How Much Should a Loyalty Reward Cost You? Margin Maths for Small Shops
By Alva Chew, founder of Stridec · Updated
In short A loyalty reward should cost you about 1% to 5% of member spend once you count your cost price and the share of points that never get redeemed. Work out the extra sales you need to cover it with one formula, then prove it with a before-and-after check on members against regulars who did not join. If the numbers do not clear break-even after two months, lower the give-back before you change anything else.
How much should a loyalty reward cost you?
About 1% to 5% of what your members spend, measured at what the reward costs you and not at its menu price. That range is our own working guide. It is not a published benchmark. It leaves room for the margin on the extra sales a reward is meant to bring, and it keeps the program from eating the profit you wanted to protect.
The earlier guide on setting your points earn rate picks the give-back rate and the threshold. This page answers the next two questions. What does that reward really cost, and how do you prove the extra visits pay for it?
What is the formula for the real cost of a reward?
Three numbers set it:
Effective reward cost = give-back rate × cost share × redemption rate
- Give-back rate is the reward value divided by the spend needed to earn it. 100 points worth US$5 at 1 point per US$1 gives 5%.
- Cost share is what the reward costs you as a share of its value. A US$5 drink that costs US$1.25 to make is 25%. Money off the bill is 100%.
- Redemption rate is the share of points that end up spent. Use 100% when you set the reward, so you know the worst case.
All the figures in the tables below are our own illustrative arithmetic. Replace them with numbers from your supplier invoices and your points ledger.
| Reward | Give-back | Cost share | Effective cost at 100% redemption | At 60% redemption |
|---|---|---|---|---|
| Free drink (US$5 menu, US$1.25 cost) | 10% | 25% | 2.5% | 1.5% |
| Free dessert (US$10 menu, US$3 cost) | 5% | 30% | 1.5% | 0.9% |
| US$5 off the bill | 5% | 100% | 5.0% | 3.0% |
| US$10 off at 300 points | 3.3% | 100% | 3.3% | 2.0% |
Source: our arithmetic, illustrative US$ figures.
The same 5% headline give-back costs you 1.5% as a dessert and 5% as money off. If your margin is thin, that gap is the reason to reward with an item.
How much extra spend does a reward need to bring in?
Enough that the margin on the extra sales covers what the rewards cost. Take US$100 of monthly spend from a member before the program, and call the extra spend x. The extra sales earn your gross margin m. The reward is paid on all of the member’s spend, including the extra, at the effective cost r.
Break-even extra spend, as a share of old spend = 100 × r ÷ (m − r)
Here r and m are percentages of spend. Check it with the first row below: 100 × 5 ÷ (70 − 5) = 7.7%.
| Reward | Effective cost r | Gross margin m | Extra spend needed to break even |
|---|---|---|---|
| 5% money off | 5.0% | 70% | 7.7% |
| 5% money off | 5.0% | 30% | 20.0% |
| 3.3% money off | 3.3% | 30% | 12.5% |
| Free drink at 10% give-back | 2.5% | 70% | 3.7% |
| Free drink at 10% give-back | 2.5% | 50% | 5.3% |
| 5% money off, 60% redeemed | 3.0% | 70% | 4.5% |
Source: our arithmetic, illustrative figures.
Our guide on whether a loyalty program is worth it runs a simpler version of this test with the same 5% reward and a 70% margin and gets 7.1%. This version is a little higher because it also pays the reward on the extra spend, which is how the reward behaves in practice.
Gross margin here means sales minus the cost of goods, as a share of sales. Use the margin for what members typically buy. A café selling coffee and a clothing shop sit far apart on this line, and the right-hand column moves with it. At a 30% margin, a 5% money-off reward needs a 20% rise in member spend, which few programs deliver. The honest fix is a lower give-back or a cheaper reward item, not a hope that members will spend more.
What redemption rate should you plan on?
Plan on full redemption when you design the reward and expect less when you forecast. The research you can open for free shows a wide gap between earning and spending.
| Finding | Scope | Source |
|---|---|---|
| Customers redeemed in 3% of purchases, and 60% of customers never redeemed over 43 months | One Latin American home and hardware retailer, points earned at 1% of spend | Stourm, Bradlow and Fader, 2015 |
| Customers who redeemed at least once had baskets 25% larger and visited 141% more often than those who never redeemed | Same retailer | Stourm, Bradlow and Fader, 2015 |
Source: Stourm, Bradlow and Fader, Journal of Marketing Research, April 2015, read 20 November 2026.
Two cautions. That retailer paid only 1% back, a far thinner reward than the 3% to 10% in this guide, so your redemption could be much higher. And the second row compares different kinds of customer. It does not show that redeeming made them spend more. Your own redemption share is one division in your ledger: points redeemed to date over points issued, grouped by the month they were issued. Our guide to unredeemed points on the books shows how to build that cohort view.
If your redemption turns out lower than you planned, the program costs less than the worst case. Don’t use that as a reason to make rewards harder to reach. A reward that few customers can reach does not bring the extra visits you are paying for.
How do you prove the reward pays for itself?
