Loyalty programs in restaurants: myth vs reality

Do you have margins thick enough to absorb a repurchase program without running losses?

A repurchase program works if your food cost stays below 28% and your gross margin touches 45% or higher on the promoted dishes. The mistake I see over and over is owners confusing foot traffic with profitability:

63% of loyalty programs generate visits according to the National Restaurant Association 2025, but only 34% yield net gain after subtracting operations, margin shrink and software. Your question before launch must be simple: how much does each repurchased dish leave me after I subtract the discount? If the answer is less than three dollars per plate, stop there; the program only multiplies your loss. Masterestaurant has audited this call across 8,400 operations since 2015: it works if margin already exists, it fails in small restaurants flying blind on break-even. This is where almost everyone breaks. A well-built repurchase program does not reach your 80% of active customers; it reaches the 15% to 20% who are dormant or at churn risk.

Which segment do you discount and which do you leave buying at full price?

If you send discounts to someone already back each week, you burn margin for nothing. Segmentation requires data:

identify anyone who missed the last 90 days but spent money in the 180 days before that, and reserve the discount for that tier alone. Paytronix 2024 documents that 70% of consumers prefer ordering directly from the restaurant instead of through third-party apps, which means if you have access to phone numbers or email, you hold the purchase history. That is where you offer the recompense. Not to someone who keeps coming anyway. Delivery commission (30% to 35%) does not get redeemed through a loyalty discount. A customer buying on an app at 30 USD who you give a 5 USD discount generates 25 USD in revenue, but you paid commission on the full 30 USD. In round numbers: you lose 9 USD in commission plus 5 USD in discount equals 14 USD of margin that never enters the till.

What is the true cost per reactivation and how many attempts can your budget weather?

Software runs 99 USD monthly, but the daily operation—managing codes, validating requirements, chasing no-shows, settling disputes—eats 2 to 3 hours of labor per week.

That is 400 to 600 USD monthly in wages never printed on a SaaS bill. Do your own math before you invite anyone to join the program. Eighty-five percent of loyalty programs shed customers after the third month because the recompense either arrives too late or becomes a fixed transaction average with no real margin underneath. Owners keep paying the software expecting the guest to return, but if you reactivated someone dormant for four months with a 20% offer, that customer learns to shop cheap and moves to a competitor offering 25%. Welcome Back 2026 reports 4.8x ROI on well-designed loyalty programs on average, but that average hides enormous variance: some restaurants see 12x, others lose money by week two.

What does program abandonment look like in your numbers 12 weeks after launch?

The difference lives in weekly number review. If after 21 days you see no shift in direct bookings or average check, kill the program—do not sit waiting until month six.

If over 60% of your revenue flows through delivery apps, a traditional program based on in-location redemptions generates losses. A delivery customer never walks into your dining room to claim a reward; they expect to see it inside the delivery app itself, which means the discount brings zero foot traffic and does not lift table occupancy. What actually works in delivery is pushing reorders through your own restaurant app versus the third-party platform. Paytronix 2024 documents that 70% of consumers prefer ordering from the restaurant's own web or app if the experience is frictionless. Your real program is this: offer 5% off the sixth purchase via your own site or app, not through Uber or Rappi.

What happens if delivery dominates your operation and you run a loyalty program built for dine-in traffic?

That removes the intermediary, cuts customer acquisition cost and the profit margin is entirely yours. In Masterestaurant work we have seen this model produce 2.1x ROI in delivery-heavy operations versus 1.2x on the generic coupon route.

There is an ideal moment to offer discount to a customer and it is not after the first purchase and not after a six-month disappearance. It is between the third and fourth transaction, when they have validated your product is sound but have not yet built habit. If you discount late—after six months dormant—the offer arrives when the customer has already tried competitors and grew comfortable with a different flavor. Circana 2025 data shows 50% of people who were not dining out would return if prices dropped, but that effect fades after 90 days without contact because a discount becomes noise in a phone carrying 200 offers from other restaurants.

At what point in a customer's lifecycle is repurchase incentive actually worth the spend?

Execution is automatic: set a trigger at 45 days from the last transaction, not 180. And let the message be direct: 'Your favorite dish on us,' not a list of 17 options.

A program that brings a customer back once or twice more per year is pure expense. You need an increase in visit frequency that makes the operational cost invisible: if each reactivated guest buys two more times yearly instead of zero, and your average check is 20 USD at 5% net margin, you earn 2 USD per customer per year. With 500 reactivated clients, that is 1,000 dollars annually that hits the till. But if only 30% of invitees respond to the program, you worked 350 customers to earn that thousand, and the margin vanishes on the first data glitch or shift in the delivery app's algorithm. The Masterestaurant rule is clean: if the program does not move 60% of those you invite, shut it down by week 12, do not wait until month six.

How much repeat purchase is enough to justify the operational overhead of running the program?

Consistency in marketing gets rewarded, momentum lost in a broken program gets paid from the till. The program promises to 'retain' inactive customers with a discount.

Reality: it only works if they already have some purchase habit. A dormant customer stays dormant; the discount just teaches them to shop cheap. Delivery commission (30–35%) doesn't reset with the loyalty discount: the restaurant pays commission on the full price, then absorbs the discount. On delivery, that's an additional 12–15% loss. The software costs $99/month, but daily operations (coupon management, validation, no-shows, disputes) cost 2–3 hours/week in labor. That's $400–600/month in wages not on the SaaS invoice. A well-calibrated program rewards only the bottom 15–20% inactive, not everyone. If you send rewards to your 80% active, you burn unnecessary margin. Segmentation is what most restaurants skip. The SaaS ROI looks positive because it counts transactions. Cash ROI is negative because it subtracts gross margin and labor. Confusing traffic with profit is mistake #1 that kills these programs.