What a repeat purchase program really costs: 2026 price tiers and the invoices nobody sends you

What does a repeat-purchase program cost per month in 2026?

As of August 2026, a properly built repeat-purchase program runs between 85 and 640 USD per month per location, and that spread covers the three layers almost nobody adds up together:

software license, redeemed incentive, and operating hours. The 180-seat grill house in Guadalajara that sorted out my thinking on this paid 149 USD for the platform and believed that was the cost; the 340 desserts redeemed each month at 2.80 USD of raw material added 952 USD, and the eleven weekly hours the floor manager spent enrolling guests at the register had no budget line either. The license was 13% of real spend. When you budget only the monthly invoice, you budget the cheapest and most replaceable piece of the system, and you leave out the 87% that actually decides whether the program adds to margin or eats it. Three tiers with different logics govern the 2026 market, and the tier matters more than the brand.

What each price tier includes?

Between 85 and 180 USD a month you get digital stamps, phone-number lookup at the register, and an exportable database: enough for a single location under 90 seats chasing frequency, nothing else.

From 180 to 380 USD, behavioral segmentation arrives, along with automated lapsed-guest campaigns and POS integration, which is the line where the program stops being a digital stamp card and starts moving reservations. Between 380 and 640 USD you find multi-unit attribution, the link between dining room and owned delivery, and APIs to match the check against the app. That top tier only pays off with more than three units, or with owned delivery weighing above 25% of sales. Five variables drive the final number, and none of them appears on the vendor's pricing page. Location count adds somewhere between 15% and 40%, since almost every platform charges per site rather than per group.

The five factors that move the real price

Active member volume comes second: going from 2,000 to 10,000 records usually jumps a tier even when no feature changes. POS integration tacks on 40 to 120 USD monthly whenever the system is closed. The messaging channel — WhatsApp or transactional SMS — bills per send and can double the license at a location pushing four campaigns a month. And the fifth, the priciest and the least visible, is the labor cost of enrolling: eleven weekly hours at a floor manager's wage run about 380 USD that nobody books as repeat-purchase spend. A repeat-purchase program is not fixed marketing spend, it is variable cost, and filing it alongside the signage and the menu photos is the structural mistake that eats the most margin. Signage costs the same with a full room or an empty one; the redeemed incentive comes off the plate, not out of the bank, and it grows at exactly the pace of the visits you wanted to trigger.

A budget that breathes versus a frozen one

There sits the paradox to resolve before signing: a successful program is more expensive than a mediocre one, and if you budgeted it as a fixed line, success will look like a variance. The right move is capping it in sales points — somewhere between 1.2% and 2.4% of channel sales — rather than in dollars, so the line breathes with the check average and the season. You do not define the incentive in the marketing meeting; you define it by crossing real food cost per item against the contribution margin that item drags along when it lands on a full check. A 15% discount on the bill punishes today's entire margin and changes nothing about tomorrow's frequency; a dessert at 22% food cost given away on the fifth visit buys the next visit at a known cost, and that visit brings a check that is rarely just the dessert.

The incentive is decided on the menu engineering sheet

At Masterestaurant we handle it this way from the first sheet: the incentive is picked for its food cost, its pull-through, and what it does to the kitchen at peak, never for how well it reads on a poster. The hard rule stands here too — no program item goes above 32% food cost, which is a ceiling, not a target. The program pays off once the contribution margin of the incremental visit clears the cost per recovered visit, and below that point you are buying expensive visits from guests who were coming back anyway. With a 24 USD average check and 65% contribution margin, each incremental visit leaves 15.60 USD; if total program spend divided by attributed visits lands at 4.10 USD, you have a business. One industry figure is no longer arguable: 39% of US restaurant visits come from loyalty program members, double the 2019 share, according to Restroworks (2025).

The exact point where the program pays for itself

What would happen if you switched the program off for a quarter? The redemptions would stop first, which reads as accounting relief, and sixty days later the frequency of everyone already enrolled would go with it — the asset that took two years to build. Negotiate three things in this order and leave list price for last: term, member tier, and messaging cost. Ask for a twelve-month contract with an exit at six, and a member tier carrying 40% of headroom over your current base, because the tier jump is what blows up the invoice in year two. WhatsApp sends get negotiated on annual volume rather than monthly, and there is usually 20% to 30% of air in there. The big lever, though, is not with the vendor: move enrollment from the cashier to a table QR and those eleven weekly hours drop to two, roughly 310 USD a month recovered without touching the contract.

How to negotiate the license and cut operating cost?

Measure cost per recovered visit every sixty days and renegotiate holding that number, not the feeling that the program is going well.

Before signing, demand in writing that you own the guest data and can export it in full to CSV at no cost, because the day you switch vendors that database is the one thing money cannot rebuild. Check as well whether the platform bills per registered member or per member active in the last ninety days: at a location with 8,000 records and 1,900 actives, that difference is a whole tier. And ask who pays for integration work when the POS releases a new version, since that invoice always shows up and never appears in the proposal. On Monday, before booking one more demo, open the register and calculate three numbers: license spend, last month's redemptions valued at raw material cost, and staff hours spent enrolling.

What to check before signing and what to do Monday?

If the sum passes 2.4% of sales, the vendor is not your problem. A badly built budget treats the repeat purchase program as fixed marketing spend and files it beside the signage and the menu photos;

a well built one treats it as a variable cost that breathes with the ticket, because every redemption comes out of the plate rather than out of the bank. In the mistaken version the incentive gets decided in the marketing meeting; under the Masterestaurant method it gets decided on the menu engineering sheet, crossing real per-product food cost against the contribution margin that product drags along when it lands inside a full check. The costliest difference is a matter of timing. A percentage discount punishes margin today and changes nothing about frequency tomorrow, whereas a low-food-cost product buys the next visit at a known price, and that visit brings an entire check behind it.

Where the two roads genuinely split?

Owners who price this well accept paying more for software when that software joins dining room and delivery under one guest ID, because guest lifetime value split across two databases is worth less than the same figure consolidated, even when the gross number looks identical.

And there is an asymmetry almost nobody weighs: getting it wrong on the expensive side means three months of a pricey licence, roughly 900 USD; getting it wrong on the cheap side means two years of a program quietly giving margin away, somewhere between 14,000 and 22,000 USD in a mid-sized location.