What missed calls cost you.
A formula you can run on your own phone log this week — and an honest read of the data's limits.
Daniel M. Ortiz · September 2026
Somewhere in a sales deck right now, a terrifying percentage of restaurant calls is going unanswered. Ask where the number comes from and the trail usually dies in another deck. We won't repeat any of those figures here.
What this page does instead: it names the one recent dataset worth knowing about, tells you who produced it and why that matters, and then hands you a five-line formula — so the only number you act on is one you computed from your own phone log.
What the call data says
In a 2026 call study self-published by Revmo AI, a voice-AI vendor, covering 12,091 recorded calls, missed-call rates split sharply by segment:
| Segment | Calls missed |
|---|---|
| Quick-service restaurants | 40.1% |
| Fast-casual restaurants | 24.8% |
| Full-service restaurants | 9.0% |
Share of inbound calls missed, by segment. Revmo AI, “State of Restaurant Calls 2026,” 12,091 recorded calls — a vendor's own study; see the caveat below.
Read that table with both eyes open. Revmo sells software that answers restaurant phones. When the company selling the fix publishes the study measuring the problem, the study is marketing with a methods section — possibly careful marketing, but you can't tell from the outside, and neither can we.
So treat these numbers as directional: a reason to pull your own call log this week. Nothing more.
The loss figure making the rounds
The same release estimates that a single 50-location quick-service chain loses about $3.65 million a year to missed calls — roughly $73,000 per location. That is Revmo's arithmetic, not ours, and it leans on the assumption every missed-call headline leans on: that a missed call equals a lost order.
It doesn't, reliably. Some missed callers ring back. Some walk in anyway. Some were a vendor, a wrong number, or your own bartender. We won't print a per-location loss figure as a fact about your restaurant, because nobody can know that number from the outside. You can know it from the inside, though — and the arithmetic takes an evening.
Run the formula on your own phone log
weekly calls × % unanswered × % revenue-intent × average order value × 52
Weekly calls
Pull your phone system's log for the last 90 days and divide the inbound count by thirteen. If there's no log to pull, write that down as finding number one.
% unanswered
The share of those calls that rang out, hit voicemail, or met a busy signal. Most phone portals surface this in two clicks.
% revenue-intent
The honest step, and the one the scary headlines skip: how many missed calls were someone trying to give you money — an order, a booking, an eight-top? Listen to a week of voicemails and count. Vendors, robocalls, staff, and wrong numbers don't count. Be conservative here; the whole result is only as honest as this line.
Average order value
From your POS. Use the phone-order average if you can separate it; the dine-in check average will flatter the result.
× 52
Fifty-two weeks makes it an annual figure — a number in your own handwriting, from your own log, that no deck can argue with.
What the formula leaves out — on purpose
Two real effects it can't capture: the caller who never tries again (the formula understates), and the caller who walks in anyway (it overstates). They pull in opposite directions, and no arithmetic run from a phone log resolves them. The formula's job isn't precision — it's to replace someone else's national average with your own measured leak. And a call that goes unanswered while every table is full was never a table you could seat; count your peak-hour misses with that in mind.
And if your number comes out small: believe it. Spend the attention somewhere else, whatever anyone's deck says. A small number here is good news about your phone, and it's yours.
How we test this in a walkthrough
In an Operations Assessment we don't estimate the phone — we test it: six calls at six different hours, mid-Friday service and after close included, with what happened written down word for word and read back at the readout. It's one of the twelve outside-in tests every walkthrough includes, and all twelve are published.
If your number came out big and you want the whole operation walked — phones to back office: