Repeat guests consistently spend more per order than first-time diners, and that gap is what should reorder your budget.

Finding 1 — The Bogotá dashboard: 4,100 sign-ups, zero pesos tracked

A points program with 4,100 sign-ups and no attributed-sales report is not loyalty, it is an expensive database. An owner in Bogotá showed me that dashboard in May, two years after opening, with 9.2% of quarterly sales spent on paid media while the healthy budget for an established restaurant runs 3% to 6% according to Toast (2025), and only a newly opened venue justifies reaching 10%. Sign-ups measure themselves because the POS counts them the moment it captures an email; repeat purchase has to be hunted down by matching the customer identifier against twelve months of tickets, and nobody does that by default. Meanwhile, Paytronix (2024) reports that 81% of U.S. loyalty members buy more frequently than non-members. The figure exists. The operator simply does not know whether it applies to him.

Finding 2 — What is a repeat visit actually worth?

Every repeat visit is worth 1.67 acquisition tickets, and that number reorders the entire budget. The logic is simple:

if existing customers spend on average more per order than first-timers, your effective acquisition cost gets divided by the number of visits you achieve in the year, not by the number of emails you captured. An operator who moves a customer from two to five annual visits did not improve loyalty by an abstract margin: he made his marketing cheaper by that same proportion, because the same media peso now buys five tickets instead of two, with fatter tickets on top. Paytronix (2025) measures that fattening separately: 38% more spend per visit versus the walk-in guest. Run the math on last year's own sales before approving the next campaign.

Finding 3 — The flat discount bleeds margin exactly where turnover is highest

A 15% discount on your signature dish bleeds contribution margin precisely on the product that leaves the pass most often, whereas a frequency reward buys the habit without touching list price. Think it through with the optimal food cost the National Restaurant Association places between 28% and 35%: if the dish costs you 32 and you shave 15 points off the price, you did not lose 15% of profit, you lost close to half that dish's gross margin, and you lost it on the customer who was going to buy it anyway. Fourth visit free, a Tuesday reservation with dessert included, the second coffee: those incentives only pay when behavior changes. Paytronix (2024) documents that loyalty members visit more than 40% more often. That differential is what you are buying, not the customer's affection.

Finding 4 — What this analysis measures and what it cannot measure

This work synthesizes public data from Paytronix, Restroworks, Toast, the National Restaurant Association, TouchBistro, LoyaltyPass and Grand View Research published between 2024 and 2026, organized by operating segment —fast casual, full service and QSR— crossed with size of one venue, three to ten, and multi-unit. There is no proprietary sample here and no Masterestaurant survey, and saying so matters: what Diego F. Parra and the team contribute is the READING, meaning which decision each figure triggers and what healthy range you should land in given your size. Where public data stops applying matters just as much. Almost all these measurements are American, and the Latin American operator works with a delivery market that Grand View Research values at US$32.420 billion in GMV for 2025, with a commission structure and a price elasticity none of those sources contemplates.

Finding 5 — The sequencing error: customer data first, reward second

Identify the customer on the ticket first, design the reward second, never the other way around, and the reverse order is the mistake that repeats most in independent restaurant loyalty programs. An enthusiastic operator launches the stamp card, hands out 4,100 sign-ups and only then discovers his POS does not tie the email to the transaction, which leaves him with the program's cost and no numerator to calculate it against. Without that link you cannot answer the one question that counts: what share of your sales comes from people who already know you. LoyaltyPass (2026) reports that loyalty members visit twice as often as digital-only customers, a 2x differential that collapses to nothing if you cannot separate one group from the other inside your own system. Verify the link this week. It costs one call to your POS vendor.

Finding 6 — What happens if your regulars stop coming back next year?

Suppose your repeat customers stop returning and you have to replace those sales by buying new traffic.

With the additional spend that the regular contributes over the first-timer, replacing a hundred repeat visits demands more first-timer visits to match the cash, and each of those is paid at full acquisition price. An established restaurant spending 4% of sales on marketing, inside Toast's 3% to 6% range (2025), jumps to the 10% Toast reserves for openings and still falls short. That is where loyalty stops being a marketing initiative and becomes a cash-flow defense, and Inc. identifies cash flow as the leading cause of financial stress and closure in small businesses. Here is the trade of the craft: the program that looks like soft spending is the only policy you hold against the hard kind.

Finding 7 — Social media brings the first-timer, the program keeps him

Social media solves discovery and does not solve repeat purchase, and confusing the two functions is what inflates budgets without moving the till. Restroworks (2025) measures that 41% of Gen Z uses TikTok to find restaurants; that traffic arrives, orders once and leaves, unless something captures it. On the spending side, Socially Powerful reports US$10.520 billion in U.S. influencer marketing during 2025, a jump of 23.7%, with an average cost per collaboration of US$202 according to Collabstr. Paying those US$202 to land a first-timer who never returns is buying half the business. Tie every discovery campaign to a capture mechanism —email, phone, POS identifier— and measure the second visit, not the reach.

Finding 8 — Healthy ranges by size: one venue, three to ten, multi-unit

The healthy range for member share of sales shifts with size, and Paytronix (2024) documents that top operators reach 37% of transactions through loyalty members. For a single venue, aiming at 15% or 20% of identified transactions in the first year is already a serious result, because your customer universe is the neighborhood and natural frequency sets the ceiling. Between three and ten venues the technical link becomes the bottleneck rather than the incentive: you need a unique identifier that travels across locations, and without it the same customer shows up as three different people. In multi-unit, Paytronix's 37% is reachable because volume allows segmenting and testing. The frequency is there. Capturing it is the work.

Finding 9 — Three differences between a loyalty program and real customer loyalty

The first difference is arithmetic, and almost nobody runs it. If a repeat guest spends more per order than a first-timer, each returning visit is worth more than one acquisition ticket, and your effective customer acquisition cost divides by the number of yearly visits you win. An operator who moves a guest from 2 to 5 visits per year did not improve loyalty by 150%: he cut his marketing cost by the same proportion. That calculation reorders the budget, not the sign-up rate. The second is incentive design. A flat 15% off the flagship dish destroys contribution margin exactly where volume lives, while a frequency reward —the fourth visit, the Tuesday reservation— buys the habit without touching list price. Paytronix data (2025) shows members already spend 38% more per visit than walk-ins; discounting for someone who was going to spend more is paying for behavior you already had. The third difference is channel ownership. Converting delivery into direct orders, with your own name and phone number, is the asset that separates a restaurant with real growth from one that merely bills.