Key takeaways
- Handle time is talk, hold and after-call work together; wrap-up such as updating the customer's record is part of the call[2].
- Report the median and the slow tail for each call type, not one average across emergencies, status checks and price shoppers.
- Short is bad when it skips the readback, the urgency question or the fee expectation; those minutes prevent wrong addresses and wasted trips.
- For emergencies, borrow the 9-1-1 idea of a dispatchable point: get the details the on-call tech needs first, then finish the call[1].
- Minutes drive staffing: in the Erlang C Example, three people answering 20 calls an hour fall from about 84% to 21% answered within 20 seconds as handle time goes from four to eight minutes.
A nine-minute call can be the best call of the day
A nine-minute booking call can be the best call your office takes all day. The CSR confirmed the address and gate code, learned it was the second backup this month, asked where the cleanout is, set the arrival window, explained the diagnostic fee and read it all back. The tech arrived with the right machine and found the root intrusion on the first visit. A three-minute call on the same job would have sent the wrong truck.
Handle time still matters, because it decides how many people you need on the phones and how long the next caller waits. The goal is not short calls. It is no wasted minutes inside calls that do everything a booking needs.
Measure it the way call centers do: talk time, plus any hold, plus the after-call work before the CSR is ready for the next caller. Research on call-center operations treats that wrap-up, updating the customer's history or processing what the caller asked for, as part of the service[2]. Many phone systems report talk time only, so check what yours calls handle time before you compare anything.
Average handle time = (talk time + hold time + after-call work) รท calls handled. Work it out per call type and per person, over at least two normal weeks.
Anatomy of a plumbing booking call
| Segment | What it covers | What skipping it costs | Example time |
|---|---|---|---|
| Greeting and reason | Company name, the CSR's name, why they are calling | A caller unsure they reached the right company | 10โ15 s |
| Urgency check | Active water, sewage, no water, gas smell, who is affected | An emergency booked as routine, or a night callout for a drip | 20โ40 s |
| Address and access | Street, ZIP, unit, gate or lockbox, parking, pets | A truck at the wrong house or outside the service area | 45โ60 s |
| Problem detail | Fixture, symptoms, age of equipment, what they already tried | The wrong tech or parts, and a second trip | 60โ120 s |
| Window and fee | Arrival window, diagnostic or trip fee, who will be home | Cancellations at the door and price surprises | 45โ90 s |
| Readback | Name, number, address, problem and window repeated back | Typos that surface when the tech is lost | 20โ30 s |
| After-call work | Notes, lead source, tags, telling dispatch | Jobs that reach the board late or incomplete | 30โ90 s |
Add up the Example ranges and a complete new-customer booking lands somewhere between about four and seven and a half minutes, before price objections or a caller who needs walking through a shutoff. Existing customers with a record on file should run shorter, because address and access become confirmations rather than questions.
- Caller
- Your team
- System
01Caller
Caller explains the problem
Kitchen sink backs up when the dishwasher runs, second time this month.
02Your team
Urgency check
No active flooding, no sewage on the floor: routine, a same-week slot is fine.
03Your team
Address and access
Street, ZIP, unit, gate code and parking, typed while the caller talks.
04Your team
Window and fee
Thursday 8โ11 a.m., diagnostic fee explained, someone home confirmed.
05Your team
Readback
Name, number, address, problem and window repeated; caller corrects the unit.
06System
Job on the board
Notes, source and tags saved; dispatch can see it before the next call.
Example. The readback caught a wrong unit number, the cheapest fix there is.
Why one average misleads
One average across every call type hides most of what is useful. A shop that books emergencies, answers status checks and fields price shoppers will see call lengths from under a minute to well over ten. A few long calls pull the average up, a run of quick status calls pulls it down, and the same number can describe very different weeks.
- Report the median for each call type, not only the mean.
- Report the slow tail too, such as the 90th percentile, because training gaps and system problems show up there first.
- Split by call type: new booking, emergency, existing-customer status, reschedule, price question, commercial.
- Split by person and by hour. Handle time often climbs late in a long shift and during the morning rush.
