Average Talk Time (ATT)
Average Talk Time (ATT) is the average length of time an agent spends actually talking to the customer on a call — the live, voice-to-voice portion of the interaction.
It deliberately excludes hold time and after-call work, which makes it a precise, narrow measure rather than a catch-all.
Because it isolates the conversation itself, talk time is one of the most useful diagnostic signals in a contact centre.
It's also one of the most misused — too often turned into a target agents are pushed to "beat", which is exactly where it goes wrong.
This guide explains what ATT is, how it sits alongside Average Handling Time (AHT) and After-Call Work (ACW), how to calculate it, and why the right way to use it is as a diagnostic — never a stopwatch.
What it is
The average time an agent spends speaking with and listening to the customer on a call — excluding hold time and after-call work.
Why it matters
It isolates the conversation itself, so it surfaces coaching opportunities, process problems and training gaps that AHT alone hides.
What this guide covers
What ATT is, how it differs from AHT and ACW, how to calculate it, common pitfalls, and why it belongs on a dashboard — not on a target sheet.
What is Average Talk Time?
In plain English
Average Talk Time is the average number of seconds or minutes an agent spends in live conversation with the customer — talking and listening — across a set of calls.
It starts when the agent and customer are connected and stops when they're not actively speaking (a hold) or when the call ends.
It says nothing about the time the customer waited on hold, or the wrap-up the agent does afterwards.
Most telephony and contact centre platforms capture talk time automatically as a standard call-event timestamp, which is why it's one of the cleanest, most reliable numbers you have.
The catch is that "clean" doesn't mean "complete" — talk time measures one slice of the interaction, and reading too much into that slice is where teams get into trouble.
What it is
The live, voice-to-voice portion of a call — the conversation itself, averaged across calls. A precise diagnostic of how long agents spend talking to customers.
What it isn't
It is not the same as handle time. It excludes hold and after-call work, and it is not a productivity score or an efficiency target to be driven down.
Why It Matters
Talk time matters because it isolates the part of the call you can actually coach: the conversation.
When you separate it from hold and wrap-up, you can see whether a long handle time is being driven by the agent's conversation, by clunky systems, or by a slow back-office process.
🎧 For contact centre leaders
ATT is a window into call flow and process friction. A talk time that drifts up across a team usually points to a process or systems problem, not lazy agents — and that's a far more useful place to look.
🧭 For team leaders & coaches
Comparing an agent's talk time against the team — alongside quality and resolution — highlights who may need conversation-control coaching, and who is simply handling more complex calls.
📊 For WFM & planning
Talk time feeds the workload that drives staffing models. Understanding what's inside it (versus hold and ACW) makes forecasts and Erlang-based staffing far more accurate.
ATT, AHT and ACW: How They Fit Together
Talk time is one ingredient in a bigger measure. The clearest way to understand it is to see exactly where it sits in the anatomy of a call.
Talk Time
The live conversation — agent and customer speaking and listening. This is what Average Talk Time measures.
Hold Time
Time the customer spends on hold during the call while the agent checks something. Excluded from talk time.
After-Call Work (ACW)
Wrap-up the agent does after the customer has gone — logging notes, updating records. A separate metric in its own right.
Average Handling Time (AHT)
The sum of all three — talk + hold + ACW — averaged across calls. The composite metric that contains talk time.
The relationship in one line
AHT = Talk Time + Hold Time + After-Call Work. Talk time is the conversation only; AHT is the whole interaction.
Track them separately and you can see which part is moving and why.
Collapse them into one number and you lose that diagnostic power. For the full picture, see AHT and ACW.
How to Calculate Average Talk Time
The calculation is simple — the discipline is in being clear about what you're including and what you're not.
The formula
Average Talk Time = Total Talk Time ÷ Number of Calls Handled.
For example, if your agents logged 50 hours (180,000 seconds) of talk time across 1,200 calls in a day, your ATT is 180,000 ÷ 1,200 = 150 seconds, or 2 minutes 30 seconds per call.
A few things to be deliberate about:
Exclude hold and wrap
If your platform bundles hold into "talk", you're not measuring talk time — you're measuring something muddier. Confirm what your system reports before you trust the number.
Segment before you average
An all-in average across sales, billing and complex technical calls is close to meaningless. Calculate ATT by call type or queue so you're comparing like with like.
Common Pitfalls
Almost every problem with talk time comes from the same root cause: treating a diagnostic as a target.
Setting a talk-time target
The moment "keep talk time under X" becomes a goal, agents rush customers, talk over them and end calls prematurely. Quality and first-contact resolution fall — and the cost simply moves to call-backs.
Chasing a generic benchmark
There is no universal "good" talk time. A 90-second number that's great for a password reset is a red flag for complex financial advice.
Comparing your ATT to an industry average you found online tells you almost nothing.
⚠️ Talk time is a diagnostic, not a target
This is ACXPA's position: pushing agents to cut talk time is one of the most damaging things a contact centre can do. Shorter calls are not better calls.
Use talk time to investigate why conversations run long — poor systems, unclear processes, repeated questions — and fix the cause. Never reward agents for getting customers off the phone faster.
💡 Pair it with quality and resolution
Talk time only means something next to outcomes. Read it alongside first-contact resolution, quality scores and customer feedback.
A rising talk time with rising resolution may be a good thing; a falling talk time with falling resolution almost never is.
Reading productivity against quality is exactly what ACXPA's Coaching & Performance Hub is built for. The Agent Performance Matrix plots each agent's productivity score against their quality score on a four-quadrant view, so a single number like talk time is never read in isolation. Its Simple tier is included with ACXPA Business Membership; the per-call competency assessment is a separate ACXPA Certified licence.
Frequently Asked Questions About Average Talk Time
What is average talk time?
Average Talk Time (ATT) is the average length of time an agent spends in live conversation with the customer on a call — talking and listening. It excludes hold time and after-call work, so it measures only the conversation itself, averaged across a set of calls.
How is average talk time different from AHT?
Average Handling Time (AHT) is a composite metric: talk time plus hold time plus after-call work, averaged across calls. Average Talk Time is just the talk component on its own. In other words, ATT is one ingredient inside AHT, which is why tracking them separately tells you which part of the call is driving your handle time.
Does talk time include hold time?
No. By definition talk time excludes hold time — the period when the customer is waiting while the agent checks something. Be careful, though: some platforms bundle hold into "talk" in their reporting, so confirm exactly what your system measures before relying on the number.
What is a good average talk time?
There's no universal answer, and that's deliberate. A "good" talk time depends entirely on call complexity: a simple password reset and a complex insurance claim have completely different healthy ranges. The right benchmark is your own data, segmented by call type, read alongside quality and resolution — not a generic figure from elsewhere.
Should I set a target for average talk time?
No — and this is ACXPA's clear position. Talk time is a diagnostic, not a target. The moment you set a talk-time goal, agents start rushing and ending calls early to hit it, which damages quality and first-contact resolution. Use talk time to investigate why calls run long and fix the underlying cause; never make a shorter call the goal in itself.
Why is my team's talk time increasing?
A rising talk time across a whole team is usually a process or systems signal, not an agent-behaviour problem. Common causes include slow or clunky systems, unclear processes that force agents to dig for answers, new products agents aren't yet confident on, or a shift in the mix of calls toward more complex enquiries. Investigate the cause before assuming it's a coaching issue.
How do I reduce average talk time the right way?
By removing friction, not by pressuring agents. Streamline systems so information is faster to find, simplify processes, improve knowledge bases and training, and fix the root causes that make conversations longer than they need to be. Done this way, talk time falls as a by-product of a better experience — for both the customer and the agent.
Where to Next
Summary: Average Talk Time
Average Talk Time is the average time an agent spends in live conversation with the customer — the talk component only, excluding hold and after-call work.
It sits inside Average Handling Time (AHT = talk + hold + ACW), and tracking it separately lets you see which part of a call is driving your handle time and why.
Used well, talk time is one of the most useful diagnostics you have: it points you to process friction, systems problems and coaching opportunities.
Used badly — as a target to drive down — it does real damage, pushing agents to rush customers at the cost of quality and first-contact resolution.
ACXPA's position is simple: talk time is a diagnostic, not a target. Read it alongside quality and resolution, segment it by call type, and never reward agents for getting customers off the phone faster.
Fix the causes of long calls, and a healthier talk time follows on its own.
