Average Handle Time (AHT)
Also called: AHT, handle time, average handling time
Average Handle Time is the average total time one call costs an agent, measured from the moment it connects to the moment the record is finished: talk time, plus any hold, plus the after-call work of writing the notes and booking the appointment. It is a capacity measure, not a quality measure.
What it means
Average handle time is the total time one call occupies an agent, averaged across calls. It starts when the call connects and it does not stop when the homeowner hangs up: it stops when the agent is genuinely ready for the next call.
That makes it a capacity number. It answers how many calls a given number of agents can absorb in a given number of hours, which is a real and useful question when staffing a storm week. It does not answer whether those calls went well.
Talk time plus after-call work
Handle time has two halves, and only the first one is visible to anyone listening.
The first half is the conversation: talk time, plus any hold while the agent looks something up, plus transfer time. The second half is the after-call work — writing the notes, correcting the spelling of the street, attaching the date of loss, setting the appointment in the calendar, flagging a bad number so nobody dials it again.
The second half is where handle time is usually mismeasured. A phone system reports talk time by default because talk time is what it can see, and a team that quotes its talk time as its handle time is quoting roughly half of the real figure. Worse, the after-call half is the half that does not disappear when it is left undone. It moves: into a supervisor cleaning records at the end of the shift, into the roofer arriving at an address that was typed wrong, into a second call to ask the homeowner something that was already said once.
So when two people compare handle times, the first question is never whether one is higher. It is whether both include the wrap.
Why pushing it down is usually the wrong instinct here
Average handle time came out of high-volume customer service, where a call is a cost and the aim is to resolve it and end it. An appointment-setting campaign inverts that. The call is not the cost; it is the product. A shorter call is only better if it produced the same appointment.
The reason this matters is that handle time is trivially easy to reduce, and every quick way of reducing it removes something load-bearing:
Skip the qualifying questions and the call gets shorter immediately. It also books inspections on rentals, on properties outside the storm footprint, and on roofs where a claim is already settled — journeys the crew makes and cannot sell.
Skip the read-back at the end and the call gets shorter again. The appointment is now one the homeowner half agreed to, at a time they did not repeat back, and the first anyone hears of it is when nobody answers the door.
Stop taking notes during the conversation and the talk time falls while the wrap grows, so the honest figure is unchanged and the record is worse.
Push hard enough and the number improves while the business behind it does not. That is the specific failure to watch for: an agent who is measured on handle time and nothing else has every incentive to end calls, and ending calls is not the job.
There is a legitimate version of the same work. Handle time that is long because the agent is hunting for the address of a property in a different system, or reading a script that asks the same thing twice, is waste, and removing it costs nothing. Fixing the tooling shortens calls without shortening the conversation.
What it should be read alongside
Handle time on its own cannot tell you whether it is too high or too low, because it has no direction. Read it against the outcomes it trades with.
The share of calls that produce a booked inspection says whether the extra minutes are buying anything. Handle time up and bookings up is qualification working; handle time down and bookings down is qualification being skipped.
The share of bookings that are actually kept says whether the appointments are real. This is the pair that catches the damage fastest, because cutting the confirmation shows up here and nowhere else.
Contact rate belongs in the same view when the campaign is outbound, since handle time is a per-conversation figure and reach decides how many conversations there are to average over.
Occupancy and queue wait belong there too, because those are the questions handle time is genuinely good at answering. If callers are waiting, handle time is part of the arithmetic that says whether the answer is more agents or shorter wrap.
Common mistakes
Quoting talk time and calling it handle time, which understates the cost of every call and makes two teams’ figures incomparable.
Setting a target for it on an appointment-setting campaign. A target on a capacity measure, in a job that is paid for outcomes, gets the capacity measure and nothing else.
Reading a fall as an improvement without checking what got shorter. Handle time going down is a question about which part of the call was removed, and the answer is occasionally efficiency and usually the qualifying or the confirming.
On the call
Your handle time is talk plus hold plus everything you do after the homeowner hangs up, so the honest way to shorten it is to capture the address, roof age, date of loss and claim status while they are still speaking rather than reconstructing all four from memory afterwards. Do not shorten it by cutting the read-back at the end. Confirming the day, the time, the address and who will be home is the part of the call that decides whether the appointment survives to the inspection, and it is the first thing an agent under time pressure drops.
This is how our agents handle it on storm appointment calls and day-to-day roofing answering service intake.
Questions people ask about average handle time
- What is actually included in average handle time?
- Three things, and the third is the one people forget: the talk time, any hold or transfer time, and the after-call work of writing the notes, tidying the record, and booking the appointment. A figure built from talk time alone understates what a call really costs, which is why two teams quoting handle times are often not measuring the same thing at all.
- Is a shorter average handle time better?
- Not on an appointment-setting campaign. Handle time is a capacity measure, so it tells you how many calls a team can take, not how many of those calls turned into inspections that happened. The work that lengthens a call is usually the qualifying and the confirming, and those are the two things a booked appointment depends on.
- How should average handle time be read then?
- Alongside the outcome rates it trades against, never on its own. Read it with the share of calls that book and the share of bookings that are kept, and treat a fall in handle time as a question rather than an achievement: if calls got shorter and bookings got fewer, the saving was taken out of the qualifying.
Back to the full roofing claims glossary.