Agent Productivity in Call Centers — How to Measure It, Why It Varies, and What Actually Improves It
Agent productivity is a measure of how much work an agent completes at an acceptable quality level in a given period. It is not a single number — it is a combination of speed, accuracy, and consistency that must be evaluated together. An agent who handles 12 calls per hour but resolves only 50% on the first attempt is less productive than an agent who handles 9 calls per hour with 80% first call resolution.
Most call centers track operational productivity at the team or site level — daily service level, aggregate AHT, total calls handled. That is necessary but insufficient. The operation is a collection of individuals, and productivity problems almost always trace to specific agents, specific patterns, or specific process gaps that aggregate metrics obscure.
This post focuses on agent-level productivity: how to measure it for individual agents, how to diagnose why it varies, and what to do about it.
Measuring agent productivity
The metrics that matter at the agent level
Not every operational metric is useful for evaluating individual agents. Some metrics (service level, abandonment rate) are outcomes of staffing decisions, not agent behavior. The metrics below are the ones an individual agent can influence.
| Metric | What it measures | How to calculate | Target range |
|---|---|---|---|
| Calls per hour | Volume throughput | Calls handled ÷ hours logged in | Varies by call type — typically 6–12 |
| AHT | Time per interaction (talk + hold + ACW) | Total handle time ÷ calls handled | Varies by call type — typically 240–480 seconds |
| FCR | Resolution effectiveness | Calls resolved without callback within 7 days ÷ total calls | 70–80% |
| QA score | Process compliance and interaction quality | Average of evaluated calls scored against rubric | 85%+ |
| Adherence | Schedule compliance | Time in correct state ÷ scheduled time | 90%+ |
| Occupancy | Time on calls vs. available time | (Talk + hold + ACW) ÷ (talk + hold + ACW + available time) | 75–85% (not a target to maximize) |
| Auxiliary time | Time in non-call states (coaching, training, system issues, personal) | Total aux time ÷ logged-in time | Fewer than 15% of logged-in time |
Why you must measure speed and quality together
Any single metric can be gamed. An agent told to reduce AHT will rush calls, skip documentation, and transfer instead of resolving. An agent evaluated only on QA scores will spend excessive time per call to ensure perfect compliance. Productivity is only meaningful when speed and quality are measured together.
| Speed metric | Quality check | What the combination tells you |
|---|---|---|
| AHT below target | FCR above target | Genuinely productive — fast and effective |
| AHT below target | FCR below target | Rushing — agent is cutting corners, creating repeat calls |
| AHT above target | FCR above target | Thorough but slow — may need process help or skill development |
| AHT above target | FCR below target | Struggling — both slow and ineffective, likely a training or fit issue |
This 2×2 matrix is the most useful diagnostic for individual agent productivity. It immediately separates agents who need process optimization from those who need coaching from those who need fundamental skill development.
Building an agent scorecard
Rather than reviewing agents against a single metric, use a weighted scorecard that combines the metrics that matter for your operation.
Example scorecard (100 points):
| Metric | Weight | Scoring |
|---|---|---|
| FCR | 25 points | 80%+ = 25, 70–79% = 20, 60–69% = 15, below 60% = 10 |
| AHT | 20 points | Within 10% of target = 20, 10–20% over = 15, 20%+ over = 10 |
| QA score | 25 points | 90%+ = 25, 85–89% = 20, 80–84% = 15, below 80% = 10 |
| Adherence | 15 points | 95%+ = 15, 90–94% = 12, 85–89% = 9, below 85% = 6 |
| Calls per hour | 15 points | At or above target = 15, within 10% = 12, 10–20% below = 9, 20%+ below = 6 |
Score interpretation:
| Score | Category | Action |
|---|---|---|
| 85–100 | Top performer | Recognize, retain, consider for mentoring or advancement |
| 70–84 | Solid performer | Standard coaching cadence, address any single weak metric |
| 55–69 | Developing | Targeted improvement plan on weakest 1–2 metrics |
| Below 55 | At risk | Intensive coaching, daily check-ins, evaluate fit |
The weights should reflect your operation's priorities. A BPO with strict SLA penalties for service level might weight AHT and calls per hour more heavily. An operation where repeat contacts are the primary cost driver should weight FCR highest.
Why agent productivity varies
In any call center with 30+ agents handling the same call types, productivity will vary significantly — typically a 30–50% gap between the top and bottom performers in calls per hour, and 15–25 points in FCR. Some variation is normal. The question is whether the variation is within an acceptable range and what is driving the outliers.
The five performance patterns
Most agent productivity issues fall into one of five patterns. Each has a different root cause and requires a different intervention.
Pattern 1: Slow and thorough
Profile: AHT 20–40% above target. FCR at or above target. QA scores high. Adherence good.
What is happening: The agent is doing the work correctly but takes too long. They may be over-documenting, providing more information than the customer needs, or navigating systems slowly.
