Eighty percent of restaurant loyalty programmes fail for the same reason: the mechanic is badly chosen. Not the gift, not the visual, not the tool: the mechanic itself. A well-placed classic stamp card brings back 30 to 45% of members; a "10% off forever" brings back 4%. That is ten times fewer, with a margin bled ten times harder. This article reviews the 7 mechanics that genuinely work in the restaurant business, and the 4 that quietly destroy your profitability.
Everything that follows is based on what I see happening with restaurateurs running Pépite Pass (stamp or points on Apple/Google Wallet, no app), as well as on public industry literature (Toast, Square, Paytronix, Bond Brand Loyalty). No figure is invented: when I give a range, it is a field observation, not a marketing promise.
1. Why the mechanic matters more than the gift
A customer does not come back to win €14 of pizza. They come back because they have an open account with your restaurant: 6 stamps out of 10, 320 points out of 500, a Silver status that could tip over into Gold. This psychological loop, called the "endowed progress effect" in the literature (Nunes & Drèze, 2006), is what makes people come back, not the content of the reward.
Practical consequence: a badly designed mechanic with a big gift works less well than a well-designed mechanic with a small gift. The proof:
- A restaurant offering 15% off for life after sign-up: observed 90-day return rate ≈ 4 to 7%.
- A restaurant with a 10-visit stamp card = 1 free dish (equivalent value across 10 tickets): 90-day return rate ≈ 32 to 41%.
- Same cost for the restaurateur. Five to ten times more returns generated.
The whole point of this article fits on one line: which mechanic turns a visit into an open account, instead of a simple rebate?
2. The 7 mechanics that genuinely work
1. The classic stamp card: the 10th meal free
Principle. The customer earns a stamp on every visit. On reaching 10, they unlock a fixed reward (often the most emblematic dish or item of the restaurant). The counter resets to zero and starts again.
Concrete example. Pizzeria in Lyon, average spend €18, digital stamp card on Apple/Google Wallet, reward at 10 stamps: 1 free pizza (food cost ≈ €2.80). Over 6 months: 412 sign-ups, 38% of customers on 3 stamps or more, 14% having reached the 10th. Frequency increase among members: +28% vs non-members.
Observed return rate. 30 to 45% of members come back at least 3 times within 90 days. It is the most universal mechanic, and the least risky. If you do not know where to start, start with this.
Why it works.Visual (boxes filling up), reachable (10 is psychologically short), addictive (each stamp is a mini-reward in itself). It is the "endowed progress" effect in its purest form.
2. Points per euro spent
Principle. €1 spent = 1 point (or 10 points, the scale does not matter). A reward is unlocked from a points threshold, often set at 5 to 7% of the cumulative value.
Concrete example. Parisian bistro, average spend €32, €1 = 1 point, reward at 200 points = house aperitif on the house (value €12, food cost €1.80). The customer pays €200 of meals (6 to 7 visits) to unlock their aperitif. Reach rate: 22% of members over 6 months.
Observed return rate. 25 to 38%. Slightly lower than the stamp card, but with a powerful side effect: average spend up by 8 to 15%. The customer adds the dessert or the glass of wine to scrape together a few more points.
When to choose it. Restaurants with a wide ticket range (dessert, wine, coffee = options). Pointless in fast food where the ticket is already optimised.
3. VIP tiers: bronze, silver, gold
Principle. Three status levels reached by accumulating visits or spend over a rolling 12 months. Each tier unlocks permanent privileges as long as the status is active: booking priority, favourite table, attention from the chef, off-menu tasting.
Concrete example. Fine-dining restaurant in Bordeaux. Bronze (3 visits/year): house amuse-bouche on arrival. Silver (6 visits/year): a bottle of still water on the house + priority slots. Gold (10+ visits): a personal word from the chef + a complimentary tasting once a year. In the programme's second year: 18 Gold customers, who generate 14% of total revenue.
Observed return rate. Hard to measure as a rate, but the real KPI here is the concentration of revenue: your top 5 to 10% of customers generate 25 to 35% of turnover after 18 months.
