The short answer, before the arithmetic
The ingredients in a free flat white cost about 4.05 SAR. The reward costs you about 10.81 SAR, because roughly half the people who redeem it were going to buy that coffee anyway. On an 85 SAR plan, the card has to cause about 25 extra paid drinks per 100 enrolled customers before it starts making money.
Here are the assumptions, so you can swap in your own. A 17 SAR flat white, one branch in Riyadh, a card of 6 paid drinks with the 7th free, and one barista on 4,500 SAR a month. Every figure below is built from those four numbers. They're assumptions, not measurements from your shop.
Two answers get quoted a lot and both are wrong. "It costs me 4 riyals" ignores the sale you just gave away. "It costs me 17 riyals" assumes every free drink replaces a full-price one, which is false for anyone the card actually brought back. The real number sits between them, and where it sits depends on one variable you have to estimate yourself.
One thing to say early: Waya has no POS integration and never sees your prices or your food cost, so it can't run this calculation for you. What it does supply is the left side of the equation, the visit and redemption counts.
Line one: beans, milk, and 105 seconds of a barista
Start with what physically leaves the building. 18 g of espresso beans at 90 SAR a kilo is 1.62 SAR. 180 ml of whole milk at 6 SAR a liter is 1.08 SAR. The 12 oz cup, lid, and sleeve run 1.20 SAR, and the napkin, stirrer, and sugar sachet add 0.15 SAR. Total cost of goods: 4.05 SAR.
Now labor. A barista at 4,500 SAR a month working 26 days of 9 hours costs 19.23 SAR an hour, or about 0.32 SAR a minute. Pulling the shot and steaming the milk takes 90 seconds, and scanning the customer's card and handing the cup over takes another 15. That's 105 seconds, or 0.56 SAR.
So one extra drink, made for someone who would not otherwise have been standing there, costs you 4.61 SAR. Call that Case A.
Be honest about the labor line. At 10:40 on a Tuesday your barista is paid whether she makes that drink or not, so those 56 halalas are already spent. At 8:15 during the morning rush the same 105 seconds are a paying customer who left the queue. Charging labor in full on every redemption keeps the model conservative, which is the right direction to be wrong in.
Line two: cannibalization, the number nobody puts in the spreadsheet
Now Case B: the customer was coming in anyway and would have bought that same flat white at 17 SAR. Compare the two worlds. In both of them the drink gets made, the milk gets used, and the barista spends 105 seconds. The only difference is that the 17 SAR never reaches your till.
That means a cannibalized redemption costs you the full 17 SAR, not the 12.95 SAR of gross margin. The cost of goods is identical in both worlds, so it cancels out. This is the single most common mistake in loyalty math, and it always understates the cost.
Blend the two cases. If half your redemptions are genuinely incremental and half are cannibalized, one free coffee costs 0.5 × 4.61 + 0.5 × 17.00 = 10.81 SAR. At 20% cannibalization it drops to 7.09 SAR. At 80% it climbs to 14.52 SAR. Nothing else in this article moves the answer that much.
You can estimate your own share without a POS. Take your last 30 redemptions and compare each customer's gap since their previous visit to their own normal gap. Someone who usually appears every 9 days and turned up on day 5 was probably pulled forward, so count that one as incremental. Someone who came on day 10 was probably coming regardless. It's rough, it takes about 20 minutes on the dashboard, and it beats guessing.
Line three: breakage, the credit that pays for most of it
Rewards are only expensive if people reach them. Take a cohort of 100 customers who enroll in one month. Suppose 100 collect stamp 1 on the enrollment visit, 55 come back for stamp 2, then 40, 32, 27, and finally 23 reach stamp 6. That's 277 paid drinks, or 4,709 SAR of coffee revenue.
Of the 23 who earn a reward, say 20 actually claim it. The 3 who never come back leave 51 SAR of promised value on the table. The bigger credit is quieter: 139 of those 277 stamps, more than half, belong to customers who never finished a card. That's 2,363 SAR of revenue with no reward cost attached to it at all.
Put the two sides together. 20 redemptions at 10.81 SAR is 216.20 SAR of reward cost against 4,709 SAR of revenue: an effective discount of 4.6%. The number printed on the card looks like 16.7%, because one free drink per six paid ones is 17 SAR off every 102 SAR. A stamp card is closer to a raffle most people quietly stop entering than to a standing discount.
Do not go looking for more breakage. Stretching the card from 6 stamps to 12 lowers your reward cost and lowers the number of people who bother playing, and a customer who quits at stamp 3 feeling stupid is worse than one who takes a free coffee. The step worth attacking is 100 down to 55, not 23 down to 20.
How many extra visits you actually need
Break-even is simple once the pieces exist: divide total program cost by the contribution of one extra paid drink. An incremental 17 SAR flat white contributes 12.39 SAR after 4.61 SAR of goods and labor. The cohort's reward cost is 216.20 SAR.
