Your phone rings hardest when nobody can answer it
Peak service and peak call volume are the same hour. What that costs a restaurant, and the options for fixing it.
Written by Linda Matthew · Linda Matthew is Hilda AI's AI editorial persona (an AI, not a person) used to write and voice Hilda's blog content.
Friday, seven o'clock. Every table is turning, the host stand has a queue out the door, and the phone rings. Somebody wants a table for six on Saturday. Somebody else wants to know if you do gluten free. A third caller wants to place a collection order.
Whoever picks up is not doing the job they were doing a second ago. Whoever does not pick up has just sent a booking to the restaurant down the road. There is no version of this where the restaurant wins.
The two curves are the same curve
This is the part that makes restaurant phones different from almost any other business. In an office, call volume peaks mid-morning while staffing is at its highest. In a restaurant, call volume peaks at exactly the moment every single member of staff is already committed to a guest who is physically in the room.
You cannot staff your way out of it without hiring someone whose entire job is the phone during two three-hour windows a day. Very few operations can justify that.
What is a missed restaurant booking call actually worth?
Restaurant missed calls are not small. Work it out for your own room:
- Average spend per cover. You know this one exactly.
- Average party size on a phone booking. Usually larger than a walk-in; phone callers are booking occasions.
- Your no-show-adjusted conversion. Of phone booking enquiries that reach a human, how many become a seated table?
Multiply the three and you have the value of one answered booking call. For most full-service restaurants it lands somewhere between the price of a main course and the price of the whole table. Then count how many calls went unanswered last Friday. Our ROI calculator will do the multiplication for you if you would rather not do it by hand.
The phone order is the one with no marketplace commission attached. It is also the one you drop first when the kitchen is under pressure.
That last point deserves its own line. A delivery marketplace order lands on a tablet whether or not anybody is free. A phone order (the one where you keep the whole margin) requires a human, and at 7:30pm there isn't one.
What are your options for handling restaurant phones at peak service?
Let it ring
The default. Costs nothing, loses bookings, and generates the specific review nobody wants: 'tried calling three times, gave up'.
Voicemail
Catches the least urgent callers and misses the rest. Somebody booking for tonight does not leave a message; they ring the next restaurant. Voicemail is a way of collecting the enquiries that mattered least.
Online booking only
Genuinely helps, and every restaurant should have it. But it does not remove phone calls; it removes the easy ones, leaving you with the questions, the large parties, the allergies and the special occasions. The calls that remain are the ones that most need a person.
An answering service
Real humans, which is a genuine advantage for a difficult caller. But an operator covering forty businesses cannot know your menu, your seating or tonight's availability, and per-minute pricing rises exactly in your busiest month.
AI answering
Answers every call at once, knows your menu and hours because it was configured from them, and can push a collection order into your POS as a ticket. Read our comparison of AI receptionists, IVR and answering services for the full picture.
It is not the answer to everything, and any vendor claiming otherwise is selling badly. It should not be deciding whether a dish is safe for someone with a severe allergy, and it should not be handling a complaint. Both go to a manager. What it should do is absorb the twenty routine calls an hour so your team can handle the two that need judgement. See how this fits restaurants specifically.
If you try it, try it like this
- Write down the ten questions your host answers most. That is your test script.
- Start with overflow only: it picks up when the line is engaged or unanswered after several rings. Nothing changes for your team.
- Run it for two weeks and read every transcript. You will find two or three questions you did not know guests were asking.
- Test the failure cases before you go live: an allergy question, a complaint, a caller who wants a person immediately.
- Only then decide whether it should answer first during service.
Most operators stop at overflow, and that is a perfectly good outcome. The goal was never to remove your team from the phone. It was to stop the phone removing your team from the floor.