Average Talk Time

Average talk time (ATT) is the mean number of seconds an agent spends actually speaking with a caller across a set of calls. It counts conversation only, so hold time, transfer time and after call work all sit outside the figure.

That narrow scope is the whole point. ATT isolates the part of a call an agent controls with their own voice, which is why supervisors read it per agent, per queue and per shift instead of as one site-wide number.

The figure feeds average handle time (AHT), staffing forecasts and coaching plans. In a call center, it’s the first number a team lead checks when a queue slips, because it moves faster than customer satisfaction (CSAT) scores.

Low ATT isn’t automatically good. A rushed 90-second call that ends in a callback costs more than a patient four-minute call that resolves the issue on the first try.

Key takeaways

• ATT counts agent and caller conversation only, dead air included, but never hold or wrap-up time.
• Healthy ATT depends on channel, industry and call type, so there’s no single benchmark to chase.
• Trends beat snapshots. A week-on-week climb says far more than any single day’s figure.
• Sudden spikes usually point at a script change, a slow system or a training gap, not lazy agents.
• Pair ATT with first-call resolution so speed never gets rewarded on its own.

How it works

ATT is total talk time divided by the number of calls handled in a period. Platforms log it automatically at agent, team and queue level, and they exclude hold, transfer and after-call work, each of which runs on its own separate clock.

The clock starts the moment agent and caller are connected and stops at disconnect. Silence counts. Dead air while an agent hunts through a knowledge base sits inside ATT, which is why a long figure often points at tooling rather than talkativeness.

Hold and transfer waits break the timer — the caller isn’t speaking to that agent. Wrap-up starts after the caller drops. Stack all of it together and you get AHT, the number tied to service level agreement targets.

Benchmarks vary widely by call type. Tech support calls often average 6–8 minutes, simple billing questions in a back office queue land closer to 2 minutes, and sales calls sit somewhere in between.

Sampling has changed too. Speech analytics tools now score 100% of calls in most enterprise deployments, up from the old 2–5% manual rate — so supervisors can see whether a drift comes from dead air, preamble or complex issues.

Examples

ATT earns its keep when you read it as a trend against one queue’s own history. The cases below show it working as a diagnostic in four different ways: a script defect, a coaching signal, a forecasting input and a side effect of automation.

A US healthcare payer running member services through a Philippines-based BPO tracked ATT weekly across 400 agents. It crept from 4:20 to 5:15 in a single month, an extra 55 seconds on every call.

The vendor traced the climb to redundant identity checks in a new prior-authorization script and cut 40 seconds by removing them. ATT dropped back into range with no headcount change at all.

A fintech in Financial Services reads ATT next to customer experience scores. Agents below the queue median whose CSAT stays high get flagged as top performers, not rushers.

A Healthcare BPO on the Top 40 BPO companies in the Philippines list plans headcount six months out on ATT trends. Rising ATT against flat volume means refresher training, not more seats.

A Software Development desk pairs ATT with chatbot deflection rates. When the bot absorbs tier-one questions, ATT rises because only complex tickets reach a human — a healthy sign, not a red flag.

Precedence Research puts the global business process outsourcing (BPO) market at USD 347.95 billion in 2025, projected to grow 10.05% annually through 2035.

The Information Technology and Business Process Association of the Philippines (IBPAP), which publishes the sector’s IT-BPM industry roadmap to 2028, reports Philippine IT-BPM employment at 1.9 million and targets 2.5 million by 2028.

IBISWorld’s telemarketing and call centers report sizes the US industry at USD 30.9 billion across 46,650 businesses in 2026.

Directories such as Clutch let buyers filter vendors by track record and industry specialty, though no directory publishes ATT itself. You get that number from the vendor’s own reporting during a pilot.

Related terms

These terms sit closest to ATT in a contact center’s reporting stack. Some name the operation the metric is measured inside, others name the metrics and service lines it gets read against, and each carries its own glossary entry.

• Call Center: the operational hub where inbound and outbound voice calls are handled by agents.
• Service Level Agreement: the contracted response, resolution and quality thresholds a vendor commits to.
• Customer Service: the wider function ATT feeds into, spanning voice, chat and email channels.
• Outsourcing: the practice of moving business functions like call handling to external providers.
• Omnichannel: a support model where ATT sits alongside chat and email response metrics.
• Virtual Assistant: a remote worker whose voice and admin tasks get measured by similar timing metrics.
• Knowledge Process Outsourcing: a higher-skill outsourcing tier where call complexity and talk time typically run longer.

FAQ

What’s a good average talk time?

There’s no universal number. Tech support often runs 6–8 minutes, billing queries land near 2 minutes and sales calls sit between them, so measure against your own queue’s baseline rather than a generic figure.

How is ATT different from AHT?

ATT counts only the agent and caller conversation, while AHT adds hold time, transfer waits and after-call work on top. AHT is the fuller cost picture; ATT isolates the talk portion a coach can act on directly.

Does lower ATT mean better performance?

Not on its own — a short call that gets escalated or triggers a callback costs more than a slightly longer one resolved first time. Pair ATT with first-call resolution and CSAT before you reward speed.

How does location affect ATT?

Whether you use offshoring, nearshoring or onshoring, ATT tracks training, tooling and script quality far more closely than geography. Language proficiency and accent familiarity shave or add seconds per call, but they rarely explain a minute-wide gap.

Can AI cut average talk time?

Real-time agent assist, auto-summaries and knowledge-base retrieval shave 30–60 seconds off many calls by clearing the dead air inside them. The Ultimate Guide to Outsourcing covers how BPOs stack these tools alongside human agents.

Should finance and accounting teams track ATT?

Yes: a Finance and Accounting BPO handling accounts payable or collections calls uses ATT to size teams and spot script friction, exactly like any other voice queue.

Track your own average talk time trends and compare vendors on quality, not just cost. Explore BPO providers in the OA directory to shortlist call center partners with proven ATT discipline.