Example: in a week of 40 new bookings, 32 take between four and seven minutes, six take 10 to 15 minutes (two commercial accounts, a caller who needed a shutoff walk-through, three price conversations), and two take under two minutes because the caller dropped mid-intake. The mean comes out at about 6.3 minutes and the median a little over five. Coach on the median, investigate the two short calls, and staff on the mean, because the queue feels every minute of the long ones.
Variability matters for staffing as well as coaching. In queueing models, waits get worse as calls get longer and also as their lengths become more variable, and they improve as more people answer in parallel[2]. Trimming the slow tail can do more for the queue than shaving seconds off typical calls.
When a short call is a bad call
Short calls look efficient on a dashboard. Before praising a falling average, check for these patterns:
- No readback. Unit numbers and callback digits get transposed, and the tech finds out at the curb.
- No urgency question. A slow leak behind a wall gets a Thursday slot, or a dripping tap gets a night callout.
- No fee expectation. The caller hears about the diagnostic fee from the tech and cancels at the door. See explaining the service call fee.
- No booking at all. The CSR takes a message and promises a callback, which shortens this call and creates two more. Only 19% of US adults say they generally answer calls from unknown numbers[3], so that callback often goes nowhere.
- Rushed intake on busy mornings. Handle time drops exactly when mistakes cost the most.
Check short calls against outcomes: wrong-address trips, door cancellations, second visits and calls back to fix a booking. The plumbing call intake checklist lists the questions that should survive any push for speed, and CSR booking rate is the number to watch next to handle time.
Borrow the dispatchable idea for emergencies
Emergency calls need their own target, and public safety has a useful one. NENA's 9-1-1 call processing standard says calls involving an imminent threat to life, serious injury or significant risk to property should be processed and dispatchable within 60 seconds, and other emergencies within 90 seconds. A call is dispatchable once the essential details, such as location, incident type and urgency, are known[1].
A plumbing office can borrow the structure without pretending a burst pipe is a 9-1-1 call. Define the dispatchable point for your emergencies: address and ZIP, what is happening, whether water or sewage is still running, and a callback number. Get there first, start the on-call path, then finish the rest, including walking the caller through a water shutoff. The whole call may run long. The minute that matters is the one in which the tech can be called.
Example target: emergency calls dispatchable within 90 seconds of answer; total handle time is whatever it takes to protect the home and finish the record.
When a long call is a problem
Long calls deserve attention when the extra minutes are spent on the office rather than the caller:
- Hunting through screens for the customer, the schedule or the price book.
- Typing notes twice, once in the phone system and again in the field software.
- Putting callers on hold to ask the dispatcher about a window.
- Answering the same policy questions from scratch because nothing is written down.
- Transferring between people who each ask for the address again.
Each of these is an office fix, not a talk with the CSR. A screen that opens the customer record on the first ring (recognizing repeat customers the moment they call), a written window policy and one place for notes usually take more time out of calls than any script.
Handle time in the staffing math
Handle time is half of every staffing calculation. Calls per hour times handle time gives the work arriving, and the Erlang C model turns that into waits for a given number of people[2]. The model is sensitive to load, so small changes in handle time move the answer a lot when people are busy. The full method is in how many CSRs a plumbing company needs.
Source: Manufacturing & Service Operations Management (Gans, Koole and Mandelbaum), 2003[2]
Show the data
| Average handle time | Answered within 20 seconds |
|---|---|
| 4 min | 84% |
| 5 min | 73% |
| 6 min | 58% |
| 7 min | 41% |
| 8 min | 21% |
In the Example, three people answering 20 calls an hour hold about 84% within 20 seconds at a four-minute handle time, 58% at six minutes and only 21% at eight (modelled). Same desk, same calls; the difference is minutes inside each call. That is why handle time creeping up after a new service line or a new software screen shows up as voicemail before anyone notices the calls got longer.
