Average handle time explained: talk, hold and after-call work, the formula, when a low AHT is a bad sign, pairing it with FCR and quality, and SLA targets.
What is average handle time?
Average handle time, or AHT, is the mean amount of agent time consumed by one customer contact, including talk time, time the customer spends on hold and the after-call work the agent completes once the conversation ends. It is calculated by adding those three components across all handled contacts and dividing by the number of contacts handled.
Average handle time is the oldest and most widely used efficiency metric in contact centers, and also the most misused. It tells you how much agent capacity each contact consumes, which makes it indispensable for staffing and forecasting and for pricing in an outsourcing contract. It tells you nothing, on its own, about whether the contact was any good. Treated as a target that agents must hit, it reliably produces rushed calls, premature transfers and repeat contacts. Treated as a planning input and read alongside resolution and quality, it is one of the most useful numbers an operation has. This guide defines the metric, explains its components and formula, shows when a low AHT is a bad sign, describes how to read it with first call resolution and quality, and sets out how to place it in a service level agreement. The shorter metric page is at average handle time.
What are the components of handle time?
Handle time has three components, each of which can be measured and managed separately and each of which has different causes when it grows:
- Talk time. The time the agent and customer are connected and speaking, from answer to disconnect, excluding holds. It grows with issue complexity, customer verbosity, agent uncertainty, slow systems the agent is navigating while talking and scripts that require long disclosures.
- Hold time. The time the customer is placed on hold during the contact while the agent looks something up, consults a colleague or waits for a system. It grows with knowledge gaps, poor search tools, approval requirements and back office dependencies. Hold time is also the component customers resent most, and it is covered in reducing hold times.
- After-call work. Also called wrap-up or after-contact work: the time after disconnect that the agent spends updating the record, writing notes, completing forms, sending follow-up emails or submitting tickets before becoming available for the next contact. It grows with clumsy systems, duplicate data entry, free-text note requirements and unclear disposition codes.
In digital channels the same idea applies with different labels. For chat, handle time is the duration from assignment to close, complicated by the fact that an agent may handle several chats at once. For email, it is the active working time on the message rather than the elapsed time to reply. Programs that blend channels should report handle time per channel rather than as one blended figure, because the figures are not comparable.
How is average handle time calculated?
Average handle time is calculated by summing total talk time, total hold time and total after-call work for all contacts handled in the period, and then dividing that sum by the number of contacts handled in the same period. The result is usually expressed in minutes and seconds. Abandoned contacts, which no agent handled, are excluded from both the numerator and the denominator. Transfers need a decision: a transferred contact may be counted once with the combined time of both agents, or as two handled contacts, and the two choices give different averages.
Two refinements matter in practice. First, the mean is sensitive to a small number of very long contacts, so reporting the median or the distribution alongside the mean shows whether a rising average reflects a general slowdown or a handful of outliers. Second, the average for a queue blends contact types that have legitimately different durations; a password reset and a billing dispute should not be held to one figure. Reporting by contact type, or at least by queue, keeps the metric meaningful. Workforce planning uses the figure directly, as described in call center workforce management and in the call center staffing calculator.
When is a low average handle time a bad sign?
A low average handle time is a bad sign when it is achieved by ending contacts before the issue is resolved, transferring customers rather than helping them, skipping verification or disclosures, or recording less than the next agent will need, because each of those saves seconds on the current contact and creates minutes on a later one. Handle time is a measure of consumption, not of value, and it is easy to reduce the consumption by delivering less.
The warning signs are consistent. Handle time falls while first call resolution falls or repeat contacts rise. Transfer rates climb. Quality scores drop on resolution and closing items while remaining fine on greeting and tone. After-call work shrinks and the next agent complains that notes are empty. Customer survey verbatims mention being rushed or cut off. A sudden fall in one team's handle time after a target was introduced, with no change in tools or process, is almost always a behavior change rather than an efficiency gain. The opposite failure also exists: a handle time that is high because agents cannot find information, cannot act without approval or are fighting slow systems, in which case the fix is structural and the agent is the wrong place to apply pressure.