Compare members with similar regulars who did not join, over the same weeks. A simple before-and-after on members alone will overstate the result, because your best customers join first and would have kept coming anyway. A study of one convenience store franchise by Liu (Journal of Marketing, 2007) found that light and moderate buyers bought more after joining, while heavy buyers claimed most rewards without changing how they shopped.
Run this check after eight weeks of the program:
- Pick the members. Take everyone who joined in the first two weeks and has at least one visit before joining.
- Pick a comparison group. Take regulars of similar spend who did not join. A paper tally or POS report of repeat customers works.
- Measure both groups twice. Average spend per person for the 8 weeks before launch and the 8 weeks after.
- Find the lift. Members’ change minus the comparison group’s change.
- Convert the lift to profit. Multiply the extra spend by your gross margin.
- Subtract the rewards redeemed at their cost to you.
Here is an illustration, using our own made-up figures for a café.
| Step | Members | Comparison regulars |
|---|---|---|
| Average spend in the 8 weeks before | US$150 | US$150 |
| Average spend in the 8 weeks after | US$165 | US$153 |
| Expected without the program | US$153 | |
| Lift per member | US$12 | |
| Gross profit on the lift at 70% | US$8.40 | |
| Points earned, 1 per US$1 | 165 | |
| Points redeemed at 50% | 82.5 | |
| Reward cost as US$5 off per 100 points | US$4.13 | |
| Reward cost as free drinks at 25% cost share | US$1.03 | |
| Net gain per member, money off | US$4.27 | |
| Net gain per member, free drinks | US$7.37 |
Source: our arithmetic, illustrative figures.
In this illustration both reward types pay, and the free drink pays more. If the lift had been US$5 instead of US$12, gross profit would be US$3.50, and the money-off version would lose US$0.63 per member while the free drink still gains about US$2.47. Keep the dated sheet for each check so you can compare the next one with it.
Eight weeks is a short window, and a small group will give a noisy answer. Treat a lift of a few dollars on a few dozen members as a hint and run the check again at 16 weeks.
How soon should a member reach the first reward?
Within a few weeks for a regular. A field experiment at a Columbia University café by Kivetz, Urminsky and Zheng (Journal of Marketing Research, 2006) found interpurchase times fell by about 20%, roughly 0.7 days, as members approached a free coffee. A first reward that sits months away loses that pull, and the extra visits the maths depends on never arrive.
| Finding | Scope | Source |
|---|---|---|
| Interpurchase time fell 20% (about 0.7 days) from the first to the last stamp on the card | Café reward programme at Columbia University | Kivetz, Urminsky and Zheng, 2006 |
Source: Kivetz, Urminsky and Zheng, “The Goal-Gradient Hypothesis Resurrected”, read 20 November 2026. The reward in the study was a stamp card, so use it as evidence for nearness to a reward and not for the stamp format.
What should you change if the numbers fail?
Change one thing at a time, in this order:
- Switch money off to an item that costs you less than its menu price. This cuts effective cost without touching how generous the reward looks.
- Raise the threshold a little, for example from 100 points to 120, and keep the earn rate at 1 point per US$1.
- Lower the give-back only after the first two. Regulars notice a cut, so give notice and honour points already earned.
- Check the margin you used. If the extra sales were mostly low-margin items, the lift is worth less than you assumed.
Points make each of these a single change. A stamp card or a tier scheme that fails the same test has several rules to adjust, and customers have to relearn each one.
How Gimify fits
Gimify is a points-only program where points are earned in proportion to each purchase, so the give-back you set is the only rate you have to cost. Staff add and redeem points on any phone at the counter, and the free plan covers one outlet with unlimited members.
Questions owners ask
How much should a loyalty reward cost me?
Aim for an effective cost of about 1% to 5% of what members spend. Effective cost means the give-back rate, multiplied by your cost price share if the reward is an item, and by the share of points you expect to be redeemed. Money off costs its full face value, so it needs a lower give-back.
How do I know if my loyalty program is making money?
Compare members with regulars who did not join over the same weeks, before and after joining. The extra spend from members, multiplied by your gross margin, should exceed what the redeemed rewards cost you. Count only the lift above the comparison group.
How much extra spend do I need to cover a reward?
Divide the effective reward cost by your gross margin less that cost, then multiply by 100. A 5% money-off reward at a 70% gross margin needs about 7.7% more spend from members. At a 30% margin it needs 20%.
Should I assume every member redeems their points?
Use full redemption as your worst case when you set the reward, then plan on less. In one retailer's programme studied by Stourm, Bradlow and Fader in 2015, 60% of customers never redeemed in 43 months. Your own ledger will tell you your number.
Is a free item or money off cheaper for me?
A free item usually costs less because you pay your cost price, not the menu price. Money off costs the full amount on every redemption. The trade is that money off is simpler to apply across a mixed menu.
When should I cut the reward?
If after about two months the members' lift over your comparison group does not cover the redeemed rewards, lower the give-back or move to a cheaper reward item. Give customers notice, because regulars who have been saving will notice a cut.
Read next
Start free
Gimify is free for one outlet with unlimited members and no time limit. Unlimited outlets are one flat monthly price.
Start free See plans