Guide
Measuring Call Handling Performance
A handful of metrics tell you whether your phone line is working. Several widely used ones actively mislead, particularly average handling time, which rewards rushing customers off the phone.
Worth tracking
- Answer rate, including out-of-hours
- Enquiries resolved on the first call
- Time to first human contact
- Enquiry mix and how it shifts
- Questions nobody could answer
Measure Outcomes, Not Activity
The metrics that survive scrutiny describe what happened to the customer: were they answered, was the matter resolved, and how quickly did a person get involved when needed.
Activity metrics describe what your team did, and they are easy to improve in unhelpful ways. Average handling time falls when staff cut conversations short, which looks like efficiency and produces callbacks.
If you track only one thing, track answer rate including out-of-hours. It is the metric most businesses have never measured and the one that reveals the largest gap.
Which Metrics to Use
| Metric | What it tells you | How it misleads |
|---|---|---|
| Answer rate | Whether callers reach anyone at all | Excludes out-of-hours unless you include it deliberately |
| First-contact resolution | Whether the matter was actually dealt with | Requires defining “resolved” consistently |
| Time to first human contact | Whether enquiries are acted on quickly | Can look good while nothing is completed |
| Escalation rate | Whether the flow scope is right | Low is not automatically better; over-escalation is safer |
| Unanswered questions logged | Exactly what to add to your content | Only captures calls that reached the system |
| Average handling time | Very little on its own | Improves when staff rush people, which is worse service |
| Call volume | Capacity planning input | Rising volume may mean poor self-service elsewhere |
Capture the Baseline First
Once anything changes, the previous state becomes unrecoverable. Whatever you have not measured before go-live cannot be compared afterwards, and arguments about whether a change worked are impossible to settle.
Four figures are enough: unanswered call volume, out-of-hours volume, current no-show rate if you take appointments, and time from enquiry to first contact.
Record them for a month rather than a week, so a single unusual period does not distort the picture.
A Review Rhythm That Works
Weekly at first
Read transcripts for the first month. Patterns appear faster in the words than in the numbers.
Monthly thereafter
Review the metrics, the unanswered questions and the escalation reasons together.
Change one thing
Adjust a single element so the next month’s data tells you whether it helped.
Quarterly once stable
Reduce frequency when the flows settle, but do not stop entirely.
Questions People Ask
What is a good answer rate?
Better than yours currently is. We avoid publishing benchmarks because they vary enormously by sector and rarely describe any particular business.
Should escalation rate be low?
Not necessarily. Over-escalation costs a minute; under-escalation can cost considerably more. Start conservative.
Why avoid average handling time?
Because it improves when conversations are cut short, which produces callbacks and unhappy customers.
How do we define resolution?
Write the definition down and apply it consistently. An inconsistent definition makes the metric worthless.
Who should review the data?
Whoever owns the content, alongside someone who takes calls. The second perspective catches things the numbers miss.
Record Four Numbers Before You Change Anything
It takes an afternoon and makes every later conversation about performance a factual one.