How should AHT be used with first call resolution and quality?
Average handle time should be read as one side of a triangle whose other sides are first call resolution and quality score, with the goal of lowering handle time only while the other two hold or improve; a change that lowers handle time at the expense of either is a cost shift, not a saving. Looking at the three together converts handle time from a target agents hit into a diagnostic managers use.
In practice this means several things. Reports show the three metrics side by side at agent, team and program level. Coaching on handle time is reserved for agents whose resolution and quality are strong and whose time is high, because they are the ones with genuine efficiency to gain, usually in system navigation or call control. Agents with low time and weak resolution are coached on thoroughness, not speed. Structural causes, such as hold time driven by a slow knowledge search or after-call work driven by duplicate data entry, are fixed at the system level rather than coached at all. The methods for raising resolution without raising time are in first call resolution, and the metrics set as a whole is laid out in customer service metrics.
Occupancy and schedule adherence complete the picture. Handle time measures how long each contact takes; occupancy measures how much of an agent's logged time is spent handling contacts; adherence measures whether agents are logged in when the schedule says they should be. All three feed the staffing model, and an operation that drives handle time down while occupancy is already very high is pushing agents toward burnout. The relationship is explained in call center occupancy.
How do you set average handle time in a service level agreement?
Average handle time is set in a service level agreement as a planning assumption or a reporting metric with a tolerance band rather than as a hard target with penalties, because a penalized handle time target gives the provider a direct incentive to cut contacts short. Where the contract is priced per hour or per agent, handle time determines how many agents the volume needs, so the parties agree an assumed figure per contact type from a measured baseline, review it at agreed intervals and adjust staffing or price when the real figure departs from the assumption for structural reasons such as a product change or a new disclosure.
Where the contract is priced per contact or per resolution, the provider carries the handle time risk and the client's interest shifts to protecting quality and resolution; in that model the agreement should put its teeth into first call resolution, quality score and customer satisfaction and treat handle time as a reported figure. In either model the agreement should define the formula, the treatment of transfers and abandons, the reporting granularity by contact type and channel, and who owns the levers that drive handle time, such as system performance, knowledge content and approval thresholds. Pricing structures are compared in call center pricing models, and the overall agreement structure in call center service level agreement.
A sensible negotiation sets the baseline during the first months of live operation, freezes the definition, and agrees a joint improvement plan that names the structural changes each side will make. That turns handle time into a shared engineering problem rather than a lever one side pulls on the other. If you want to see how handle time, resolution and quality would be defined and reported for your program, and how the staffing model would be built from them, request a proposal with your contact types, channels and current figures.
Average Handle Time Explained: Frequently Asked Questions
What does average handle time include?
Average handle time includes talk time, the time the customer spends on hold during the contact and the after-call work the agent performs after disconnecting, such as updating records and writing notes. It excludes time in queue before the agent answers and excludes contacts that were abandoned before an agent handled them.
How is average handle time calculated?
Add total talk time, total hold time and total after-call work for all contacts handled in a period, then divide by the number of contacts handled in that period. Report it by contact type and channel rather than as one blended figure, and look at the median or distribution alongside the mean to spot outliers.
Is a low average handle time always good?
No. A low handle time achieved by ending contacts early, transferring instead of resolving, skipping verification or leaving thin notes shifts work to a later contact and lowers first call resolution. Handle time should fall only while resolution and quality scores hold or improve; otherwise it is a cost shift rather than an efficiency gain.
How do average handle time and first call resolution relate?
They pull in opposite directions when managed carelessly: pressure on handle time tends to lower resolution, and effort on resolution tends to raise handle time. Reading them together, with quality score as the third measure, lets managers coach efficiency only where resolution is already strong and fix structural causes such as slow systems and approval limits.
Should average handle time be a penalty metric in an outsourcing contract?
Generally no. A penalized handle time target gives the provider an incentive to cut contacts short. Treat handle time as a planning assumption for staffing and price, reviewed at intervals against a measured baseline, and put contractual teeth into first call resolution, quality and customer satisfaction, which protect the outcome the handle time is supposed to serve.
