Restaurant memberships and subscriptions: what churn shows and the pitch hides

When a subscription falls short (and the number that gives it away)?

Redemption frequency, not sign-up count, is what exposes a badly calibrated subscription: when your members double their weekly visits and the redeemed item carries a food cost above 26%, every month of membership drains money.

At the three Bogotá coffee shops, the twenty most loyal subscribers jumped from 2.3 to 6.1 visits per week against a monthly fee of 89,000 pesos and a direct input cost of 1,180 pesos per redeemed cup: 4,100 dollars lost in eleven weeks. Market average sits near 2.19 restaurant visits per week per diner, according to Revenue Management Solutions via Nation's Restaurant News, so tripling that figure among your most loyal base is not commercial success, it is a leak. Panera sustains its Unlimited Sip Club past a million members because coffee runs at 18 to 26% food cost with a daily redemption cap; Sweetgreen pulled its own after two years because a bowl cannot carry that arithmetic.

Option 1 — Prepaid balance with a bonus: the right entry door

Load 100, receive 115. That is the recurring-revenue model I would install first in almost any operation, because it pulls cash forward without committing the kitchen, and a 15% bonus costs you less than the 38% churn of a poorly designed subscription. Implementation cost approaches zero if your POS already issues gift cards, and the staff learning curve barely fills a week: load balance, deduct, read the unredeemed-balance report. Who is it for? Operations with an average check of 12 to 30 dollars and an office clientele, the crowd that eats near work four days a week and values not pulling out a card. Its structural advantage is that you sell MONEY, not dishes: the margin on each transaction stays yours, the bonus spreads across the whole balance, and redemption can never exceed what you already collected.

Option 2 — Annual benefits club: turning scarcity into money

A single annual payment, somewhere between 69 and 119 dollars for a preferred table, two tastings a year and reservation priority, monetizes scarcity without giving away product. The difference against a consumption subscription is clean: here you charge for ACCESS, and access carries a marginal cost near zero as long as the waitlist is real. Technology cost stays low —one member flag in your reservation engine— but the curve runs about three weeks, most of it spent teaching the host to honor priority without burning the table that walked in without a membership. Who is it for: white-tablecloth restaurants with a verifiable waitlist at least three nights a week. If your dining room does not fill, this club has nothing to sell, and charging for it anyway turns a privilege into a promise your floor cannot keep on a Tuesday in November.

Option 3 — Recurrence through catering and corporate orders

Before redesigning how you charge, check whether your recurrence already lives somewhere else: the repeat corporate order. Some 46% of restaurants offer catering, and those running it as a formal program grow revenue 5.1% against a 3.3% average, according to Technomic and Checkmate. A lunch contract for a 40-person office, three days a week, delivers the same cash predictability as a membership, with two advantages no consumer subscription offers: volume is known 48 hours ahead, and price gets negotiated by batch rather than by craving. Entry cost here is commercial, not technological —somebody has to knock on office-building doors— and your kitchen needs a production line that does not fight with midday service. Who is it for: kitchens sitting idle between 9 and 11 in the morning.

Option 4 — Recurrence with no fee: the properly costed visit point

A classic recurrence program —accumulate, redeem— remains the lowest-risk option, and that is precisely its virtue: the reward gets paid out of margin already banked. If your tenth visit gives away an 8-dollar dish at 30% food cost, the real cost is 2.40 dollars spread across nine prior tickets, meaning 27 cents per visit, under 3% of the check. Set that against a monthly fee where you absorb the risk of a customer consuming more than forecast. The fair objection is that it pulls no cash forward and creates none of the psychological commitment that comes from paying in advance. True. Yet with more than 40% of adults ordering delivery or takeout three to five times a month according to UpMenu, the frequency battlefield is already contested, and no monthly contract is required to win visit number eleven.

How to choose: the three conditions that make a subscription viable?

Memberships pay off when three conditions hold at once, and none of them bends: a product consumed out of HABIT, food cost between 18 and 26%, and an identified base above 400 recurring customers.