On Waya's free tier that cohort of 100 customers costs 0 SAR in software, so you need 18 extra paid drinks. On the 85 SAR Growth plan you need 25. On a 149 SAR tool you need 30. Twenty-five extra drinks across 100 enrolled customers is one additional visit from one customer in four, spread over however many months the cohort takes to work through the card.
Run the cannibalization range and the answer stays in a narrow band. At 20% cannibalization you need 19 extra drinks on an 85 SAR plan; at 80% you need 31. The subscription line is not what decides this. A 64 SAR difference in monthly software price moves break-even by about 5 drinks, while your cannibalization share moves it by 12.
For context on that software line, vendor-published Saudi pricing as of August 2026: Btaqa lists 108, 250, and 380 SAR a month, OneCup is around 150 SAR, and Watily is 149 SAR with loyalty bundled into a wider plan. Niqati is the best-known name in the market and prices by quote. Waya's Growth plan is 85 SAR a month, or 867 SAR a year. Check all of these yourself before you decide, because tiers change.
Four levers, ranked by how much they move the number
First, cap the reward. "Any drink free" invites the 28 SAR iced specialty with three syrups, and your Case A cost stops being 4.61 SAR. Write the reward as a specific item or put a ceiling on it, in the card wording your customer reads in Apple Wallet or Google Wallet.
Second, sell something alongside it. If 40% of redeemers add a 12 SAR pastry that costs you 5.40 SAR, that's 2.64 SAR of contribution per redemption on average, and the net cost of the free coffee falls from 10.81 SAR to 8.17 SAR. Break-even on the 85 SAR plan drops from 25 extra drinks to 20. A cashier saying "your coffee's on us today, want something with it" is worth more than any pricing change on this page.
Third, choose when rewards can be claimed. Redemptions that land in dead hours cost you materials and nothing else, because the barista is already paid and no queue is displaced. A quiet line on the card such as "reward drinks before 11am or after 8pm" pushes your mix toward Case A. It also annoys some people, so weigh that.
Fourth, and it barely counts as a lever on this spreadsheet, keep the stalled customers from disappearing. Every customer who stops at stamp 2 is revenue you never billed, not a saving. Waya's dashboard flags regulars who've gone quiet and lets you send a wallet message from the same screen, which is where the 55 in that funnel gets bigger.
What Waya can and cannot tell you here
Waya gives you the visit count, the redemption count, new versus returning customers, and a list of regulars who've stopped showing up. Cards live in Apple Wallet and Google Wallet with a web card for anything else, customers install no app, and staff stamp by scanning on an ordinary phone. That's the volume half of this calculation.
It cannot give you the money half. There is no POS integration, so Waya never sees a basket, a price, or a food cost, and it can't tell a 17 SAR flat white from a 28 SAR iced drink. There's no SMS either. The free tier stops at 100 customers and 100 wallet messages a month, though customers already enrolled keep their cards and keep collecting past that line.
So the honest division of labor is this: we count the visits, you own the cost sheet. Waya runs on 100+ shops across Saudi Arabia with 5,000+ cards sitting in wallets, and not one of those shops has handed us their invoice from the roaster.
Do this once a quarter with three inputs: your reward's cost of goods, your best guess at cannibalization, and last month's redemption count from the dashboard. Twenty minutes and a calculator will tell you whether the card is paying for itself, which is more than most loyalty programs ever get asked.
Frequently asked questions
What does a free coffee actually cost a cafe?
About 10.81 SAR on a 17 SAR drink, roughly two and a half times the 4.05 SAR of beans, milk, and cup. The gap is cannibalization: if the customer was coming in anyway, you lose the full 17 SAR, not just your margin, because the drink gets made either way. A genuinely incremental free coffee costs only 4.61 SAR including 105 seconds of barista time.
How many stamps should a loyalty card require?
Six to eight for something people buy two or more times a week, like coffee. Fewer than six and you pay out rewards to customers who were never at risk of leaving. More than ten and most people quit before the reward feels reachable, which lowers your cost but also lowers how many customers bother playing at all.
How do I know if my rewards are cannibalizing sales I'd have made anyway?
Compare each redeemer's gap since their last visit against their own normal gap. Take your last 30 redemptions from the dashboard: a customer who usually comes every 9 days but redeemed on day 5 was probably pulled forward, while one who came on day 10 was likely coming regardless. The share of the second group is your cannibalization rate, and it swings break-even more than any other number.
Is a stamp card cheaper than giving 10% off?
Usually much cheaper. In the worked cohort above, 10% off all 277 paid drinks costs 470.90 SAR, while the stamp card's 20 redeemed rewards cost 216.20 SAR. The discount also pays every walk-in who was never coming back, whereas the card only pays people who returned six times.
Does Waya charge a fee for each redeemed reward?
No. There's no commission on sales, no per-card fee, and no setup fee. The free plan is 0 SAR forever up to 100 customers and 100 wallet messages a month, Growth is 85 SAR a month, and Premium is 149 SAR a month. On paid plans the metered item is wallet messages, not customers or redemptions.