The minutes cost money too. Office and administrative support jobs made up about 12% of all jobs at plumbing, heating and AC contractors in May 2025[4], and much of that time is spent on the phone. Example: at 50 calls a day, each minute of average handle time is about 18 hours of phone work a month (50 calls ร 21.7 working days รท 60). At the June 2026 loaded cost of $36.73 per hour worked for private-industry office and administrative support workers[5], that is roughly $660 a month per minute. The wage alone understates it: customer service representatives at plumbing, heating and AC contractors earned a median $22.22 an hour in May 2025, before benefits and payroll taxes[6].
Targets by call type
| Call type | Must happen before hanging up | Example target | Red flag |
|---|---|---|---|
| New residential booking | Urgency, address and access, problem, window, fee, readback | 5โ7 min including notes | Under 3 min with no readback |
| Emergency | Dispatchable details first, on-call path started, shutoff guidance | Dispatchable within 90 s | A long hold before the address is taken |
| Existing customer, new problem | Record confirmed, problem, window, readback | 3โ5 min | Address typed from scratch |
| Status or arrival check | Answer from the board and set the next update | Under 2 min | A hold to ask dispatch every time |
| Price question | Fee explained, what the visit includes, offer to book | 3โ5 min | Ends with no booking offer |
| Commercial service call | PO or work order, site contact, access hours | 5โ8 min | Missing PO found at invoicing |
Commercial calls are where long handle times are most often justified: purchase orders, site contacts, access windows and not-to-exceed limits all have to be captured before the truck rolls. The commercial contractor example shows how much of that intake can be standardized.
Coaching without rushing callers
Pull the slow tail
Each week, list the longest tenth of calls for each type and listen to five of them.
Name the wasted minute
Was it searching, retyping, a hold for dispatch, or a caller who genuinely needed the time?
Fix the office first
Screens, written policies and one place for notes come before any script change.
Check the fast calls too
Listen to five of the quickest bookings and score the urgency, fee and readback steps.
Track outcomes, not just minutes
Wrong-address trips, door cancellations and booking rate tell you whether a shorter call is a better call.
Recordings make this routine; using call recordings to coach CSRs covers how. For after-hours and overflow calls, an SI Employee runs the same intake on calls your office forwards, reads the details back and emails a summary with the recording and transcript, so the after-call work is already done when someone opens it (see how forwarded calls are handled).
Frequently asked questions
What is a good average handle time for a plumbing booking call?
There is no published plumbing benchmark worth trusting. A complete new residential booking with an urgency check, address, problem, window, fee and readback takes several minutes. Set your target from two weeks of your own calls, by call type, and judge it against outcomes such as wrong-address trips and door cancellations.
Should after-call work count in handle time?
Yes. The CSR is not free for the next caller until the notes are done, and call-center research treats that wrap-up as part of the service[2]. Leaving it out understates the staff you need.
Does a script make calls longer?
A good one often makes them shorter, because the CSR stops improvising the order of questions and the readback catches errors in seconds. Scripts that bolt marketing lines onto every call do make calls longer; cut those first.
How do I measure handle time if my phone system only shows talk time?
Time a sample by hand for a week, from answer to the moment the CSR is ready for the next caller. Many field service tools also log when a job record is created and saved, which gives you the after-call portion.
Will a lower handle time raise our booking rate?
Not by itself. Booking rate rises when callers get a clear window and price expectation, and that takes time. A lower handle time is a staffing gain, so protect the booking steps while you chase it.
References
National Emergency Number Association (2026). NENA Standard for 9-1-1 Call Processing (NENA-STA-020.2-2026) (opens in a new tab) โ
Manufacturing & Service Operations Management (Gans, Koole and Mandelbaum) (2003). Telephone Call Centers: Tutorial, Review, and Research Prospects (opens in a new tab) โ
Pew Research Center (2020). Most Americans don't answer cellphone calls from unknown numbers (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2025). Occupational Employment and Wage Statistics, May 2025: All occupations, plumbing, heating and air-conditioning contractors (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2026). Employer Costs for Employee Compensation, Table 4: Private industry workers by occupational and industry group (June 2026) (opens in a new tab) โ
U.S. Bureau of Labor Statistics (2025). Occupational Employment and Wage Statistics, May 2025: Customer service representatives, hourly median wage, plumbing, heating and air-conditioning contractors (opens in a new tab) โ