Miss a single one and you should pick prepaid balance or the annual club instead. I work this inside the Masterestaurant Restaurant Model Canvas, not inside the media plan, because recurring revenue touches purchasing, shifts and break-even long before it touches a campaign. Run it backwards: if tomorrow 30% of your subscribers decided to redeem the maximum allowed, could your kitchen produce it and your margin absorb it? If the answer is no, you do not have a subscribable product, you have a promise with a ceiling made of air. And here is the paradox: the customer who wants your membership most is exactly the one who costs most to serve, so the daily redemption cap is not fine print, it is the whole model.

The pricing mistake I keep running into: setting the fee against the check instead of the cost

Nearly everyone sets the monthly fee by looking at menu price, and that is where the wreck begins. Your fee should come out of the COST of maximum theoretical redemption, multiplied by expected frequency plus a 25% cushion, and only afterwards compared against what a customer would willingly pay; if those two numbers never meet, the product is not subscribable, full stop. With menu prices up 9.8% since February 2025 in Colombia according to ACODRES, a fee set in 2025 and left untouched eats your margin by itself, with nobody doing anything wrong. Write in an annual review clause from day one and an explicit daily limit —one unit per day, not genuinely unlimited— because «unlimited» in the headline with a cap in practice is exactly what holds up the Panera model.

When NOT to change anything?

Stay where you are if your restaurant sells occasions rather than habits: celebrations, anniversary dinners, tourism.

A white-tablecloth grill with a 60-dollar check and one visit per customer every four months has nothing to subscribe, and building a membership there only adds an administrative layer your host never asked for. Do not move either if your identified base falls short of 400 recurring customers: below that threshold, twenty intensive subscribers can swing your entire monthly food cost, exactly as happened in Bogotá. And if your POS cannot separate member consumption from regular consumption, do not start: without that split you cannot measure anything and you will fly blind for months. In a sector projecting 1.55 trillion dollars in sales for 2026 according to the National Restaurant Association, the pull to copy the neighbor's model is enormous; resist it until you hold the frequency data from YOUR own house.

The four real alternatives, with price and learning curve

ALTERNATIVE 1 — Prepaid balance with bonus (load 100, receive 115). Implementation cost close to zero if your POS already issues gift cards; one week of learning curve. Best for operations with a 12-30 dollar average ticket and an office crowd. Verdict: this is the correct entry door to recurring revenue, since it pulls cash forward without committing kitchen capacity, and a 15% bonus costs far less than the 38% churn of a miscalibrated subscription. ALTERNATIVE 2 — Annual one-payment benefits club (69-119 dollars for preferred seating, two tastings a year and reservation priority). Low technology cost, three-week curve driven by reservation handling. Best for tablecloth restaurants with a genuine waitlist. Verdict: it converts scarcity into money without giving away product, and it remains the only format I have watched hold margins above 70% on the club line. ALTERNATIVE 3 — Packaged-product subscription for takeaway (sauces, whole-bean coffee, sourdough, assembly kits). Startup cost between 2,000 and 6,000 dollars for packaging and logistics; two-month curve.

The four real alternatives, with price and learning curve — in practice

Best for brands with strong culinary identity and a dark kitchen or production area running below capacity. Verdict: it widens the revenue structure toward a 45-60% industrial margin and never depends on the guest walking in. ALTERNATIVE 4 — Frequency-based repeat purchase with variable reward (the tenth monthly visit counts, the first weekly one does not). Almost zero cost, two-week curve. Best for any operation above 900 monthly tickets. Verdict: it delivers roughly 80% of the loyalty effect of a membership at 5% of its risk, and it is what I tell operators to test before touching any monthly plan. The decision tree I use with clients fits into four questions: does your hero product sit under 26% food cost? do you have 400 identified recurring guests in the POS? is consumption habit rather than occasion? can you set a redemption ceiling without the guest feeling lied to? Four yeses, pilot the membership in one unit for ninety days. Three or fewer, go straight to alternative 1 or 4.