Restaurant advertising: the numbers before and after the spend gets reorganized

Cost per seated new guest is the only figure that should chair the meeting

Divide the month's ad spend by the new guests who actually sat down, and you get the number that orders everything else. At the 92-seat steakhouse that opens this analysis, 4,800 USD a month produced barely 340 verifiable new guests: 14.1 USD per seated cover, against a 21 USD average check and a 68% contribution margin, meaning 14.28 USD of margin per visit. The house was paying almost a dollar for every dollar it took in, and the campaign stayed alive because nobody looked at that ratio. The National Restaurant Association documents in its State of the Restaurant Industry 2025 that operators allocate between 3% and 6% of sales to marketing; the trouble is not the percentage, it is that most cannot attribute even a third of that line to a specific transaction. Without guests in the denominator, every campaign looks like a winner.

Repeat visits turn an unsustainable acquisition cost into a profitable one without touching your bid

A guest who returns 4.1 times a year is worth 131 USD in lifetime value against the 58 USD of one who comes 1.8 times, and that 73 USD gap is what lets you outbid your competition in the same channel without losing money. Go back to the steakhouse: at 14.1 USD acquisition and a frequency of 1.8, the LTV/CAC ratio read 4.1 on paper yet collapsed because 44% of those guests never returned. The repeat-visit lever exists and it has been measured. Per the National Restaurant Association 2025, 78% of consumers are more likely to visit a restaurant where they earn points, and 62% —PepsiCo Partners 2025 via Restroworks— increase their visits when the offer is tied to a specific time slot. Moving frequency from 1.8 to 2.6 is worth more than doubling your Meta budget. That 12% of budget feeding the email and WhatsApp list produced nearly half of all measurable reservations: six times the return per dollar versus paid impressions.

Owned channels return a multiple no auction can match

No channel magic there, just prior permission. Sakari measured in 2025 that 84% of consumers had opted into SMS from at least one business, an audience you already own and are not using; and 57% scanned a QR code at a restaurant in the past month, per Sunday 2025, which means the table itself is the cheapest data-capture point in your operation. A QR on the check asking for an email in exchange for a returning-guest benefit costs you the paper it prints on. Every 100 emails captured weekly add up to 5,200 a year, and that is where clean margin lives. Before you put a single peso into promoting your restaurant inside a third-party portal, measure what survives of the order. Rezku published in 2026 that app commissions run from 15% to 30% per order, and CloudKitchens puts the real effective commission, surcharges included, between 35% and 45%.

When delivery eats 40% of the order, advertising that feeds it destroys margin?

On a 25 USD ticket with a 30% food cost, an order leaving 4.25 USD of gross margin supports no customer acquisition cost above 2 USD.

Here is the tension almost nobody resolves: the portal DOES bring incremental volume and genuine visibility, but paying for that visibility only makes sense if you use it to migrate the guest onto your own channel. An insert in the bag carrying a direct-order incentive is advertising, and it is the only kind that pays for itself in that channel. You can buy the impression, but you cannot buy the second when a guest opens your listing and finds 3.6 stars with unanswered reviews. BrightLocal reports in its Local Consumer Review Survey 2025 that 96% of consumers are willing to write a review and 89% expect a reply to positive ones as well as negative ones. Translated into cash: if your click-to-reservation conversion runs at 4% with a neglected listing and climbs to 6.5% once reviews get answered, your customer acquisition cost drops 38% without moving a dollar of budget.

Online reputation is the filter deciding whether your advertising converts or evaporates

That is the correct order of investment. Diego F. Parra insists in MASTERESTAURANT method audits on reviewing the listing before the pixel, because driving traffic toward a weak reputation is paying for the privilege of being turned down. No range in this piece works as an absolute number, it works as a contrast against your own spreadsheet. Small restaurant, 40 to 60 seats and sales under 45,000 USD a month: allocate 3% to marketing —roughly 1,350 USD—, put 60% into owned channels and reputation, and cap acquisition at 6 USD per cover. Mid-size operation, 90 to 140 seats: 4% to 5% is the healthy band, and paid advertising tolerates up to 10 USD acquisition provided annual frequency clears 2.5 visits. Group of three or more locations: negotiate 5% to 6% with a centralized content budget, because three weekly Reels sustained across a quarter amortize their production across every location and the cost per site falls to a third.

How to read these numbers in YOUR operation: three size scenarios?

Run your own ratio before moving one line. Take the steakhouse: 2,880 USD a month leaving Meta and the portal, entering the database, time-bound offers and a points program.

Month one, new reservations fall between 20% and 30%, because cold traffic genuinely goes dark and you have to absorb that dip. Month two, frequency starts to move and repeat visits cover the hole. By month three, with frequency at 2.4 and the 82% of consumers who per Savings.com 2025 say coupons help them against high prices, the same cover volume holds with 1,920 USD less in paid spend. Annualized, that saving is 23,040 USD owing nothing to any auction. I got this wrong for years: I thought switching off the ads meant retreating, when it meant no longer subsidizing impressions that never touched a table.

Where these benchmarks come from and what you should NOT ask of them?

The market figures used here are public and verifiable:

National Restaurant Association 2025 for marketing allocation and loyalty programs, BrightLocal 2025 for reviews, Sakari 2025 for SMS, Sunday 2025 for QR codes, Rezku and CloudKitchens 2026 for delivery commissions, PepsiCo Partners and Savings.com 2025 via Restroworks for promotions. The operating ranges —3% to 6% of sales, the acquisition ceilings per scenario, an LTV/CAC floor of 3— are field reading from the Masterestaurant method applied to those public datasets, not primary research with a sample. Three honest limits: consumer studies are mostly US market, commission data shifts by city and contract, and no declarative survey predicts behavior with precision. Use them as a measuring stick against your own books, never as a target. The denominator moves from impressions to guests. As long as reach heads the report, every campaign looks like a winner; once the number presiding over the meeting is customer acquisition cost in dollars per seated cover, half the budget lines collapse on their own in the first review.

Four differences that explain the jump from 2.4x to 12x

Repeat business enters the advertising equation. A guest returning 4.1 times a year carries 131 USD of guest LTV against 58 USD for someone visiting 1.8 times, and that 73 USD gap is precisely the margin that lets you outbid your competition on the same channel without losing money. Audiovisual content stops being decoration and becomes acquisition inventory. Three Reels a week sustained across a quarter build a retargetable audience that cheapens every later campaign; sporadic production accumulates nothing and forces you to buy cold audiences at list price every month. Online reputation gets managed like a paid channel. As Kim Kaupe, ZinePak co-founder and a frequent consumer-sector speaker, argues, the social proof other customers generate keeps working while the business sleeps, and in restaurants that is literal: BrightLocal's Local Consumer Review Survey 2024 measured that 75% of consumers read restaurant reviews regularly, more than any other local category.