Root cause: Usually one of:
- Agent is relatively new and has not yet built muscle memory for system navigation
- Agent is a perfectionist who over-explains or over-documents
- Agent's call mix is weighted toward complex call types (check this before assuming an agent issue)
Intervention:
- Review 3–5 calls with the agent, noting specific moments where time could be saved
- Focus on system navigation efficiency — keyboard shortcuts, screen flows, where to find information quickly
- Define documentation standards so the agent knows what level of notes is sufficient
- Set a specific, incremental AHT target (reduce by 30 seconds over 2 weeks, not "get to target immediately")
Pattern 2: Fast and sloppy
Profile: AHT at or below target. FCR below target. QA scores below average. Calls per hour high.
What is happening: The agent is prioritizing speed over quality. They may be skipping verification steps, providing incomplete resolutions, or transferring calls that they could resolve.
Root cause: Usually one of:
- Agent was coached heavily on AHT and interpreted it as "get the customer off the phone fast"
- Agent avoids complex calls by transferring or providing partial answers
- Agent lacks the knowledge to resolve certain call types and defaults to quick workarounds
Intervention:
- Stop coaching this agent on AHT — they do not have an AHT problem
- Review calls where the customer called back within 7 days. Identify what was missed on the original call
- Coach on resolution completeness: "Before ending the call, confirm: have I fully resolved the issue? Does the customer have all the information they need?"
- Monitor FCR weekly for 4 weeks to track improvement
Pattern 3: Inconsistent
Profile: Metrics fluctuate significantly day to day or week to week. Some days are excellent, others are poor. No stable pattern.
What is happening: The agent has the skill but not the consistency. Performance depends on mood, energy, time of day, or call difficulty.
Root cause: Usually one of:
- Agent is affected by high occupancy and burns out during peak periods
- Agent performs well on familiar call types but struggles when encountering unfamiliar ones
- Schedule does not match the agent's energy — poor performance concentrates in specific shifts or days
Intervention:
- Compare performance by shift, day of week, and time of day. If the pattern correlates with schedule, address the schedule
- Identify which call types produce poor performance. If specific types are the problem, provide targeted training on those types
- If the agent performs well during lower-volume periods but poorly during peaks, the issue may be occupancy-driven burnout rather than a skill gap
Pattern 4: Declining over time
Profile: Agent was previously a solid performer but metrics have been trending down over 4–8 weeks. AHT increasing, FCR decreasing, or adherence dropping.
What is happening: Something changed — and it is not usually the agent's skills.
Root cause: Usually one of:
- A process or system change made the job harder (new CRM, added documentation requirements, new call type)
- The agent is disengaged or considering leaving — declining performance is often the first indicator of pending attrition
- Personal issues affecting focus and energy
- A schedule change that the agent finds unsustainable
Intervention:
- Check whether the decline coincides with any operational change. If it does, the agent may need retraining on the new process, or the process itself may be the problem
- Have a direct conversation. "Your performance was strong through February and has changed in the last month. What has shifted?" The agent often knows
- If the cause is engagement, address the underlying issue (schedule, workload, recognition, growth opportunity) before coaching on metrics
Pattern 5: Consistently low
Profile: All metrics below target. AHT high, FCR low, QA scores low. Performance has been poor since onboarding or has been poor for 3+ months with no improvement despite coaching.
What is happening: The agent does not have the foundational skills to do the job at the required level.
Root cause: Usually one of:
- Training did not prepare the agent adequately — they graduated from training without the skills needed
- The agent was hired without the right aptitudes for the role (communication skills, system proficiency, problem-solving ability)
- The agent has been in a struggling state for so long that demoralization is compounding the skill gap
Intervention:
- Define a specific, measurable 30-day improvement plan with clear targets (e.g., "AHT from 520 to 420 seconds, FCR from 55% to 65%")
- Provide intensive support: side-by-side coaching, buddy shifts with a top performer, daily metric review
- If the agent shows meaningful improvement over 30 days, continue the support. If the metrics do not move, the issue is likely fit rather than skill, and the conversation shifts to whether the role is right for the agent
What actually improves agent productivity
Process fixes (highest impact, fastest results)
Process problems disguise themselves as agent performance problems. Before coaching any agent, check whether the process is forcing low productivity.
| Process problem | How it shows up in agent metrics | Fix |
|---|---|---|
| Too many disposition codes | ACW is high for all agents, not just some | Reduce to 15–20 codes organized by category |
| Agents switch between 3+ systems per call | AHT is high across the board, especially hold time (looking up information) | Consolidate or integrate systems so agents have a single screen |
| No standard resolution paths | FCR varies widely between agents on the same call type | Create decision trees or job aids for the top 10 call types |
| Unclear documentation requirements | Some agents write paragraphs, some write nothing — ACW varies from 20 to 120 seconds | Define what must be documented for each call outcome |
| Excessive approval requirements | Hold time is high — agents put customers on hold to get supervisor approval for routine actions | Increase agent authority for common resolutions (credits up to $X, plan changes, waived fees) |
Coaching that changes behavior (medium impact, 2–4 weeks)
Coaching is the primary tool for improving individual agents. But most coaching fails because it focuses on the metric ("your AHT is too high") rather than the behavior that drives the metric.
