Average Talk Time (ATT)
Definition
Average Talk Time (ATT)
Average Talk Time (ATT) is the average number of seconds a call center agent spends actively speaking with a caller, measured from the moment the two connect to the moment either party hangs up. It excludes hold time and wrap up work.
Talk time is the single largest slice of a call’s cost — it drives staffing math more than any other timer on the dashboard. Shave ten seconds off a 300-second average and you free about 3% of the talk capacity behind every contact.
ATT is often confused with Average Handle Time (AHT), which bundles talk, hold and after-call work together. ATT isolates the speaking portion alone, and that changes what a manager coaches.
ATT is a standard field in every call center reporting suite, so the number is easy to pull and easy to misread. Read it next to resolution and quality data or do not read it at all.
Key takeaways
- ATT measures only the live conversation portion of a call, not hold time and not wrap-up.
- Lower ATT lifts capacity and cuts cost per contact — but only when service quality holds steady.
- Benchmark bands vary by sector: retail sits near 3 to 4 minutes, technical support closer to 6 to 9.
- Voice analytics and agent-facing knowledge bases cut ATT faster than raw pressure on agents.
- Pair ATT with First Call Resolution to avoid the fast-but-repeat-callers trap.
How it works
ATT is captured by the phone system on every completed call, then averaged across an agent, team, queue or reporting window. The timer starts when the two parties connect and stops when one hangs up, ignoring hold and wrap.
The formula is a straight ratio:
ATT = Total talk seconds ÷ Total calls answered
Most cloud phone platforms, including Genesys, Amazon Connect, Five9 and NICE CXone, expose the raw talk-seconds field on every call detail record. Analysts slice it by skill, shift, campaign or tenure band.
Work the capacity math through and the appeal becomes obvious. Take 200 seats, 60 calls per agent per shift, and ten seconds saved on each one. That is 120,000 seconds, or roughly 33 agent-hours recovered in a single day.
Managers usually watch ATT alongside a handful of sibling timers. The table below shows what each one counts, which is where most reporting arguments start.
| Metric | What it measures | Includes hold? | Includes wrap? |
|---|---|---|---|
| Average Talk Time (ATT) | live conversation seconds | No | No |
| Average Handle Time (AHT) | full cost per contact | Yes | Yes |
| Average Hold Time | time the caller waits mid-call | Yes | No |
| After-Call Work (ACW) | wrap-up once the caller hangs up | No | Yes |
| Average Speed of Answer (ASA) | queue wait before the agent connects | No | No |
| Occupancy | share of paid time spent on contacts | Yes | Yes |
Because Average After-Call Work Time sits outside ATT, a team can post a flat talk time while total handle cost climbs. Check both, or the saving is imaginary.
A lower ATT is not automatically better — rushed agents create repeat callers, and repeat callers push total volume up even as individual talk times fall. Metric design matters here.
Poorly built key performance indicators reward gaming instead of service. An agent told to hit a talk-time target will find a way to hit it, usually by closing calls early.
Examples
ATT swings widely by call type — a retail order-status call can close in two minutes, while a technical escalation runs past ten. Five benchmarked slices show how far the number moves once intent and complexity change.
Retail and e-commerce (2024). UK research firm ContactBabel, publisher of the Contact Centre Decision-Makers’ Guide, tracks retail interactions in the 3 to 4 minute band once self-service absorbs order-status queries.
Technical support. Enterprise help desks such as Cisco’s Technical Assistance Center (TAC) routinely land between 6 and 9 minutes, because troubleshooting needs screen-share, log checks and back-and-forth diagnosis.
Insurance claims (2023 to 2024). Property claim intake at carriers like Progressive and Allstate averages 7 to 12 minutes, as the agent walks the caller through incident details, damage photos and next-step scheduling.
Offshore Business Process Outsourcing (BPO) seats in the Philippines. Manila-based providers running U.S. retail programs typically target 3 to 5 minutes on tier-1 voice lines, using tight scripting and knowledge-base search.
Utilities and telco billing. No public benchmark covers this slice cleanly, but most tier-1 billing queues land between 4 and 6 minutes once identity checks and payment arrangements are counted.
Read together, the bands make one point: comparing raw ATT across sectors tells you almost nothing. The comparison only works inside a single call type.
| Call type | Typical ATT band | Main driver |
|---|---|---|
| Retail order status | 3 to 4 minutes | self-service deflection |
| Manila tier-1 retail voice | 3 to 5 minutes | scripting and knowledge base |
| Enterprise technical support | 6 to 9 minutes | diagnosis and screen-share |
| Property insurance claims | 7 to 12 minutes | evidence capture and scheduling |
Workforce planners feed these bands straight into forecasts, so a wrong talk-time assumption shows up as an understaffed queue. Call center agents then absorb the gap through overtime.
Related terms
ATT sits inside a cluster of contact center measures, each covering a different slice of the same conversation. The cluster stops at the call boundary: it covers timers and outcomes per contact, not channel mix or campaign economics.
- Average Handle Time (AHT): the full talk, hold and wrap total behind every contact.
- First Call Resolution (FCR): whether the caller’s issue got fixed on the first contact.
- After-Call Work (ACW): the logging and coding time an agent spends once the caller hangs up.
- Customer Satisfaction (CSAT): the quality gauge that talk time alone cannot see.
- Key Performance Indicator (KPI): the broader class of measures ATT belongs to.
- Call Center Agents: the frontline staff whose talk time patterns feed the average.
Reading them together prevents the classic trap of optimizing one at the expense of the rest.
FAQ
What is a good average talk time in a call center?
Benchmarks vary by sector. Retail and e-commerce lines usually land between 3 and 4 minutes, while technical support and insurance claims often run 6 to 10 minutes. The good number is the one that holds service quality steady as it falls.
Is ATT the same as AHT?
No. Average Handle Time adds hold time and after-call work on top of talk time, while ATT isolates the live conversation alone. Confusing the two produces bad coaching targets and workforce forecasts that miss.
How can a contact center reduce average talk time?
The reliable levers are agent-facing knowledge bases, real-time voice analytics that surface next-best actions, and routing that puts the caller on the right skill first. Coaching against outlier calls beats coaching against the average.
Does lower ATT always improve customer experience?
Not automatically. Harvard Business Review’s 2010 study Stop Trying to Delight Your Customers found customer effort beats speed as a loyalty driver. Callbacks from rushed calls hurt the experience.
How often should managers review ATT?
Most contact centers review team-level ATT weekly and individual-agent ATT monthly, since coaching calls for enough volume behind the number to separate signal from noise.
Ready to benchmark your own contact center against Manila-based BPO partners? Compare vetted providers on the Outsource Accelerator hub.
Independent
