Customer loyalty: the traditional method vs the Masterestaurant method

When the stamp card stops earning its keep?

A stamp card stops working for a restaurant the day more than 60% of redemptions come from guests who already visited three or four times a month before the program launched, and that number lives in your POS, not in your agency's deck.

Matching each card's issue date against the consumption history tied to the same phone number takes about two hours, and almost nobody does it. An earlier symptom gives it away: stamps climb month over month while average ticket and daily covers stay flat. At that point the program bought applause, not visits. The cost is real, since Bain & Company puts the price of acquiring a new customer at 5 to 25 times the cost of retaining an existing one, a figure quoted constantly to justify the card and almost never to ask whether the retained guest needed the discount at all. A flat 10% to 15% off the check genuinely works in one case: the restaurant under twelve months old that needs to build a customer base fast and has no frequency history to measure anything against.

Flat percentage off the check: who it fits, what it drains

There the program acts as cheap advertising. Switching costs nothing because there is nothing to dismantle, and running it is where the pain shows up: on a 28 USD ticket with a 68% contribution margin, that 15% takes 4.2 USD out of a 19 USD contribution, roughly 22% of what the dish leaves behind. For an owner with two years of data and recurring guests already identified, the same mechanic is pure margin transfer, because the guest who spends most is the one capturing the biggest discount. Setting a reward with a known cost —a dessert at 1.4 USD food cost, a house drink— flips the arithmetic of the traditional program, and that flip is the entire point. Under a percentage discount, a 40 USD check costs you 6 USD in reward; under a fixed reward it costs 1.4 USD, and if the guest climbs to 55 USD because someone came along, the reward still costs 1.4 USD.

Fixed-cost rewards: the option that gets cheaper as you sell more

The program turns CHEAPER exactly when it sells most. This fits the restaurant with a broad menu and low-food-cost, high-margin desserts or drinks, which describes most urban casual dining. Switching costs one afternoon reconfiguring the POS and reprinting material, plus the awkward conversation with the guest holding accumulated stamps that you have to honor through expiration. Paying for frequency means the reward triggers only when a guest beats THEIR own baseline of monthly visits, not when they collect stamps. If someone has come three times a month for a year, their fourth visit that month is the first one you should pay for, and the previous three cost you nothing. Diego F.

Paying for incremental frequency: what it is, who can run it

Parra runs it this way inside the Masterestaurant framework precisely because customer loyalty is an acquisition-cost decision, not a campaign: if a Google Ads lead in restaurants and food runs 30.27 USD according to WordStream's 2025 benchmarks, and ChowNow places the cost of acquiring a new guest between 30 and 80 USD, a 1.4 USD reward that buys one extra visit from a current guest wins that comparison twenty to one. It does require identifying the guest at every visit, and that is the real filter. Measuring frequency demands three things a cardboard card never asks for: an identifier per guest, a POS that keeps history, and somebody who reads the report monthly. Cheapest identifier is still the phone number captured at booking or at check close, with response rates that keep the channel viable —Constant Contact reports SMS marketing conversion between 21% and 30%— and it needs no proprietary app or custom software spend.

What measuring visits instead of stamps demands from your floor?

The mistake I see repeated most is launching the program without defining the baseline it will be compared against, because without that starting number every result looks good.

Measure three months of per-guest frequency first, design the reward after. Reversed, you launch something nobody can evaluate, and that is the most common ending of all. Rewarding with visibility rather than product is the least explored option and the only one that takes nothing off margin: the guest posts a photo, tags the restaurant, and gets booking priority, the table they want, or early access to a seasonal menu. The economics back it, because Loop.fans measures user-generated content converting 4 times better than brand photos, Emplifi logged UGC posts in the third quarter of 2025 outperforming non-UGC posts by more than 10 times, and Instagram engagement grew 28% among active users according to Restroworks. It fits restaurants with photogenic plates and a dining room worth showing.

Social redemption: the alternative that never touches food cost

It fails in high-turnover fast food and in pure delivery, where there is no experience to post and the reward reads as asking a favor, which is exactly how guests receive it. Cutting the reward for the three-visit regular so you can pay only for the fourth creates real friction, and anyone claiming otherwise has not been on the floor when a regular asks why the bar just moved. The paradox resolves by separating recognition from discount: the loyal guest gets recognition through treatment —their name, their table, the dish the way they like it, the chef's recommendation before anyone else— and the incremental guest gets paid in product. That costs nothing and works better, because high-frequency guests rarely come for the 10%; they come because nowhere else treats them that way. Turn that treatment into a discount and you have put a price on a relationship that had none, and afterward there is no way to withdraw it without it landing as punishment.

The uncomfortable tension: your loyal guest notices

That damage never shows up in a report. Keep the stamp card if your restaurant bills under 15,000 USD a month, has no POS with per-guest history, and the current program costs less than 2% of sales, because the administrative cost of migrating eats the improvement. Keep it too if you are in peak season or mid menu change: two simultaneous changes make it impossible to know which one moved the number, and that diagnosis is worth more than the savings. There is a third case, the most honest one: if your problem is traffic rather than repeat business —daily covers falling while the recurring base holds steady— no loyalty program fixes it, because you do not have a loyalty problem, you have an acquisition problem and you are spending on the wrong side. Measure covers per recurring guest this month before touching anything. FIRST, what the reward actually is.

Four differences you can read in the P&L

Traditional programs discount the bill, so cost scales with the ticket: the better the restaurant does, the more expensive the program gets. The Masterestaurant method fixes a reward with a known cost — a dessert at 1.4 USD of food cost, a house drink — and that cost holds even when the guest spends 40 USD, which makes the program cheaper precisely when sales are strongest. SECOND, who gets paid. Handing 15% to everyone who shows a card transfers margin to the guest who was already loyal, since that guest collects the most stamps. Working on monthly frequency with a reset, the reward unlocks only once a guest clears THEIR own visit baseline, and at that point you are buying a visit that did not exist; it is the only way to know whether customer loyalty created repeat business or merely subsidized it. THIRD, who owns the data. A paper card leaves you nothing: no phone, no last-visit date, no favorite dish.

Four differences you can read in the P&L — in practice

Without that there is no remarketing, no 45-day win-back for lapsed guests, and online reputation stays a matter of luck. With an identified base, a win-back campaign to 600 lapsed guests costs under 30 USD in messaging and recovers 40 to 70 visits. FOURTH, the channel. I got this wrong for years, recommending programs that ran identically on platforms and in the dining room, and it was an expensive mistake: when the reward redeems inside the aggregator, you pay the reward AND the commission. Restricting redemption to owned channels — your ordering site, WhatsApp, the floor — turns the program into the cheapest lever for direct delivery conversion available today.