When to choose it. High average spend (€35 and up), a clientele that likes to feel recognised, front-of-house staff able to spot the customer on arrival (Pépite Pass shows the tier directly on the Wallet card).
4. The birthday reward
Principle. During the month (or the week) of the customer's birthday, a reward is automatically offered: dessert, aperitif, dish depending on your margin. Communicated by Wallet push or SMS.
Concrete example. Brunch spot in Lille. Birthday reward: 1 free brunch for 2 brunches paid. Push sent on the 1st of the birthday month. Over 1 year: 38% of members who filled in their date came back within the 30 days following the push. Average spend on the birthday visit: +22% (because they come as a group and order more).
Observed return rate. 30 to 45% of customers who left their birthday date. It is one of the best-ROI mechanics because it triggers a group visit.
Trap to avoid. Only ask for the month and the day (not the year), otherwise people lie or fill in nothing. And do not make the reward expire in 48h: 7 to 10 days is the right compromise between urgency and diary feasibility.
5. Customer referral
Principle. A loyal customer sends a personal link to a friend. When the friend signs up and makes their first visit, both receive a reward. The referrer must see their benefit immediately, otherwise they will not refer a second time.
Concrete example. Burger spot in Marseille. Referrer and friend each receive 2 bonus stamps (i.e. 20% of a head start on the next reward). Activation: 11% of members refer at least once over 12 months, with an average of 1.4 active referred friends per referrer. Cost of acquiring a new customer through referral: ≈ €4 of raw ingredients vs €8 to €15 on Meta ads.
Observed return rate. 8 to 15% of loyal customers refer. It stays modest, but the recruited friend has a higher retention rate than a customer found through ads (social recommendation effect).
Critical condition. The referral link must be shareable in 2 clicks from the Wallet card (WhatsApp / SMS / iMessage). If the referrer has to open a website and copy-paste a code, the rate drops to 1-2%.
6. Proximity pushes (geolocated)
Principle. You send a Wallet push to all customers who have installed the card. Apple Wallet and Google Wallet show a free notification. Immediate effect on footfall in the targeted time slot.
Concrete example.Italian restaurant in Toulouse, sluggish lunch service on a rainy Wednesday. Push at 11:15: "Today, house burrata free with any lunch menu". Result on the slot: +14 covers vs the average of a rainy Wednesday (measured over 8 weeks). Estimated ROI on the cost of the push (zero): infinite.
Observed return rate. Variable, but 3 to 8% of installed cards generate a visit within 48h of a well-calibrated push. On 500 installed cards: 15 to 40 extra visits, at €0 marginal cost.
Golden rule. A maximum of 2 pushes per month per card. Beyond that, people uninstall. And each push must give something specific (dish, offer, time slot), not a "see you soon" that interests nobody.
7. Waking dormant customers (at 60 days of absence)
Principle.A customer who has not come in for 60 days automatically receives a push "we missed you, here is a gift". The reward has to be strong enough to break the inertia: it is the only mechanic where overpaying the reward is profitable (you are paying to recover a lost customer, not to retain an active one).
Concrete example.Bakery/sandwich shop in Nantes. Push at 60 days: "We missed you, come and try our new dishes, your drink is on us". Over 6 months: 27% of dormant customers come back at least once, of whom 60% fall back into a normal frequency afterwards. Cost: €1.20 of drink per return vs a lost customer worth €240 of annual revenue.
Observed return rate. 20 to 35% of dormant customers reactivated. It is the mechanic with the best absolute ROI, often neglected because it fires quietly in the background, without glory.
To go further on this precise point, I put numbers on the impact of a lost customer in this article: how much a lost customer really costs in the restaurant business : the calculation is instructive.
A specific case to settle? Try Pépite Pass for free
3. The 4 mechanics that flop (and why)
Now the side nobody ever tells you about. These four mechanics are sold everywhere, burn through cash, and bring nobody back. If you have one in place, switch it off before reading on.
Flop no. 1: the permanent -10% (or -X%) for life
Principle. The customer signs up, presto, -10% on all their orders forever.
