How to run a one-week call audit at a service company

What a one-week audit answers

Most owners know roughly how busy the phones are. Few know how many callers hung up, how many wanted a price and nothing else, or how many real jobs went to voicemail on Saturday morning. A one-week audit answers those questions with your own data, and it costs nothing but a few hours of tagging.

• How many unique people called, and how many reached a person?
• Of the people who wanted service, how many booked?
• When did the unanswered and abandoned calls happen?
• How quickly, if ever, did someone call the missed callers back?
• Which calls were not jobs at all, and how much time did they take?

Before day one: collect every phone line

An audit that covers only the main office line will undercount. In small trades the owner's cellphone is often the real front desk. More than half of US plumbing, heating and air-conditioning contractor establishments, 58.3%, had fewer than five employees in 2023^[3], and in shops that size calls get forwarded to whoever is free.

• Main number and any tracking numbers from ads, your website or vehicle wraps.
• Cellphones that receive forwarded calls, including the owner's and the on-call tech's.
• Answering service or overflow reports for the same seven days.
• Voicemail boxes, with time stamps, plus your web form and text inbox for the same week.

Pick an ordinary week. Avoid the week of a heat wave, a hailstorm or a holiday unless you specifically want to study a surge. Tell the team the audit is about the system, not about catching individuals out, or people will start answering differently for seven days.

The five outcome buckets

Add one optional tag, Existing job, for customers calling about a visit already booked. Those calls matter for workload, but they do not belong in your booking-rate math.

Seven days, step by step

1. 01

   Day 0: set up the sheet

   One row per call: date, time, line, caller number, duration, answered by, bucket, job type, notes, and callback time if it was missed. Freeze the header and add a drop-down for the bucket column so tags stay consistent.

2. 02

   Days 1 to 7: tag at the end of each day

   Ten minutes at close is enough for most shops. Listen to recordings only for calls you cannot place from the log. If a caller tried three times, tag the person once and note the attempts.

3. 03

   Every day: record callbacks

   For each unanswered caller, write down when someone called back and what happened. This is the column most audits skip and the one that explains most lost work.

4. 04

   Day 8: clean the data

   Merge duplicate callers, fill missing job types and move existing-job calls out of the booking-rate denominator.

5. 05

   Day 8: build the one-page summary

   Use the formulas in the next section and sort unanswered calls into hour blocks. Keep the raw sheet as a baseline for the next audit.

The one-page summary

• Answer rate = unique callers who reached a person ÷ unique callers.
• Booking rate = booked ÷ (booked + lost + price shopper). Spam and existing-job calls are left out.
• Missed bookable calls = unanswered callers whose voicemail, callback or later call showed a real job.
• Callback speed = median minutes from missed call to first return attempt.
• Recovery rate = missed bookable callers who booked after a callback ÷ missed bookable callers.
• Top three hour blocks for unanswered calls, with the count in each.

Write two sentences under the numbers: the biggest leak and the first fix. A summary that ends in a decision gets acted on; one that ends in a chart gets filed. If you want to put a dollar figure on the leak, the missed-call cost formula uses the same inputs.

Do not skip the weekend

Many service companies close the office on Saturday and assume demand drops too. Homeowners see it differently. Only 30% of employed Americans worked on an average weekend day in 2025, compared with 81% on a weekday^[1]. Weekends are when people are home, notice the dripping ceiling or the dead opener, and finally have time to call. Your audit should show whether those calls go anywhere.

Example: what one week revealed

Example: a fictional three-trade company (HVAC, electrical and generators) logged 312 unique callers in a September week. After removing 41 spam and vendor calls and 66 existing-job calls, 205 callers wanted new service. The office answered 171 of them and booked 112. Twenty-eight price shoppers and 31 lost calls made up the rest of the answered group. Of the 34 unanswered bookable callers, 19 were reached by callback, but only 7 booked, because the median callback came 3 hours and 10 minutes after the missed call.

Example summary line: answer rate for new-service callers 83%, booking rate 65%, recovery rate 21%. Biggest leak: 7:30 to 9 a.m. on weekdays and Saturday 9 to noon. First fix: move one CSR's start time to 7:15 a.m. and route Saturday calls to the on-call rotation instead of voicemail.

Callbacks are where audits surprise people

Slow follow-up is common across industries, not just home services. In an audit of how US companies handled web leads, 37% responded within an hour, 24% took more than 24 hours and 23% never responded at all^[2]. Phone callbacks fail for an extra reason: people do not pick up. Only 19% of US adults say they generally answer calls from unknown numbers^[4].

1. Return every missed bookable call within 15 minutes during open hours, and say the company name in the first three words of the voicemail.
2. If the first callback fails, try once more within the hour, then stop.
3. Record the outcome in the same sheet so the next audit shows whether speed improved.

Reading the hour-by-hour pattern

Once unanswered calls are sorted into hour blocks, the shape usually points to one of four causes. Match your pattern before you spend money on a fix.

• A spike at opening time: the phones start before the office does. Customers call on their way to work, and techs are calling in with questions at the same moment. A staggered start fixes it.
• A spike at lunch: everyone eats at once. Rotate breaks and forward the line to a second person rather than letting it roll to voicemail.
• Misses spread evenly across the day: the office is simply understaffed for the volume, or the person answering also does dispatch, billing and parts runs.
• Misses concentrated after close and on weekends: this is a coverage decision, not a staffing tweak. Compare an on-call rotation, an answering service and an AI employee on cost per booked job.

Look at abandoned calls separately from calls that went to voicemail. A caller who hangs up after 40 seconds in a queue is telling you something different from a caller who never heard a ring at all.

Mistakes that spoil the data

• Counting call attempts instead of people, which inflates volume and hides repeat callers who gave up.
• Leaving spam in the denominator, which makes the booking rate look worse than it is.
• Tagging from memory at the end of the week instead of daily.
• Ignoring the owner's cellphone and the on-call phone, where after-hours work often lands.
• Auditing once and never repeating it. Run the same audit each quarter so you can see whether a fix worked.

Once the summary is done, compare your answer rate with the targets in our answer rate benchmark guide, and use the booking-rate figure to coach the team on the calls they answered.