Contact Center KPIs Explained: The Metrics That Matter
A guide to contact-centre KPIs in five families: access, efficiency, quality, resolution and capacity, with definitions, pitfalls and how to choose.
Open any contact-centre system and you will find dozens of metrics. Most are useful in the right hands and distracting in the wrong ones. The skill is not in knowing them all but in choosing a small set that matches what your callers and your business need, and in reading them sensibly.

This guide explains the main KPIs, grouped by what they tell you, with notes on how each can mislead. Detailed articles are linked where they exist, such as Answer Rate and Abandonment Rate, First Call Resolution and Average Handle Time. For setting levels, see KPI Targets: How to Set Them.
The five families of KPI
What each family tells you
- AccessCan callers reach you? Answer rate, abandonment, speed of answer, service level.
- EfficiencyHow much work does each call take? Handle time, after-call work, occupancy.
- QualityHow well are calls handled? Quality scores, compliance, customer ratings.
- ResolutionIs the problem solved? First-call resolution, repeat calls, transfers.
- CapacityDo you have the people you need? Adherence, utilisation, forecast accuracy.
A balanced set includes at least one measure from each family. A centre that watches only access and efficiency tends to rush callers; one that watches only quality may leave queues too long.
Access measures
| KPI | What it measures | Watch out for |
|---|---|---|
| Answer rate | Answered calls divided by offered calls | Definitions of offered and answered |
| Abandonment rate | Callers who hung up while waiting | Short abandons and callbacks |
| Average speed of answer | Mean time from entering the queue to being answered | Averages hide long waits |
| Longest wait | The longest any caller waited | One call can distort a day |
| Service level | Share of calls answered within a set time | The target chosen is a judgement |
Access measures describe what callers experience first. They are also easiest to improve by throwing resources at them, which is why they should be read with efficiency. For depth, see Service Level 80/20 and Average Speed of Answer.
Efficiency measures
| KPI | What it measures | Watch out for |
|---|---|---|
| Average handle time | Talk time plus hold plus after-call work, per call | Rewards rushing if used alone |
| After-call work time | Time spent on notes and tasks after a call | Falls when notes are poor, not only when work is efficient |
| Occupancy | Share of logged-in time spent on calls and related work | Very high values lead to fatigue |
| Transfer rate | Calls passed to someone else | Some transfers are right; many point to routing problems |
Efficiency figures are the most tempting to push. A shorter call is not better if the caller must ring back. See After-Call Work Time, Occupancy and Utilisation and Transfer Rate for each measure in detail.
Quality and customer view
Quality measures ask how well a call was handled, not how fast. They come from two sources: an internal review of sampled calls against a scorecard, and the customer's own opinion, from a short survey after the call.
- Quality scores. A reviewer rates a sample of calls against agreed criteria such as accuracy, tone, following process and accurate notes. See Quality Management for Calls.
- Customer satisfaction. A short rating after the call. See CSAT After Calls.
- Effort and recommendation scores. How easy was it, and would you recommend us. See Customer Effort Score and Net Promoter Score for Support.
Quality figures are noisy. Small samples swing, survey responses come mainly from people with strong views, and reviewers differ. Use calibration sessions so that reviewers score alike, and read the comments, not only the averages.
Resolution measures
Resolution measures ask whether the caller's problem was solved. They are among the hardest to measure well and among the most informative.
Two ways to see resolution
| First-call resolution | Repeat-call rate | |
|---|---|---|
| What it asks | Was it solved on the first contact? | How many callers came back about the same thing? |
| How it is found | Agent tags, survey answers or a check for a repeat within a set period | Matching callers and topics over a window |
| Strength | Direct and intuitive | Based on behaviour, not opinion |
| Weakness | Depends on how the outcome is recorded | Needs reliable caller identification |
Read First Call Resolution and Repeat Call Rate together. A fall in handle time alongside a rise in repeat calls is a warning that speed has been bought at the expense of solutions.
Capacity and planning measures
Capacity measures tell you whether the right people are in the right place. They are the bridge between the forecast and what callers experience.
- Forecast accuracy. How close predicted volumes were to the actual calls, by interval.
- Schedule adherence. Whether staff are working when the schedule says they should be. See Schedule Adherence.
- Shrinkage. The time agents are paid but unavailable, such as breaks, training and leave.
- Callback completion. Whether promised callbacks happen on time. See Callback Completion Rate.
Capacity measures rarely interest callers, but they explain most of what callers feel. A queue that was long at ten on Monday is usually a planning story, not an agent story. Workforce Management and Erlang C Explained give the background.
Choosing and combining KPIs
A sensible way to choose
- State the goalsWhat should callers feel, and what does the business need?
- Pick a few from each familyStart with around eight in total.
- Write definitionsWhat is counted, from which system, over what period.
- Name an ownerOne person responsible for each figure.
- Pair themPut speed beside quality and handle time beside resolution.
- Review and pruneRemove any KPI that has not led to a decision in a quarter.
Be wary of leaderboards. They make a few people feel good and many feel judged, and they invite gaming. Use team-level figures openly, and individual detail privately for coaching. KPI Dashboard for a Small Team shows a lighter-weight set.
A worked example shows the idea. Suppose handle time falls for two months while repeat calls rise. Taken alone, the first figure looks like progress. Placed beside the second, it suggests agents are closing calls early. The remedy is then coaching and better knowledge sources, not a new target. This kind of reading is what turns a list of metrics into a management tool.
Pitfalls to avoid
Common KPI mistakes
- Using a figure without a written definition
- Reporting only averages and no intervals
- Setting targets from a benchmark without checking its definition
- Rewarding speed without checking resolution
- Including test or demo calls in the data
- Changing the phone setup without noting it on the charts
- Using KPIs to rank staff instead of finding problems
Where the numbers live in Call
In Call, the Home page shows calls today, talk time, calls in queue and service level, with a view of the team and alerts. The Reports page offers half-hour intervals with offered, answered and abandoned calls, speed of answer and handle time; agent scorecards; arrivals by hour; and a report builder where you group by agent, queue, outcome or hour and export the result as a CSV file.
Queues have their own answer targets and alert limits. Call quality scoring rates AI-handled calls on several criteria, and quality scorecards support reviews of human calls. Callback lists show how many were done on time, and wallboards bring the main figures to the floor.
Frequently asked questions
Which KPIs should a small team start with?
Calls received, calls missed, longest wait and callbacks due. Add resolution and a customer rating when you can measure them reliably.
What is the difference between handle time and talk time?
Talk time is the time on the call itself. Handle time usually adds hold time and after-call work. Check how your reports define each.
Why do KPIs conflict?
Speed and quality often pull in opposite directions. Pressure to shorten calls may reduce resolution, so read paired measures together.
How many KPIs are too many?
If nobody acts on a figure, it is probably surplus. Around eight, spread across the five families, is a reasonable starting point.
Are industry benchmarks reliable?
Treat them with caution. Definitions differ, and many figures lack a clear source. Your own history is usually a better baseline.
Should KPIs be shown for each agent?
Individual figures are useful for private coaching. Public rankings can encourage gaming and resentment, so share team measures openly.