Measuring missed calls with your own phone data

Before paying anyone to answer more calls, find out how many you actually miss, when, and what happens to those callers. Most phone systems already have the reports.

Vendors in this market often open with a statistic about missed calls. A Goodcall HVAC article cites third-party research claiming HVAC contractors miss 27% of inbound calls and lose an average amount per missed call. Smith.ai’s after-hours page says 27% of leads call outside 9-to-5 business hours. We cannot verify those figures, and even if accurate they are averages across other companies. Your numbers are what matter, and you can usually get them from reports your phone system already produces.

What to measure

Break each measure down by hour of day, day of week, and month. A single monthly percentage hides everything useful.

Where to find the data

• Business phone system or VoIP provider: most have call logs, missed-call reports and queue reports. Look for terms such as call detail records, call history, queue statistics or abandoned calls.
• Call tracking tools, if you use them for marketing.
• Carrier records for basic lines.
• Manual log: if nothing else exists, log two weeks of calls by hand.

Definitions differ between systems. Write down how your system counts each measure so you compare like with like over time.

A two-week baseline worksheet

If your reports are thin, or you want a check on them, a short manual baseline is enough to start. Split the week into the blocks your business actually runs on, and record the same columns for each:

The blocks above are an example; use your real opening hours, shift changes and on-call handover times. Fill it in for two ordinary weeks, then again for one busy week if you can. The difference between the two is the part of the problem that is seasonal rather than structural.

Sort the missed calls by what the caller wanted

A missed tune-up request and a missed no-heat call on a winter night are not the same problem. For a sample of missed calls where you can reach the caller or the voicemail, record which type of call it was:

• Urgent comfort loss: no heat or no cooling.
• Safety-sensitive: callers describing a gas smell, carbon monoxide alarm or similar. These should never depend on a voicemail box; check what your greeting tells them to do.
• New work: repair requests, replacement estimates, tune-ups.
• Existing customers: technician ETA, reschedules, billing, maintenance-agreement members.
• Not a customer: vendors, sales calls, spam, wrong numbers.

The mix tells you what kind of coverage would help. Mostly existing-customer status calls suggests better proactive updates; mostly urgent calls after hours suggests an on-call handoff problem; mostly new work during the day suggests overflow coverage or staffing.

Calls are not the only lead source

Missed phone calls are one gap. Avoca’s HVAC speed-to-lead page treats missed calls as one lead source among web forms, Google Local Services Ads, Facebook, Yelp and Angi, and markets fast response to all of them. Their response-time figures are their marketing claims, not something we have tested. The measurement point stands on its own, though: if you take web-form or ad leads, record the time from submission to first human or automated reply alongside your call data. A company that answers every call but replies to web forms the next morning has the same problem in a different channel.

When voicemail may be enough

Goodcall’s own comparison of answering services and voicemail, written by a vendor that sells answering, still lists cases where voicemail can be sufficient: businesses with minimal after-hours calls and low-urgency work. Your baseline answers that question for your company. If after-hours calls are rare, most are non-urgent, and callbacks happen first thing each morning, a clear voicemail greeting plus a reliable callback queue may be the cheapest fix. If the sample shows urgent calls going to voicemail, or callers who never leave a message and never call back, it is not.

Ask any vendor to report the same measures

Once you have a baseline, any service you evaluate should be able to report the same measures for calls it handles: calls offered, answered, time to answer, messages taken, urgent calls escalated and how quickly, and callbacks completed. If a provider cannot give you those numbers under your definitions, you will not be able to tell whether it worked. Add this to your vendor questions.

How to read the results

• Missed calls cluster at lunch or shift changes: a staffing or routing fix may be cheaper than any service.
• Missed calls cluster after hours: see the overnight and weekend plan and consider after-hours coverage.
• Missed calls spike in the first hot or cold weeks: see the peak-season surge plan and consider overflow coverage.
• Few callers leave voicemail: Goodcall and Smith.ai both argue callers often will not leave messages. Check whether that is true for you by comparing voicemails left with calls sent to voicemail.
• Callbacks take hours: the problem may be ownership, not answering. See the callback queue handoff and the proposed design for missed-call follow-up.

Estimating the value — carefully

It is tempting to multiply missed calls by an average job value. That overstates the problem: some missed callers call back, some were spam or wrong numbers, some would not have booked. A more honest approach:

1. Take a sample of missed-call numbers from a busy week.
2. Check how many later reached you and booked anyway.
3. Check how many were existing customers, new prospects, or spam.
4. Estimate the value of the remainder using your own average ticket for the likely call types.

Treat the result as a rough range, not a forecast.

Repeat the measurement

Measure again after any change — new staff schedule, answering service, AI receptionist — using the same definitions. That is the only reliable way to know whether the change worked. Then use missed-call follow-up rules for the calls that still get through the gaps.

Published 2026-09-26 · Concept preview: describes ideas, not a live service.