Effective coaching structure:
| Step | What the supervisor does | Time |
|---|---|---|
| 1. Identify the metric gap | Compare the agent's scorecard to the target. Identify the 1–2 metrics that are furthest from target | Before the session |
| 2. Listen to specific calls | Pull 3–5 calls that illustrate the gap. Note the specific moments where the behavior affects the metric | Before the session |
| 3. Share the observation | "On this call at the 2:40 mark, you paused for 35 seconds after the customer finished speaking. What were you doing at that point?" | 5 min |
| 4. Identify the cause together | The agent explains what they were doing (searching for information, reading a policy, composing notes). The cause determines the intervention | 5 min |
| 5. Define the specific change | "Instead of searching the knowledge base during the call, let's try having the top 5 billing policies open in a tab before your shift starts" | 5 min |
| 6. Set a check-in | "Let's look at your AHT for the next 5 days and review again on Friday" | 1 min |
Total time: 15–20 minutes per coaching session. Frequency: weekly for developing agents, biweekly for solid performers.
What makes coaching fail:
| Coaching failure | Why it fails | Instead |
|---|---|---|
| "Your AHT is 420, it needs to be 360" | Tells the agent the gap but not what to change. The agent does not know how to reduce AHT by 60 seconds | Identify the specific behavior driving the excess time and coach on that behavior |
| Coaching on 5 metrics simultaneously | Agent cannot focus on improving everything at once. Nothing changes | Coach on 1 metric per session. Move to the next when the first improves |
| Coaching without listening to calls | Supervisor is guessing at the cause. The coaching may target the wrong behavior | Always review specific calls before coaching. The calls reveal the actual behavior |
| One coaching session with no follow-up | Behavior change requires reinforcement. A single conversation does not stick | Schedule a follow-up within 5–7 days to review whether the change is showing in the metrics |
Removing barriers (variable impact, ongoing)
Some productivity limitations are not about the agent or the process — they are about what the agent does not have.
| Barrier | How it affects productivity | Resolution |
|---|---|---|
| Slow systems | Every screen load adds 3–5 seconds. Over 60 calls, that is 3–5 minutes of wasted time | Escalate to IT. If system speed is the constraint, no amount of coaching will fix AHT |
| Incomplete knowledge base | Agents put customers on hold to ask supervisors or search for answers that should be documented | Audit the top 20 call types. Ensure resolution steps are documented and accessible |
| No decision authority | Agent must get supervisor approval for routine resolutions (credits, exceptions, plan changes) | Define approval thresholds. Agents should be able to resolve 80%+ of calls without escalation |
| Poor schedule fit | Agent consistently underperforms on specific shifts (late evening, split shifts) | Review whether the agent's assignment matches their availability and performance patterns |
Productivity by tenure
Agent productivity follows a predictable curve. Expectations and interventions should be calibrated to tenure.
| Tenure | Typical performance level | What to expect | Intervention focus |
|---|---|---|---|
| 0–30 days (nesting) | 50–60% of target productivity | Agent is learning systems, processes, and call handling simultaneously. Speed is slow. Errors are frequent | Focus on accuracy and process compliance, not speed. Side-by-side support |
| 30–90 days | 70–85% of target | Agent handles routine calls competently but struggles with edge cases and complex scenarios | Coach on the specific call types where performance drops. Build knowledge depth |
| 90–180 days | 85–95% of target | Agent is competent across most call types. AHT is close to target. FCR is improving | Fine-tune efficiency — system navigation, documentation speed, call control techniques |
| 180+ days | 95–105% of target | Fully proficient. Performance should be stable and consistent | Standard coaching cadence. If performance is still below 90% of target at this stage, reassess fit |
The tenure math matters for workforce planning. If your attrition rate is 5% per month and your average tenure at departure is 8 months, roughly 15% of your workforce at any time is below full productivity. The staffing calculation should account for this — you need more agents than the formula suggests because the formula assumes all agents are at full proficiency.
What not to do
Do not rank agents publicly. Public rankings create anxiety, discourage collaboration, and incentivize gaming. Share individual metrics with each agent privately. Team-level trends can be shared broadly.
Do not set a single calls-per-hour target without context. An agent handling password resets (90-second calls) and an agent handling billing disputes (8-minute calls) should not have the same throughput target. AHT and throughput targets must be set by call type.
Do not coach on occupancy. Occupancy is a function of staffing levels and call volume — not agent behavior. An agent cannot control how quickly the next call arrives. Occupancy above 85% indicates a staffing problem, not an agent productivity problem.
Do not ignore systemic patterns. If 60% of your agents miss the AHT target, the problem is not 60% of your agents — it is the target, the process, or the tools. Diagnose whether the problem is individual or systemic before creating individual improvement plans.
Do not measure productivity without measuring quality. An operation that tracks calls per hour but not FCR will optimize for speed and create repeat contacts that increase total workload. Productivity metrics without quality checks produce the illusion of efficiency while the total handle time — including repeat contacts — increases.
