Marketing metrics that matter: traditional method vs the Masterestaurant method

Chapter 1 — The five metrics that govern the decision, and why reach is not one of them

A useful restaurant dashboard holds five boxes: acquisition cost by channel, 90-day repeat rate, contribution margin per order after commission, conversion from the Google profile to an order, and lifetime value of the identified guest. Reach does not make the list, and the reason is arithmetic: when 99% of restaurants already run at least one social profile and 72% of people use those networks to research where to eat, per Restroworks (2025), presence stopped separating a business from the place next door. What separates them is how much of that attention ends in a transaction with margin. A complete Google profile is 7 times more likely to earn clicks, per WebFX (2026), and that click can be traced all the way to the register; an impression cannot. Measure what the register records. Every quarter you report impressions instead of identified guests costs you, on average, the full margin of the repeat visit you never activated.

Chapter 2 — The cost of measuring vanity, expressed in margin dollars

The numbers hold it up: 70% of first-time guests never come back, per Restroworks (2025), while email returns 36 dollars for every dollar spent per Litmus (2024) and as much as 42.24 dollars per the DMA (2024). A location billing 800 thousand dollars a year and serving roughly 26 thousand guests leaves, at that leakage rate, close to 18 thousand people uncaptured whose contact details cost nothing to ask for at the table. That is not a marketing loss. It is an accounting loss dressed up as follower growth, and it lands in the income statement as an advertising expense that repeats every month because it never built an owned base. A third-party delivery order with a 30-dollar ticket and 28% commission leaves 21.60 gross; at a 30% food cost, contribution margin drops to 12.60, while that same ticket through an owned channel leaves a clean 21 dollars of contribution.

Chapter 3 — Subtracting commission BEFORE comparing channels rewrites the whole ranking

The gap, 8.40 dollars per order, is precisely what funds the loyalty program almost nobody builds: loyalty programs return 4.8 times on average and 90% of operators report positive returns, per Welcome Back (2026). And the consumer already sits on the right side: 70% prefer ordering directly from the restaurant rather than through an intermediary, per Paytronix (2024), and 67% prefer doing it from the venue's own site or app, per Statista. When the report comes in gross, the third party always wins the comparison because it competes with clean hands. Below 500 thousand dollars a year, the metric that rules is conversion from the Google profile, and nothing else: with complete profiles earning 7 times more clicks per WebFX (2026), fixing photos, hours and menu pays better than any paid campaign. Between 500 thousand and 1 million, the 90-day repeat rate takes over, because there the 70% first-guest leakage reported by Restroworks (2025) already amounts to tens of thousands of dollars.

Chapter 4 — The same metric carries different weight depending on the annual revenue band

Above 1 million, acquisition cost by channel enters, once the budget is large enough for two channels to compete. Past 5 million, the axis becomes margin per order after commission, with delivery carrying between 25% and 40% of sales. And above 10 million, lifetime value of the identified guest rules, the only figure that supports a group valuation. A celebrity restaurant or a large-format themed venue above 5 million dollars a year carries a metric the mid-size operator never needs: the cost of staying relevant. Here the reservation is the product, and Toast (2025) data marks the terrain with seated reservations growing 8% year over year on a comparable basis, solo reservations climbing 22% in the third quarter and Tuesday gaining 15% year over year, the biggest jump of any day. That format pays for content production, public relations and event programming, line items that never show up as marketing in its books and that push acquisition cost artificially down.

Chapter 5 — The high end, the celebrity chef and the cost that never shows on the dashboard

When a campaign with local food creators returns roughly 8 times and adds 30% in reservations the following week, per Get Sauce (2025), the right question is not how much it returned: it is how long it lasted. Diego F. Parra builds the Masterestaurant dashboard from four sources you already pay for: the point of sale, the Google profile, the email platform and the aggregator's commission statement. Five numbers a week come out of there, not twenty. Order matters: first margin per order after commission, because it decides whether bringing more volume is worth anything; then repeat rate, the only lever that lowers acquisition cost without renegotiating fees. Email is the cheap lever on that second box, and the birthday coupon gets redeemed three times more often than a standard offer, per Stripo (2025). I got this wrong for years, recommending acquisition first and retention second: it runs the other way, because acquiring onto a base that leaks at 70% is filling a bucket with holes.

Chapter 6 — What happens to these five figures under a cost shock of 5%, 12% and 20%

Assume input costs rise 20% and you leave the price alone: on a 30-dollar ticket with a 30% food cost, raw material goes from 9 to 10.80 and contribution margin falls from 21 to 19.20 dollars, down 8.6%. Under a 12% shock the drop is 5.1%, and under a 5% shock barely 2.1%. The practical consequence lands in the third-party channel: at 28% commission, that same 20% hit leaves margin per order at 10.80 dollars, and any campaign costing more than that per new guest destroys value while the platform panel shows growth. Half of the people who stopped dining out would return with lower prices, per Circana (2025), so raising the price is not always the way out. Cutting dependence on the expensive channel is. In the first two weeks, bring the Google profile to 100% and start collecting reviews, because the local pack's top three carries 47 more reviews on average than positions 4 through 10, per BrightLocal (2025).

Chapter 7 — The first 90 days: what to install, in what order and with which cutoff figure

From day 15 to 45, capture email and phone on every order and at every table, aiming to identify 40% of the period's guests; 75% of restaurants worldwide already use QR codes for digital menus, per QR Code (2025), and that same QR captures the contact. From 45 to 90, measure repeat rate and cut the channel whose margin per order, commission already subtracted, sits below your menu average. One warning about content: 84% of guests prefer seeing food and drink photos on a restaurant's social feeds, per Toast (2024). Start this week with the profile. The unit of measurement. A traditional dashboard counts EVENTS —an impression, a click, a like— while the Masterestaurant method counts PEOPLE with a purchase history, the only thing you can sell to again. A 3-to-10-unit group reporting twelve million impressions without knowing how many unique guests it identified during the quarter does not have a digital marketing problem.

Chapter 8 — Four differences that change the quarter's outcome

It has a customer accounting problem. When commission gets deducted. If delivery revenue enters the report gross, the third-party channel always wins the comparison, because it competes with clean hands against channels that do carry their costs. Subtract commission before comparing and direct delivery conversion changes rank, and with 67% of consumers preferring to order from the restaurant's own site or app (Statista), the direct channel stops being a romantic idea and becomes the line with the best contribution margin. The horizon. Fourteen days measure a campaign, ninety days measure a business. Given that 70% of first-time diners never come back (Restroworks, 2025), an operator who closes the books at day fourteen records as a win what the full quarter shows as a leak: acquisition paid for one single check, then paid again next month for the same customer profile. Where the data ends up. Under the traditional method the data stays on the platform and has to be repurchased monthly; under the Masterestaurant method it moves into an owned asset that yields.

Chapter 9 — Four differences that change the quarter's outcome — in practice

The gap is economic and it is documented: email returns between 36 and 42 dollars per dollar invested according to Litmus (2024) and the DMA (2024), a return no paid channel sustains structurally.