Why it flops. Three combined reasons:
- Dead loss from the very first visit, before the customer is loyal.
- The customer gets used to it. -10% becomes their normal price in 3 visits, they no longer perceive any added value.
- No return mechanism. Nothing to unlock, nothing to look forward to, nothing to talk about. It is a discount dressed up as a programme.
Observed return rate. 4 to 8%. Worse than doing nothing, because you also bleed the margin of customers who would have come back anyway.
Flop no. 2: points with no expiry and no tier
Principle. The customer accumulates points indefinitely. No expiry, no intermediate tier, just one big jar filling up.
Why it flops. Without a visible and reachable tier, the brain does not perceive progress. Without expiry, there is no urgency. The customer forgets they have a card. And you accumulate an invisible accounting liability that you will have to honour sooner or later.
Observed return rate. 6 to 12%. Drops off quickly as the points base grows.
Flop no. 3: overly complicated mechanics
Principle."Earn 2 points per euro Monday to Thursday, 1 point at the weekend, a x3 multiplier if you come as a pair between 2pm and 4pm, points doubled on desserts during the sales, and a status required to access the pre-sales…"
Why it flops. If I have to reach for a calculator, I have already lost. The universal rule of loyalty: one sentence must be enough to explain everything. "10 visits = 1 free pizza": everyone understands in 1 second. That is what you need.
Observed return rate. 8 to 14%. People sign up then give up within 30 days.
Flop no. 4: voucher rewards
Principle.Instead of offering a specific dish, you offer "€10 off your next order".
Why it flops.
- A voucher gets compared to money. It gets rationalised ("it is just €10, the place next door is just as good"). A free dish gets compared to a pleasure, that cannot be rationalised.
- The voucher becomes an accounting liability for you. The in-kind reward has a raw-ingredient cost far below its perceived value.
- Zero emotional effect. Nobody tells their friends they have a €10 voucher at your place. Everybody tells them they won a free brunch.
Observed return rate. 10 to 16%. Better than the other flops, but clearly below in-kind mechanics.
4. How to choose THE mechanic for your restaurant
There is no universal right answer. The right programme depends on the average spend, the target frequency, and the positioning. Here is the matrix I use with Pépite Pass restaurateurs when we set the mechanic together.
| Type of restaurant | Average spend | Primary mechanic | Secondary mechanic |
|---|---|---|---|
| Kebab / pizza / burger | €10 - 18 | Stamp (10 visits) | Dormant wake-up 60d |
| Bakery / coffee shop | €4 - 9 | Stamp (10 visits) | Proximity push |
| Brunch / brasserie | €18 - 28 | Stamp (8 to 10 visits) | Birthday |
| Bistro / market cuisine | €25 - 40 | Points per euro | Birthday |
| Fine-dining / Michelin-starred | €50 and up | VIP tiers | Birthday |
| Food truck / festival | €8 - 14 | Stamp (5 to 7 visits) | Proximity push |
This matrix is a starting point, not a truth set in stone. I lay out the full logic in the digital restaurant loyalty card guide 2026.
5. How much the reward should be worth: the 5 to 8% rule
The question that comes up at every demo: "Léo, what do I give at the 10th visit?" Answer in one rule: 5 to 8% of the cumulative revenue to reach the reward.
The calculation fits on a napkin:
- Average spend × number of visits for the reward = cumulative revenue for 1 reward.
- Perceived value of the reward ≈ 5 to 8% of that cumulative revenue.
- Raw-ingredient cost (your real cash outflow) ≈ 2 to 3% of the cumulative revenue.
Concrete example: pizzeria at an €18 ticket, 10 stamps → €180 of cumulative revenue per reward. Ideal reward between €9 and €14 in menu value (5 to 8%). A margherita pizza at €11 with a food cost of €2.80 fits perfectly: you give 6% of perceived value, and you spend 1.5% of real cash.
What you must NOT do: below 4%, customers find it stingy and stop. Above 12%, you bleed your margin AND you attract voucher hunters who are not your core target.
