Last updated: September 12, 2026
What call center occupancy rate measures
Occupancy rate is the one number on a contact center dashboard where a rise can be bad news. It measures the share of an agent's available time spent handling calls: talk plus hold plus after call work, divided by that total plus the idle seconds waiting for the next call. Every other metric here rewards you for pushing it.
Drive occupancy into the nineties and the arithmetic reads like a win, since every paid minute converts into a handled call. What stops being described is the person in the seat, who hangs up and hears the next call land before the last message is finished.
The cost surfaces on the twelfth call of an hour, where an operator's voice has flattened and nobody asks the second question, and again in a resignation a month later. Occupancy reads 94% through all of it, so keep it beside the QA metrics that describe your program rather than on a target sheet.
The formula, and what counts as available
Occupancy is handling time divided by available time over the same interval.
Handling time is talk plus hold plus after call work, the same three pieces that make up average handle time. Available time is that total plus the seconds an agent sat logged in waiting for a call that had not arrived. Time spent off the phone belongs to neither half, which is the part most people get wrong.
Say an agent is paid for an eight hour shift, or 480 minutes. Ninety of those go to two breaks, a team meeting, and a coaching session, leaving 390 available. Talk comes to 250 minutes, hold to 18, after call work to 44, so handling time is 312. Occupancy is 312 over 390, or 80%, which the arithmetic produced rather than anybody setting it.
| Time | Where it goes | Why it lands there |
|---|---|---|
| Talk and hold | Handling | The call is live, whatever the caller hears |
| After call work | Handling | Message typing, dispatch to the on-call tech, disposition coding |
| Waiting for the next call | Available, outside handling | The seconds that keep occupancy under 100 |
| Break, lunch, meeting, coaching | Outside the fraction | Paid time that was never available, which is where shrinkage counts it |
Two floors can run the same shift and publish occupancy nine points apart on definitions alone, so check how your platform counts blended outbound work and a hold parked on a second leg.
Occupancy and utilization are two fractions wearing one name
Those two get swapped constantly, and the confusion is inherited rather than careless, since workforce management vendors define them differently.
The split that holds up: occupancy measures against the time somebody was available to take calls, and utilization measures against the time they were paid. Same numerator, different denominator.
Divide the same 312 handling minutes by the 480 paid minutes and utilization comes out at 65%. One agent, one shift, two numbers fifteen points apart, both correct. Which one gets quoted depends on what the person quoting it wants, since 80% describes a floor running hot and 65% describes a floor with room in it.
Shrinkage is the other side of the same shift
Shrinkage is the share of paid time that was never available for calls, and it picks up holidays, sick days, and training once you run it across a year instead of a day. On the shift above, 90 minutes out of 480 is 18.75%.
The three fit together cleanly. Occupancy describes what happened inside the available time, shrinkage describes how much available time there was, and multiplying occupancy by what shrinkage left behind gives utilization: 80% against 18.75% produces the 65% above.
The benchmark table on this site carries a shrinkage row with an empty figure cell, since every published version we chased led to another page quoting a page. Your own trailing quarter is the one worth planning against.
Occupancy falls out of the schedule instead of setting it
Occupancy is not a dial you turn. It arrives once you decide how many people answer a known volume at a known handle time, and the answer speed you promised decides most of it.
Staff to answer faster and occupancy drops, because more agents are waiting on the same calls. Cut the headcount and it climbs while the queue lengthens and abandons follow. You pick your service level or you pick your occupancy, and whichever one you leave alone gets decided for you, which is usually the service level a client agreement holds you to.
Queue size moves it too. Six operators cannot hold the same occupancy as sixty at the same answer speed, because a small queue has nowhere to absorb two calls arriving in the same ten seconds, which is why a six seat board benchmarked against a large center ends up understaffed.
What a good occupancy rate is
Published targets vary by several points depending on whose page you land on, and none met the evidence bar when we checked on 2026-09-12, which asks for a named publisher, a date, and a stated population. No figure gets quoted here for that reason.
What decides yours is the answer speed you committed to, your queue size, how long your calls run, and what the work costs the person taking it. Build it from your own trailing quarter, per queue.
Why chasing occupancy upward backfires
The case for pushing it is real at the bottom of the range, where a floor sitting at 55% pays for a lot of waiting and tightening the schedule buys genuine capacity.
Further up, the same move buys something else. The gap between calls is where an agent finishes the last message and resets their voice for a stranger, and squeezing it moves the work rather than removing it. After call work gets typed during the next call's ring, and the message that needed 40 seconds goes out in 20 with a field missing.
We are not going to hand you the threshold where that starts, since no traceable source gives one and it depends on the call anyway. An operator taking 45 second messages recovers faster than somebody who just came off an eleven minute Scope of Appointment. Your own floor answers it, in the month where attrition and quality scores start moving together.
Occupancy and the parts of a call that fall off first
A hot floor does not make anybody worse at the job, and the scores fall anyway, because the parts of a call that take time and are easy to skip come off first and those parts are almost always scored.
Verification goes early, since it costs ten seconds and the caller rarely objects to losing it. The recap follows, then the second probing question, and a callback number read back digit by digit becomes a number said once at speed. Each of those is a weighted line on an answering service scorecard, so the score drops while the standard and the people stayed where they were.
That makes a falling quality trend readable as a staffing symptom, and ruling it out is cheaper than building a coaching plan on top of it. Three checks separate the two.
Pull occupancy by interval for the weeks in question rather than as a monthly average, since a month at 78% hides a run of afternoons at 94%.
Score calls from the hottest intervals against calls from the quietest ones, same agents, same accounts, same scorecard. A coaching problem follows the agent into both, and a staffing problem stays with the interval.
Read which criteria lost the points. An even drop across the scorecard usually means nobody understood the standard, while points coming off capture, close, and probing with greeting and script adherence untouched is the signature of a floor working faster than the calls allow.
Rule based criteria read cleanly there, since whether a number got read back is a yes or a no, and anything resting on how rushed somebody sounded holds up less well, which Aircall and StackAI both note. When the intervals explain the drop, what happens in a coaching session waits until the afternoons look different.
A call review scores a batch of recordings from your own busiest week against your existing scorecard, which puts the quality side of that week next to the occupancy that produced it.
Frequently asked questions
What is the occupancy rate formula?
Divide handling time by available time over the same interval. Handling time is talk plus hold plus after call work, and available time is that total plus the seconds an agent waited for a call. Time off the phone stays out of both.
What is a good occupancy rate in a call center?
Published targets vary and none traced back to a named publisher with a date and a population when we checked on 2026-09-12, so this page quotes none. Your answer speed commitment, queue size, call length, and how hard the calls are decide the right number for your floor.
Does occupancy include after call work?
Yes, on both sides of the fraction. Wrap, disposition coding, message typing, and dispatch are handling time, and dropping them understates occupancy while flattering your capacity math.
Sources
- Aircall and StackAI on rule based criteria scoring more reliably than tone. Third party, unverified.
- No published occupancy rate target met the evidence bar on 2026-09-12, so none is quoted here.
- Shrinkage has no figure traceable to a named publisher with a stated population, checked 2026-09-12.
- Trinsic feature sheet and call notes, 2026-08-27 and 2026-09-03. Verified internally.
- The 480 minute example, the definition table, the interval comparison, and the order in which criteria fall off a rushed call. Editorial judgment.