What a missed dinner-rush call is actually worth
Most restaurant owners know they lose some money to unanswered phones. Almost none have done the math. A back-of-napkin model anyone can run in ten minutes.
A friend who owns a 32-seat pasta restaurant in Brooklyn told me recently that he never thinks about the phone. He has bigger problems. The bar runs out of natural wine on Saturdays. His sous chef wants Tuesdays off. The phone rings during service and someone either picks it up or they don’t. He couldn’t tell you which happens more often.
I asked him to humor me for ten minutes. We pulled his call records from his number provider (most VOIP carriers will show you this, as will Google Voice and Verizon Business) and looked at one Friday in April. Between 6 and 9 pm he received 47 inbound calls. 18 were either missed outright or hung up after more than 45 seconds of ringing. Three weeks of Fridays looked about the same.
That’s where the math starts.
The back-of-napkin model
You need three numbers. All three are knowable in about ten minutes:
- Your missed-call rate during peak hours. Pull a sample week. Count rings that went unanswered or to voicemail during your busy window. Divide by total rings in that window.
- Your average phone-order ticket. Your POS knows this. Most kitchens we’ve looked at land between $34 and $58. Phone orders skew higher than walk-ins because groups tend to call ahead.
- The share of missed callers who would have ordered. Some missed calls are spam. Some are vendors. The honest number is usually 50 to 60 percent for dinner-service restaurants, lower for cafes and bakeries. If you’re not sure, use 50.
For my friend, those three numbers were 38 percent missed during Friday peak, a $46 average ticket, and we used 55 percent as the share of real callers. He gets about 90 peak-hour calls per Friday.
That works out to about 90 × 0.38 × 0.55 × $46 = $866 in missed revenue, on a single Friday. He runs three peak Fridays a month and two slower ones, plus comparable Saturdays. Call it $7,800 a month in orders that called him and got nothing back.
His reaction was, “That can’t be right.”
It might not be exactly right. The 55 percent figure is a guess. The missed-call sample was three weeks, not three months. But cut it in half and it’s still a number that gets your attention.
Why this happens
It’s not that anyone is being lazy. The phone rings at the worst possible moment, when the floor is full and the kitchen is slammed. The person who would normally pick up is plating, or running a credit card, or apologizing to the four-top by the window. They hear the ring, they glance at it, they don’t go answer it because the four-top is right there and the phone caller is probably just asking what time you close.
Most of the time they’re right. The caller was asking what time you close. But one in three calls is someone trying to order $80 of pizza for their kid’s birthday, and now they’re calling the place down the block.
What restaurants have tried
Voicemail with a callback. Almost no one calls back after leaving a voicemail at a restaurant they’ve never been to. The activation energy is too high. By the time a host clears the floor and dials back, it’s 9:30 and the person ordered pho.
A dedicated phone person. Works at a certain scale. If you take ten or more peak-hour calls an hour, a host with a headset earns the labor cost back. Below that, you’re paying a wage for a job that’s idle 70 percent of the shift.
An outsourced answering service. The ones built for restaurants are fine. They’re also 80¢ to $1.20 a call, plus a monthly minimum, and the operator doesn’t know your menu, so order accuracy is hit or miss. Sometimes you end up with a takeout ticket that says “the pasta one” and a phone number to call back.
An online-only policy. Some restaurants have given up on the phone entirely and route everyone to a website. This works if every one of your customers already uses online ordering. For most rooms, it leaves real revenue on the table.
Why we built Ocrila
We didn’t set out to build a phone product. We were watching how good Claude had gotten at messy conversations, how cheap ElevenLabs voices had become, and how Deepgram’s word-error rate on noisy audio had quietly dropped below the level where it mattered. The three pieces became good enough at the same time. A phone host that could handle a real Friday order, an “extra-spicy half-pepperoni half-margherita on cauliflower crust for a 7:15 pickup,” was suddenly possible without half a million dollars of voice infrastructure.
So we built it. The economics are straightforward. A Starter plan is $199 a month. If you’re losing $1,000 in missed orders in that same month, it’s a fivefold return.
The numbers above might be off for your specific room. Run yours. If the result is small, ignore us. If it’s not, we’re here.