6. Should you run several mechanics at once?
Yes, but with an iron rule: a maximum of three, and each must act on a different moment of the customer journey. Stacking stamp + points + cashback + tiers + referral is a guarantee that nobody understands a thing.
The combination I recommend to 80% of restaurants:
- Classic stamp card as the main frequency engine (visible, addictive, measurable).
- Birthday as the annual emotional engine (generates a group visit).
- Dormant wake-up at 60d as a safety net (recovers 25-30% of the lost ones).
Three mechanics, three moments of the journey: recurrence, emotion, recovery. The customer only has to understand the stamp card, the other two fire on their own in the background. It is the simplest combination to run for the restaurateur and the clearest for the customer.
Premium / fine-dining restaurants: replace "stamp card" with "VIP tiers" and the rest holds up.
7. How to measure whether it works: the 3 KPIs that matter
You can spend 6 months iterating on the mechanic without knowing whether it works. Here are the only three indicators to watch.
KPI 1: the 90-day return rate
Out of 100 signed-up customers, how many came back at least once within 90 days?
- Below 20%: mechanic too weak, we rework it.
- 20 - 35%: a decent standard.
- 35 - 50%: very good, we keep it.
- Above 50%: you have struck gold, we scale it up.
KPI 2: the sign-ups to visitors ratio
Out of 100 people who go through the till, how many sign up for the card? Below 15%, it is a pitch or sign-up friction problem, not a mechanic problem. Above 30%, it is healthy. A Wallet card that opens via QR or through an NFC link at the till regularly reaches 35 to 45% take-up, because there is no app to install.
KPI 3: average spend of members vs non-members
Do your members spend more than your non-members? On a points/euros mechanic, the expected gap is +8 to +15%. On pure stamp, it is more like 0 to +5% (the stamp acts on frequency, not on the ticket). If you want a bigger average spend, go with points.
These three KPIs are what I personally track with each Pépite Pass restaurateur at the end of the first quarter. Everything else (number of pushes sent, open rate, etc.) is secondary.
8. The beginner mistakes to avoid
I have seen dozens of badly launched programmes. Here are the systematic mistakes.
- Launching without a till pitch.If the front-of-house team does not offer the card at every payment, you will get 15 sign-ups a month. With a well-rehearsed sentence ("want your loyalty card? it is just a QR to scan, no app, 5 seconds"), you climb to 200 a month.
- Changing the mechanic every 2 months. A programme is assessed over a minimum of 6 months. Before that, you do not have the data.
- Setting a reward expiry that is too short. 48h is punitive. 7 days is urgent without being frustrating.
- Forgetting the reminder push before expiry.A push "you have 3 visits left before your 10th free meal disappears" at 90 days of inactivity brings back 25-35% of dormant customers: it is free and it is massive.
- Not distinguishing newcomers from loyal customers. 2 stamps given at sign-up looks trivial, it doubles your take-up rate (the customer is immediately at 2/10 and feels an investment on your part).
- Wanting to measure everything before launching. Launch, measure the 3 KPIs above, iterate. Perfection is the enemy of getting started.
On pushes specifically, I have a dedicated article that lays out the best practices: Wallet push notifications for restaurants : when to send, to whom, with what message.
Conclusion: the mechanic is 70% of the result
If you remember one thing: the mechanic matters more than the reward. A classic stamp card with a modest reward crushes a 10%-off-for-life with a big discount, because the psychological loop of progress is what makes people come back, not the content of the gift.
Eighty percent of restaurateurs would be better off launching with the "stamp + birthday + dormant wake-up" combination: simple, measurable, robust. The remaining 20% (fine-dining, premium, wide ticket) are better off moving to VIP tiers and points per euro.
Whatever you choose, keep it simple. One sentence must be enough to explain everything. If you are unsure about the calibration, I am reachable directly on WhatsApp. I built Pépite Pass for this, literally. And if you want to test it at your place, it is a free trial, no card, no commitment. You have time to see whether it catches on before paying anything at all.
To go further on the overall retention strategy (not just the mechanic): how to build restaurant customer loyalty in 2026. And if you want to put numbers on what a single lost customer really costs you: how much a lost customer costs in the restaurant business.
