on break, etc.) per the schedule |
AHT — what the number actually tells you:
AHT is the most commonly benchmarked metric and the most commonly misused. A low AHT is not inherently good. What matters is the relationship between AHT and quality:
| AHT | FCR | What it means |
|---|---|---|
| Low | High | Efficient resolution — agents handle calls quickly and correctly |
| Low | Low | Agents are rushing — short calls but problems are not resolved, generating callbacks |
| High | High | Thorough but slow — agents resolve issues but take too long, possibly due to system limitations or process complexity |
| High | Low | Fundamental problem — agents spend a long time on calls and still do not resolve them. Training, tools, or process issues |
Occupancy — the burnout threshold:
Occupancy above 85% for sustained periods means agents are handling calls back-to-back with minimal recovery time. Short-term (during a volume spike), this is manageable. Long-term, it drives burnout and attrition. If your occupancy is chronically above 85%, you need more agents — not more effort from existing ones.
Occupancy below 70% means agents are idle for significant portions of their shift. This is a scheduling problem — either the shift is overstaffed for its volume, or volume has declined and the schedule has not been adjusted.
Workforce management benchmarks
These metrics measure how well you manage your workforce as a whole.
| Metric | Typical benchmark range | Notes |
|---|---|---|
| Annual attrition rate | 30–45% (industry average) | High-performing operations: 20–30%. BPOs often higher than captive centers |
| Unplanned absence rate | 5–8% on any given day | Higher on Mondays, Fridays, and day after holidays |
| No-call no-show rate | 1–3% | Above 3% indicates attendance policy or engagement problems |
| Training time to proficiency | 2–6 weeks | Depends on complexity — simple support may be 2 weeks, technical/regulated may be 8+ weeks |
| Shrinkage | 25–35% | Total non-productive time as percentage of paid hours (breaks, meetings, training, PTO, absences) |
Attrition — what the benchmark does not tell you
The industry-wide average of 30–45% annual attrition blends high-attrition operations (60%+) with low-attrition ones (15–20%). The average is not a target — it is a reflection of an industry that has historically underinvested in agent retention.
More useful than comparing against the industry average:
- Track your own attrition monthly and identify trends — is it rising, falling, or stable?
- Segment by tenure — high attrition in the first 90 days is an onboarding or hiring problem; high attrition after 12 months is a career progression or compensation problem
- Calculate the cost — each departure costs roughly 3–4 months of the agent's fully loaded salary in recruiting, training, and ramp-up time. See our attrition management guide for the full calculation
Shrinkage — why your number is probably wrong
Shrinkage is one of the most important scheduling inputs and one of the most commonly miscalculated. If you underestimate shrinkage, your schedules will be understaffed because you planned for agents who are not actually available.
| Shrinkage component | Typical range | Often missed? |
|---|---|---|
| Breaks (paid) | 5–7% | No |
| Lunch (unpaid, but scheduled) | 0% (unpaid) | N/A |
| PTO / vacation | 5–8% | Sometimes |
| Unplanned absence (sick, NCNS) | 5–8% | Sometimes |
| Training | 2–4% | Often |
| Team meetings / coaching | 2–3% | Often |
| System downtime | 1–2% | Often |
| Late arrivals / early departures | 1–2% | Often |
| Total | 25–35% |
Operations that only count breaks and PTO in their shrinkage calculation (10–15%) are underestimating shrinkage by half — and understaffing every shift as a result.
Cost benchmarks
These metrics measure the financial efficiency of your operation.
| Metric | Typical benchmark range | Notes |
|---|---|---|
| Cost per call | $3–$8 (inbound support) | Varies by complexity, geography, and channel |
| Cost per agent hour (fully loaded) | $15–$30 (US domestic) | Includes wages, benefits, facilities, technology, supervision |
| Overtime as % of total labor hours | Fewer than 5% | Above 5% is structural understaffing, not occasional gap-filling |
| Revenue per agent (sales/collections) | Varies by industry | Track month-over-month trend rather than comparing to industry |
| Supervisor-to-agent ratio | 1:12 to 1:20 | Below 1:12 is management-heavy; above 1:20 reduces coaching capacity |
Cost per call — what drives the number
Cost per call is the metric most often cited in benchmarking discussions, but its components matter more than the total:
Cost per call = Total operating cost ÷ Total calls handled
If your cost per call is $6 and the benchmark is $5, the question is where the extra dollar comes from:
| Cost driver | If above benchmark | Action |
|---|---|---|
| Agent wages | Higher wages in your geography or for your skill requirements | May be unavoidable — focus on other cost drivers |
| Overtime | Overtime exceeding 5% of labor hours | Fix scheduling to reduce structural overtime |
| Attrition | High turnover driving constant recruiting/training costs | Address retention drivers |
| AHT | Longer calls than benchmark for similar call types | Investigate root cause — training gaps, system limitations, unnecessary process steps |
| Occupancy | Low occupancy (idle agents) inflating cost per productive hour | Adjust scheduling to match volume |
| Supervision ratio | More supervisors per agent than benchmark | Evaluate whether the ratio is justified by quality/training needs or is a legacy structure |
BPO-specific benchmarks
BPOs have additional metrics that single-client operations do not track:
| Metric | Typical benchmark range | Notes |
|---|---|---|
| Billable utilization | 80–88% | Percentage of agent paid hours that are billable to a client |
| Non-billable time | 12–20% | Training, bench time between accounts, internal meetings |
| Client profitability margin | 15–25% gross margin | Below 15% on a sustained basis is unsustainable |
| Invoice accuracy | 98%+ | Fewer than 2% of invoices requiring correction |
| Back office hours per client | Declining over time | Track monthly to identify clients consuming disproportionate admin effort |
Billable utilization is the BPO's most important efficiency metric. Every percentage point of improvement translates directly to revenue:
- 100 agents at $15/hour average billing rate
- 2,080 paid hours per agent per year
- Each 1% improvement in billable utilization = 2,080 additional billable hours = $31,200 in annual revenue
A BPO running at 80% billable utilization versus one running at 86% — same headcount, same cost base — generates $187,200 more annual revenue per 100 agents.
Building a benchmarking practice
Benchmarking is not a one-time exercise. It is a recurring practice that compares your current performance against both external references and your own historical data.
What to benchmark monthly:
- Service level, AHT, FCR, occupancy, adherence — the operational metrics that change week to week
- Compare against your own trailing 3-month average, not just an external benchmark
What to benchmark quarterly:
- Attrition, absenteeism, shrinkage, cost per call, overtime percentage — the structural metrics that change more slowly
- Compare against the ranges in this post and against your own prior quarters
What to do with the results:
| Your position | Action |
|---|---|
| Within benchmark range, stable trend | No action needed — maintain current practices |
| Within range but trending worse | Investigate before it exits the range — early intervention is cheaper |
| Below range | Diagnose the root cause using the guidance above, build a specific improvement plan |
| Above range | Verify your measurement is accurate, then examine whether performance can be sustained or whether you are running agents too hard |
The most valuable benchmarking comparison is not your operation vs. an industry average — it is your operation this quarter vs. your operation last quarter. External benchmarks tell you where you are relative to others. Internal trends tell you whether you are getting better or worse. Both matter, but the internal trend is more actionable.
