Average speed of answer, or ASA, is the number callers feel first. It measures how long people wait in the queue before an agent picks up. A low ASA means callers reach a person quickly; a rising ASA is usually the first sign that a queue is understaffed, that handle times have grown, or that agents are not where the roster says they are. This guide explains how to calculate average speed of answer, why the average can hide bad waits, how to set a sensible target and what actually brings ASA down.
What average speed of answer measures
ASA is the average time an answered call spends waiting in the queue before an agent answers it. The clock usually starts when the call enters the queue, after the IVR menu, and stops when the agent picks up.
Two details matter when you compare ASA across systems or vendors:
- Only answered calls count. Callers who hang up while waiting are not in the ASA figure at all. That is why ASA must always be read next to the abandon rate.
- IVR time is usually outside it. If your IVR menu takes 50 seconds to get through, the caller has waited 50 seconds before ASA even starts. Keep menus short; see IVR call flow design.
For the wider KPI set and how a wallboard shows them together, see call centre KPIs and wallboards.
The ASA formula
ASA = total queue wait time of answered calls ÷ number of answered calls
A worked example. Your support queue answers 1,000 calls in a day. Adding up the wait time of every answered call gives 25,000 seconds.
ASA = 25,000 ÷ 1,000 = 25 seconds.
That looks healthy. Now look at how the same 25 seconds can be made up:
| Group of calls | Calls | Wait per call | Total wait |
|---|---|---|---|
| Answered quickly | 900 | 5 seconds | 4,500 seconds |
| Answered after a long wait | 100 | 205 seconds | 20,500 seconds |
| All answered calls | 1,000 | 25 seconds average | 25,000 seconds |
Nine callers in ten waited 5 seconds. One in ten waited well over three minutes. The average is still 25 seconds. This is the main weakness of ASA: it hides the long waits that generate complaints. Read it with the longest wait, the service level and the abandon rate, and look at it by hour, not only by day.
ASA, service level and abandon rate together
| Measure | What it tells you | What it misses |
|---|---|---|
| Average speed of answer | Typical wait for callers who got through | Long waits hidden in the average, and callers who gave up |
| Service level | Share of calls answered within your target time | How long the slow calls actually waited |
| Abandon rate | Share of callers who hung up before an agent answered | Whether the ones who stayed were answered fast |
| Longest wait | The worst experience in the queue right now | Everything else |
A queue with a good ASA and a high abandon rate is not healthy: the impatient callers left, so only the patient ones were measured. Call centre service level explains how to calculate the second measure and how to treat abandoned calls.
Setting an ASA target
There is no single correct ASA. The right target depends on what the caller is trying to do and what else they could do instead. Some common starting points, as rules of thumb rather than standards:
- Sales, bookings and admissions lines: keep ASA very short, often within 20 to 30 seconds, because a waiting buyer calls a competitor.
- Customer support: many teams aim for under a minute, with a strict limit on the longest wait.
- Emergencies, hospital codes and urgent helplines: answer almost immediately; these should not wait in a general queue at all.
- Back-office or scheduled callbacks: a longer ASA is acceptable if callers are told what to expect.
Set the ASA target together with a service level target and an abandon target, and agree all three with the business. A target that nobody is staffed to meet only produces red numbers.
Why ASA goes up
ASA rises for a small set of reasons. Check them in this order:
- Not enough agents logged in for the hour. Compare calls offered by hour with agents available by hour. Most ASA problems are roster problems; see call centre shift scheduling.
- Agents logged in but not available. Long breaks, paused status and long wrap-up after calls all remove capacity. The agent status and break codes guide covers how to track this fairly.
- Handle time has grown. A new product, a system change or a policy change can add a minute to every call. See average handle time.
- Repeat calls. Customers calling back about the same issue add volume. Better first call resolution cuts it.
- Peaks you did not plan for. Monday mornings, the first days of the month, festival seasons, results days and outages all spike demand.
How to bring ASA down
- Staff to the interval, not the day. Use the hourly or half-hourly pattern of calls and plan agents with Erlang C. A day that is fine on average can still have two terrible hours.
- Schedule breaks against the valleys. Lunch breaks at the busiest hour are a common, avoidable cause of long waits.
- Control wrap-up. Keep after-call forms short. In Voxdial Contact Center, auto closure returns an agent to the queue after a configurable time, 30 seconds by default.
- Add an overflow group. Let a second team or trained back-office staff take calls when the queue builds, with routing by agent priority.
- Offer a callback instead of a long wait. Missed and abandoned callers should land on a callback list that agents work through.
- Use AI as a backup, not a replacement. In a human-first setup, agents answer first; when every agent is busy or it is after hours, a BitVoice AI voice agent can take the call, handle simple requests and pass the caller back to an agent with a summary.
Watching ASA in Voxdial Contact Center
Voxdial Contact Center shows queue speed in two places. The supervisor dashboard shows average wait time, average response time, service level, abandoned ratio and occupancy, with a date picker to review any past day and an hour-by-hour chart of calls. The wallboard, built for a TV in the contact centre, shows average speed of answer next to calls in queue, the longest wait, the service level gauge, the abandon rate gauge, SLA breaches and callbacks pending, and it refreshes on its own.
Targets for the gauges, such as the service level threshold in seconds and the abandon rate target, are set once in Call Center Settings, so the colours on the wallboard match what your business has agreed.
Want to see live ASA, service level and abandon figures on your own queues? Talk to BitVoice for a demo.
FAQs
What is a good average speed of answer?
It depends on the line. Sales and booking lines usually aim for 20 to 30 seconds or less, and many support teams aim for under a minute. Set ASA together with a service level and abandon target rather than on its own.
Does ASA include abandoned calls?
Usually not. ASA averages the wait of answered calls only, so callers who hung up are missing from it. Always read ASA next to the abandon rate.
What is the difference between ASA and service level?
ASA is the average wait of answered calls. Service level is the percentage of calls answered within a set time, such as 80% within 20 seconds. Service level shows how many callers waited too long; ASA can hide them.
How do I reduce average speed of answer without hiring?
Staff by interval, move breaks away from peak hours, shorten wrap-up, add an overflow group and offer callbacks. An AI voice agent can also take overflow calls when every agent is busy.
