Restaurant membership and subscription models: the numbers before and after

What does it actually cost to serve what the fee includes?

Between 0.45 and 0.70 USD of product cost per specialty coffee: that number decides whether your membership makes or loses money, and almost nobody runs it before printing the card.

On a 12 USD monthly fee you can absorb roughly twenty served coffees before you go under water, provided the subscriber buys something else on the same visit; if they walk in, collect the coffee and leave, the membership turns into an expensive coupon dressed up as innovation. The industry has the muscle to carry these programs — the National Restaurant Association projects 15.8 million restaurant jobs in the US for 2026, a hundred thousand more than the prior year — but aggregate volume never rescues a badly costed recipe card. Before launching, sit down with the costing of the included product and multiply it by the maximum frequency the plan allows. That product is your ceiling of risk.

Revenue is not the fee: it is the incremental spend

The monthly fee pays for the program's plumbing; the margin comes from the ticket the subscriber adds on top. A coffee customer who walks in three times a week adds pastry around 40% of the time, and with a 3.80 USD pastry at 28% food cost you book close to 2.74 USD of margin on each of those visits: about five a month, thirteen and a half USD of contribution the fee alone would never have handed you. Put differently, a recurring program is not a revenue line, it is a frequency engine. And I got this wrong for years, recommending high fees to "lock in" the cash: a high fee scares off the frequent subscriber, who is precisely the one buying the add-on. Lower the fee, raise the frequency, collect margin on the extra. At 8% monthly cancellation the average tenure is twelve and a half months; at 15%, it drops below seven.

Churn outranks acquisition

That gap decides whether acquisition cost amortizes or burns, because a subscriber contributing 13 USD a month leaves around 162 USD in the first scenario and barely 87 in the second. If acquiring them cost you 40 USD across advertising, platform commission and a discounted first month, the good scenario returns four times the investment and the bad one, two. Chase retention before volume: a three-hundred-subscriber program bleeding 6% is worth more cash than an eight-hundred one bleeding 18% every month. The metric on your weekly board is not "sign-ups", it is CANCELLATIONS, and I want it broken down by tenure month. No included benefit may push a dish's food cost above 32%, and below a 60% contribution margin each extra visit worsens the result instead of improving it. The arithmetic is merciless: if the included dish runs at 34% cost and you hand it over with an effective 25% discount inside the fee, real margin collapses to 44% and that enthusiastic subscriber coming five times a month costs you money every time they walk through the door.

The discount has a recipe-card ceiling

Design the benefit around low-cost, high-perception items — infusions, seasonal starters, the second drink — never around the protein headliner on your menu. The 32% limit is not a Masterestaurant recommendation: it is the tolerable maximum, and working flush against the ceiling leaves zero cushion for a supplier increase. A business showing contracted recurring revenue stops being valued on an EBITDA multiple and starts being valued on the quality of its base. Diego F. Parra insists on an order we do not negotiate at Masterestaurant: first you stabilize food cost below 32% and you know your break-even, then you design the recurring program. Reversed, membership amplifies the problem. A venue at 38% food cost giving away a dish a month multiplies its leakage, and the owner finds out in quarter four, when there are already eight hundred subscribers whose price cannot be raised without breaking the promise. The database, by contrast, is a defensible asset: every visit tied to an identifier turns a mailing list into consumption history, and that gets audited.

What changes in front of a restaurant investor?

Franchising understood this earlier — the International Franchise Association counted more than 4 million franchised fast-food jobs in the US in 2025, up 2.6% — because its model lives on measured recurrence.

The three scenarios I use to land these figures split by volume, not by ambition. Small venue, up to 60 covers and a single strong service: aim for 150-250 subscribers, a fee of 9 to 14 USD, and do not launch until the costing of both included products is closed. Mid-size operation with two services and some delivery: the healthy range is 400-900 subscribers, a fee of 15 to 29 USD, and here you already need a named person reviewing churn every Monday. Group of three venues or more: above 1,500 subscribers the fee can climb to 35-49 USD because the benefit becomes multi-site, but it demands POS integration across locations or the program breaks at the counter.

How to read these numbers in YOUR operation?

Spain has the sector size to sustain it, with close to 1.89 million hospitality workers and 40,000 more in 2025 according to Hostelería de España.

Be honest about provenance: the employment and sector-weight figures come from verifiable public sources — National Restaurant Association, International Franchise Association, Hostelería de España, CANIRAC with INEGI for Mexico and its 2.1 million direct jobs, around 1% of GDP in 2024 — while the fee, churn and margin ranges in the tables are consulting criteria applied to restaurant operations, not a statistical sample. Nobody has censused membership programs in independent restaurants, and whoever sells you an "industry" churn percentage is selling you smoke. Use them as an order-of-magnitude reference to calibrate your own board, never as a promise of results. The only figure governing your decision is your own costing, and that one you already have at home. A recurring program is not judged in quarter one, it is judged once it has run twelve months on the same promise.

The twelve-month test before you scale

What happens if you launch in March, close 600 sign-ups in the first half and decide to open a second venue against that projected revenue? If real churn turns out to be 14%, by month twelve you have fewer than 200 active subscribers, contracted revenue falls by two thirds and the new lease is still signed for five years. There sits the paradox to resolve before signing anything: membership does stabilize February's cash, yes, but it also hands you a false sense of floor that pushes you to commit fixed costs against revenue that can evaporate. The rule I apply: commit no new fixed costs against recurring revenue until you have two consecutive quarters with churn below 8%. Until then, the membership cash stays in the bank. The fee is not the revenue: incremental spend is. A coffee subscriber walking in three times a week buys pastry on 40% of those visits, and that is where the margin funding the whole program lives.

The differences that decide whether the model holds

Churn outranks acquisition. At 8% monthly churn, average tenure is twelve and a half months; at 15% it drops under seven. The gap between those two numbers decides whether acquisition cost is recovered or burned. Discounting has a recipe-card ceiling. No included benefit may push a dish past 32% food cost; under 60% contribution margin, every extra visit worsens the result instead of improving it. Membership rewrites the conversation with a restaurant investor. A business showing contracted recurring revenue, measured churn and an 8% forecast error gets valued on different multiples than one showing average ticket. A dark kitchen and a dining room cannot carry the same design. Without a floor there is no suggestive selling, so the subscription leans on cadence and logistics; with a floor, the included benefit is the excuse for your team to sell everything else. Where digital menus come up: QR is a complement and the PHYSICAL menu always stays. QR gives price updates, accessibility and subscriber analytics; the printed menu governs service pace, menu narrative and suggestive selling. Both, each with its own job.