Most restaurants file the phone under operations. It sits with utilities and payroll, a cost of being open. Meanwhile the delivery marketplace commission gets filed under marketing, because someone decided commission buys customers.
That filing decision distorts every comparison you make afterward. The phone brings you customers too. Some of them are new, and being reachable is what converted them. Price it accordingly and the channel comparison stops flattering the platforms.
What you are actually paying for when the phone rings
A call arrives because someone already found you. They saw the sign, searched your name, got a recommendation, or looked at a listing. That discovery was paid for somewhere, whether in rent for a visible corner, in whatever you spend on local search, or in twenty years of goodwill.
The call is the last step in that funnel. What you spend to be reachable at that moment, a person available to pick up or software that always is, is conversion spend on demand you already generated. If the phone rings out, you paid for the whole funnel and dropped the customer at the register.
That is the framing that makes the number worth calculating. Everything you spent to make the phone ring is stranded if nobody answers, and the amount is larger than the answering cost by a wide margin.
Separating acquisition from fulfillment
The mechanical work is telling first-time callers apart from repeat ones, and most restaurants cannot do it today.
If your POS stores a phone number with each order, you can. Pull a year of phone orders, group by number, and mark each order as first-seen or repeat. It is a spreadsheet job, not a project. The output is a monthly count of new phone customers, which is the denominator you have been missing.
If you do not capture phone numbers, that is the first fix, and it is worth doing regardless of this exercise. It also makes loyalty programs on phone orders possible and gives you a list for anything you might send later.
Two cautions on the grouping. Households share numbers, so one entry can represent several people, which matters if you were planning anything personalized. And a number that appears once two years ago and once last week is not a repeat customer in any useful sense, so pick a window, twelve months is reasonable, and treat anything older as a fresh first contact.
Expect the split to surprise you in one direction or the other. Neighborhood pizza counters often find that the overwhelming majority of phone volume is repeat business, which reframes the phone as retention infrastructure. A restaurant in a tourist district or near a hotel often finds the opposite.
Doing the division
Take a month. Add up your phone-answering cost using the labor arithmetic in what a phone order costs to take, then add any spend that specifically drove calls rather than walk-ins.
Divide by first-time phone customers who ordered.
If your monthly phone cost is a few hundred dollars in loaded labor and you acquired thirty new customers by phone, your acquisition cost per new phone customer is in the low tens of dollars, and it came with a full-price ticket rather than a discounted one. Compare that against what an introductory discount costs you, or against a marketplace commission on a first order plus the fact that the platform keeps the customer.
The comparison usually favors the phone. It also usually reveals that the phone is doing more retention than acquisition, which leads somewhere useful rather than nowhere.
One warning about the denominator. If you are missing a meaningful share of calls, your first-time customer count is suppressed and your acquisition cost is overstated, because the cost of being reachable was incurred for callers who never reached you. Fix the measurement of your miss rate before you take the acquisition figure to anybody as an argument.
Why the marketplace comparison is not apples to apples
A delivery platform acquires a customer for you and then charges you a commission on every subsequent order from that customer, forever, because the relationship lives inside their app. Your acquisition cost is low and your marginal cost never falls.
A phone customer is the opposite. You spent labor to convert them, and you now hold their number, their order history, and the direct relationship. The second order costs you the same few minutes of labor as the first, with no percentage attached, and the tenth is the same again.
Over a customer's life those two curves separate dramatically, which is why the honest comparison runs on lifetime economics rather than first-order cost. That is the subject of lifetime value of phone customers, and it is where the case for defending the phone actually lives. The commission arithmetic on its own sits in voice AI versus delivery app commissions.
None of this means dropping the platforms. It means knowing which channel you would rather grow when you have a choice about where to point a customer.
The line item nobody puts on the spreadsheet
Missed calls are acquisition losses, and they are the most expensive kind because the acquisition spend already cleared.
A first-time caller who hears eight rings and hangs up does not call back. They call the next place. You paid rent, listings, and reputation to generate that call, and the last five dollars of coverage is what failed. The full accounting is in the real cost of a missed restaurant phone call, and the measurement method is in how many calls does your restaurant miss.
The uncomfortable version of the math: if you are missing calls during your busiest hour, your effective acquisition cost across all channels is higher than you think, because a share of everything you spend on being found terminates in a busy signal.
What to do with the number once you have it
Two decisions come out of this cleanly.
First, if new phone customers are a meaningful share of your new customers, the phone deserves budget as a marketing channel and should be evaluated against your other marketing spend rather than against your labor budget. A few hundred dollars a month that reliably converts inbound demand is competitive with almost anything you could buy in local advertising.
Second, if repeat callers dominate, stop calling it acquisition and start treating it as the retention channel it is. Then the question becomes how much you would pay to keep your best customers ordering directly instead of through an app that charges you for the privilege, and that number is usually higher than what coverage costs.
Pull the phone numbers out of your POS this month and do the first-seen versus repeat split. Whichever way it comes out, you will be arguing about the phone with data instead of instinct, and you will know whether the money you spend answering it belongs in operations or in marketing.