
The uncomfortable truth about most restaurant loyalty programmes is that a large share of the rewards go to people who were already loyal. That is not a reason to skip loyalty, the data alone justifies it, but it is a reason to design the programme so that it buys incremental frequency rather than simply discounting your best customers.
| Structure | How it works | Best for | Weakness |
|---|---|---|---|
| Punch card | Buy N, get one free | Simple menus, high frequency, carryout-led shops | Rewards only the already-frequent; no data unless digital |
| Points | Earn per dollar, redeem at thresholds | Varied ticket sizes, delivery-led shops | Feels like currency; customers optimise against you |
| Tiers | Status unlocks benefits | Larger files, strong regulars, higher ticket | Complexity; lower tiers feel excluded |
| Surprise & delight | Unannounced rewards on behaviour | Any size; excellent brand effect | Unpredictable cost; hard to communicate as a benefit |
This choice quietly sets what the programme optimises for. Points per dollar rewards ticket size and favours large delivery orders. Points per visit rewards frequency and favours the weekly carryout customer.
For most pizza shops, frequency is the more valuable behaviour and the harder one to buy, which argues for visit-weighted earning or a blend: a base amount per visit plus a smaller amount per dollar. That structure makes a $22 Tuesday order feel worth placing, which is exactly the occasion you are trying to create.
The threshold should sit slightly above the customer’s natural rhythm: close enough to feel achievable, far enough to require an extra occasion.
If your median engaged customer orders eight times a year, a reward at ten visits asks for two additional occasions and remains plausible. A reward at twenty is invisible; a reward at six is a discount on behaviour you already had.
A reward threshold below your customers' existing frequency is not a loyalty programme. It is a price reduction with paperwork.
Even a loyalty programme with mediocre incremental lift usually pays for itself through identification. Enrolled customers are identified customers, and identification is what makes segmentation, journeys, win-backs and honest measurement possible at all.
Value the programme accordingly. If enrolment lifts your identified-ticket rate from 30% to 65%, the programme has just made every other marketing activity you run substantially more effective, and that benefit rarely appears in the loyalty report.
Your most frequent customers will hit thresholds fastest and take the most reward value. That is arithmetically inevitable and mostly fine, retention has value, but it means the programme by itself will not grow frequency at the top.
Give your best customers non-discount benefits instead: priority at peak, first access to new items, a genuinely better experience. Those cost less and are harder for a competitor to match with a coupon.
Digital, without much hesitation. A physical card produces no customer record, cannot be segmented, cannot trigger a journey and cannot be measured. The convenience of paper is not worth losing the identity of every customer who uses it.
Well-designed restaurant programmes typically see 40–70% of earned rewards redeemed. Very low redemption means the threshold is out of reach and the programme is not motivating anyone; very high redemption with flat frequency means you are simply discounting existing behaviour.
Points should expire on inactivity, commonly 6 to 12 months without an order, for two reasons: it caps a growing liability, and expiry warnings are one of the most effective win-back triggers you will ever run. Issued rewards should have a shorter, clearly stated window.
Decide explicitly and enforce it in the ordering system. The common workable rule is that points accrue on net spend after discounts, but reward redemptions do not combine with promotional offers on the same ticket.
Keep reading
Modelling reward cost per incremental visit, carrying the liability properly, and why breakage is a forecasting input rather than a profit centre.
Campaigns & offersWhy issuing value forward, redeemable on the next order rather than this one, protects margin today, buys a second occasion, and self-selects the customers worth paying for.
Audience & dataPractical ways to turn anonymous tickets into identified customers (online ordering, loyalty, QR on the box, the till, and wifi) without annoying anyone.
Campaigns & offersDesigning a coherent set of offers with fences, stacking rules, minimum spends and a maximum discount depth, so promotions stay a tool rather than becoming the price.
marketing.pizza runs all of it, every night, across every store you have.