How much does a loyalty program cost — and does it pay off? The full ROI math

The question "how much does a loyalty program cost" has three very different answers depending on the path you take: paper punch cards for a few dollars, a custom app for $10,000–30,000, or a card in Apple and Google Wallet on a SaaS plan for a few dollars a month. But the real question isn't "what does it cost" — it's "what does it earn". In this article we break the cost of a loyalty program down to its components and calculate ROI on hard data, so you know exactly when the program starts returning money instead of just consuming it.

Three paths, three loyalty program costs

Before we calculate the return, you need to understand what you're actually paying for. The cost of a loyalty program isn't just the price of the tool — it's the sum of the upfront spend, the running cost, and the hidden cost (your time, printing, SMS fees). Here's how that looks across the three most common options for a small business.

• Paper / DIY: card printing $50–150 a year, a stamp for $10, plus a few hours of your time each month. The apparent cost is low, but you get zero data, zero communication, and huge waste (lost cards, stamp fraud).
• Custom app: $10,000–30,000 for the build, plus $400–1,000 a month for maintenance, hosting and fixes. Add 2–4 months of waiting and the risk that customers simply won't download it.
• Wallet-based SaaS (e.g. Loyalif): from ~$25 to ~$75 a month, zero developer setup, card ready in an hour, with push, automations and analytics included. No printing cost, no SMS fees.
• The hidden cost of every path: no push. With paper and most cheap apps you can't reach a customer who stopped coming in — and that's exactly where the biggest money sits.

How to calculate loyalty program ROI

Loyalty program ROI has two profit sources worth counting separately: recovered customers (win-back) and increased visit frequency among existing regulars. Take a café with an average basket of $6 and 600 active customers in the base. Here's where the return actually comes from.

1. 1Win-back: an automatic push after 21 days of absence recovers ~12% of customers who otherwise wouldn't return. Out of 600 people that's ~30 at risk per month, of whom ~9 come back. At 3 extra visits of $6 = ~$160 of recovered revenue per month.
2. 2Visit frequency: a "close-to-reward" push plus a birthday push lift average frequency by ~0.4 visits per customer per month. Even with 250 engaged customers that's 100 extra visits × $6 = ~$600 a month.
3. 3Bigger basket: a customer who knows they're "2 stamps short" adds on average +8% to the bill. Across 350 visits a month that's another ~$160.
4. 4Referrals: a referral program adds ~5–8 new customers a month with no ad spend — pure upside, because acquisition cost drops to zero.

The return in numbers: when the program breaks even

Adding just the first three lines, our sample café recovers and adds ~$920 of revenue a month. Even at a cautious 65% margin that's ~$600 of profit. Against a SaaS cost of ~$50 a month, ROI exceeds 1,100% — and the program pays for itself in the first week of the month. By contrast a custom app at $20,000 with the same profit only breaks even after ~33 months, assuming nothing breaks.

Why the cheapest option is often the most expensive

Paper looks free, but it enables none of the four revenue levers above. No push means no win-back, no data means no birthdays, no automation means no "close-to-reward" reminders. So you don't pay with the tool's price — you pay with lost revenue, and because that loss never shows up in the books, it's the hardest to notice.

“Don't ask how much a loyalty program costs. Ask how much not having one costs — and count it for every month you go without.”

If you want to test these numbers on your own base, with Loyalif you can launch a card in Apple and Google Wallet in an hour — no app, no developer costs, with push, win-back, birthdays and ROI analytics included. Calculate your return before you pay for anything: in most venues the program breaks even within the first month.