Repeat-visit program: before vs after with the Masterestaurant method

1. The blind spot no ad report will ever fix

No marketing dashboard tells you how many of Friday's guests had eaten there before, and that gap is what keeps commercial spending inflated. An owner with 180 seats and revenue above 5 million a year showed me 14 thousand dollars a month in paid media, three agencies, and an average check that had been flat for nine months. Acquiring a new customer costs between 30 and 80 dollars in restaurants, according to ChowNow, and the cost per lead for Google Ads in the category sits at US$30.27, according to WordStream (2025), so every guest who never returns is inventory bought and thrown away. That same person's second visit carries no acquisition cost on top, which is why it moves the unit economics of the whole table more than any bid optimization someone will sell you this quarter. The second visit gets decided at the table, between the greeting and the check, and only afterwards does WhatsApp or email confirm it.

2. Repeat business is an output of service, not of marketing

Changing who owns the metric changes the correction cycle: when the floor manager answers for it instead of the agency, you correct within a week because preshift happens daily, while the ad report lands once the month has already closed. Diego F. Parra insists on an order almost nobody respects —first the service structure that produces the return, then the channel that reminds people of it, and only last the media budget—, and Masterestaurant builds the program from the dining room because the server holds the final contact before the decision. A discount sent to somebody who was badly served recovers nothing: it speeds up the loss and teaches that guest your brand is worth less. Doubling paid media without touching service hands you the same revenue at twice the commercial cost. Walk the whole scenario: more people arrive, the room saturates, table turns jam, wait times stretch, and the average review drops half a point.

3. What happens if you double the media spend and leave service alone

According to Michael Luca's Harvard Business School study (2016), one Yelp star moves between 5% and 9% of an independent restaurant's revenue, so that lost half star eats the entire traffic increase. And you still carry labor cost, which runs between 25% and 35% of revenue according to the U.S. Bureau of Labor Statistics, now spread across an exhausted team. Purchased traffic amplifies whatever your operation already is; it never corrects it. Your server wants to close fast to turn more tables and earn more tips, while the repeat program asks for thirty extra seconds at the close: both positions are legitimate, and pretending otherwise wrecks the rollout. Individual measurement is the bridge. Once the repeat rate per server shows up in preshift alongside tips and turn times, it stops being an imposed chore and becomes each person's own scoreboard. That is the economic argument the floor actually understands: retaining costs 5 to 7 times less than acquiring, according to Invesp, and Bain & Company puts the gap between 5 and 25 times depending on category.

4. The tension with the server is real, and measurement resolves it

If one four-top with a 120-dollar check comes back just once more per quarter, that server produced 480 dollars of annual sales with zero CAC. Tips pay that back. Below 500 thousand dollars a year you do not need a loyalty platform: you need a notebook of regulars and a table-closing script. Concrete threshold: identify your 50 highest-frequency guests, assign each one to a server, and measure whether they return within 45 days. With a CAC of US$30 to US$80 per ChowNow, bringing 20 of those 50 back equals between 600 and 1,600 dollars of acquisition you never paid. Between 500 thousand and 1 million the threshold shifts: there is finally volume for a real database, and the decision is hiring the cheapest CRM that captures name, phone, and date of last visit, nothing else. The trap in this band is buying 400-dollar-a-month software for 900 contacts.

5. Under 500 thousand and 500 thousand to 1 million: the small band stays

Fill the database with 30% of the room first; automation comes later. Past a million in annual revenue, the decision threshold is the 60-day repeat rate: if it sits below 25%, do not open another acquisition channel until you fix it. At this size it pays to attribute the second visit by server and by daypart, because the problem is almost never the whole restaurant but two specific shifts. Above 5 million —or with more than one location— the program needs its own budget and a named owner on the org chart, and the financial contrast gets hard to ignore: 14 thousand dollars of monthly media buys between 175 and 466 new customers at ChowNow's cost, while lifting repeat rate three points on a base of 40 thousand annual visits returns 1,200 visits with no acquisition cost attached. With 62% of consumers discovering restaurants through Google, per Restroworks (2024), traffic will keep arriving; the return trip is what never arrives by itself.

6. Above 10 million and the high-end profile: governance, not campaigns

In a group or chain above 10 million you govern repeat business rather than running campaigns on it: one single table-closing standard, one guest data point captured in the same format at every location, and cross-comparison of the 60-day rate between branches to locate where the loss happens. The operating threshold I use is three points of spread between the best and worst location: above that, the problem belongs to floor management, not to the neighborhood. The high-end case deserves its own paragraph. A large-format themed restaurant, or a media chef's project billing over 5 million, lives off a first visit pushed by notoriety, and repeat business routinely collapses there. Their metric is not the review score —usually high— but visit number two. Without it the curve falls the moment novelty runs out, and novelty always runs out. The first difference is architectural: in the old model repeat business is a byproduct of marketing, and in the new one it is an output of the service process.

7. Three differences a CEO must grasp before approving budget

That changes who answers for it. When the owner of the metric is the floor manager instead of the agency, the correction cycle drops from a quarter to a week, because the preshift happens daily while the media report arrives after the month has already closed. The second is financial and shows up in unit economics. A guest brought by paid media carries US$30 to US$80 of acquisition cost, per ChowNow, while that guest's second visit carries only the cost of the message that prompted it. At the same contribution margin per plate, the repeat visit is structurally more profitable, and that gap funds the entire program without new capital and without moving break-even. The third is risk mitigation, and almost nobody prices it. A restaurant growing only through acquisition depends on platforms that set prices without asking: Google Ads cost per lead already stands at US$30.27, per WordStream (2025), and third-party delivery absorbs 30% to 40% of order value, per Restaurant Business (2024).

8. Three differences a CEO must grasp before approving budget — in practice

An owned base of returning guests is the one asset that cannot raise your commission tomorrow, and in an operational due diligence it counts for more than three good months of revenue.