Short answer: choose a reward customers want, with a predictable cost and a simple service flow. Divide the reward's actual unit cost by the qualifying revenue required to earn it. The result is the program's direct effective cost before behavior changes.
A reward may look inexpensive because the item costs little to make, yet become expensive because customers earn it too quickly. It may protect margin and still fail if nobody wants it. Design has to balance perceived value, actual cost, and distance to the benefit.
This guide provides a reproducible method rather than a universal target. Margins, purchase frequency, and spare capacity differ between businesses.
The basic reward formula
Start with two values:
- Reward unit cost (RC): the incremental cost of delivering the benefit.
- Qualifying revenue (QR): the amount a customer must spend to reach it.
The formula is:
direct effective cost (%) = RC ÷ QR × 100
Example: a café requires 9 purchases of $6 and offers a drink that costs $1.60 to make and serve.
- Qualifying revenue: 9 × $6 = $54
- Reward cost: $1.60
- Direct effective cost: $1.60 ÷ $54 = 2.96%
That 2.96% is not the program’s “profit or loss.” It is only the direct reward cost relative to required revenue. Existing discounts, tax, fees, abuse, added labor, and purchases that would have happened anyway still matter.
Use cost, not selling price
If the free drink sells for $6, the reward does not necessarily cost the business $6. Incremental cost may include:
- ingredients or merchandise;
- packaging and disposables;
- variable payment fees, when applicable;
- commission or labor that grows with service;
- expected waste;
- delivery absorbed by the business.
Fixed costs such as rent matter to the company but do not necessarily rise when one extra unit is delivered. Separate incremental cost from accounting allocation to avoid false precision.
Perceived value and actual cost differ
Strong rewards often combine high perceived value with controlled incremental cost. This is why a business’s own product may work better than cash off.
Compare:
- $5 off: $5 of perceived value and a cost close to $5.
- A drink sold for $6 and produced for $1.60: perceived value near the known price, at a lower cost.
- An upgrade during unused capacity: potentially meaningful value with little added cost.
This is not a license to inflate prices or hide restrictions. The benefit must be real, available, and described clearly.
Three worked examples
Café: free item after 9 purchases
| Input | Value |
|---|---|
| Minimum purchase for a stamp | $6 |
| Required stamps | 9 |
| Qualifying revenue | $54 |
| Cost of free drink | $1.60 |
| Direct effective cost | 2.96% |
Decision question: does the reward create a visit that would not otherwise happen, or simply pay a customer who already returns? Compare participants’ visit intervals before and after, not just completed cards.
Salon: benefit after 5 services
A salon charges $45 per service and offers a treatment with $7 of supplies plus 15 added minutes.
- Qualifying revenue: 5 × $45 = $225
- Supplies: $7
- Added-time cost: estimate from compensation and capacity
When the schedule is full, 15 minutes has an opportunity cost. During unused time, the cost may be lower. The same reward has different economics by time of day, so a quiet-period condition can be legitimate when disclosed from the beginning.
Restaurant: stamp with a minimum order
A restaurant gives one stamp for orders of at least $25. After 6 stamps, it offers a dessert that costs $5.
- Minimum qualifying revenue: 6 × $25 = $150
- Maximum direct effective cost at the threshold: $5 ÷ $150 = 3.33%
Orders above $25 also qualify, so observed average revenue may be higher and the effective percentage lower. Update the calculation with participant revenue after the pilot.
How many stamps should you require?
There is no magic number. The goal must be reachable within a horizon customers can imagine.
Use this sequence:
- estimate natural purchase frequency;
- choose when a participant should feel the first win;
- divide that period by the expected interval between visits;
- test whether the resulting cost fits the margin.
If a customer sees a barber every 30 days, ten stamps represent almost a year. At a café visited twice a week, the same ten stamps may take five weeks. An identical count creates entirely different experiences.
An early milestone followed by a main reward is an option, but every added tier increases explanation. Keep only the complexity that produces useful behavior.
Reward by visit, item, spend, or tier?
| Model | Use when | Main risk |
|---|---|---|
| Per visit | Transaction values are similar | Rewarding very low-value visits |
| Per item | A clear product recurs | Ignoring additional purchases |
| Per spend | Order values vary widely | The rule feels abstract |
| By tier | The relationship is long and measurable | Complexity and a distant benefit |
A hybrid rule can require a minimum spend for every visit. Avoid stacking so many conditions that staff need to negotiate exceptions at checkout.
Set guardrails before the pilot
Record four limits:
- cost ceiling: how much the program may consume per $100 of qualifying revenue;
- earning window: expected time to the first reward;
- service time: acceptable additional seconds per transaction;
- misuse handling: returns, duplicates, and manual stamps.
These limits let you stop or adjust a pilot before a badly calibrated promotion becomes permanent.
How to measure whether the reward worked
Cards issued and stamps delivered measure the flow, not the outcome. Track:
- activation rate: participants receiving a first stamp;
- completion rate: participants reaching the reward;
- time to completion: exposes a target that is too remote or too easy;
- visit interval: the primary signal for frequency programs;
- qualifying revenue per participant: the real base for effective cost;
- cost of rewards redeemed: benefits delivered, not merely promised;
- post-redemption return rate: whether the relationship continues.
Compare equivalent periods and, when possible, a similar group that did not participate. Seasonality, price changes, and concurrent campaigns can imitate program improvement.
A 30-day pilot
- Record a baseline for the prior four weeks.
- Choose one reward and one rule.
- Limit the pilot to one location, shift, or customer group.
- Observe the first enrollments in person.
- Review adoption and operations weekly without changing the rule every day.
- At the end, calculate effective cost from actual revenue and redemptions.
- Keep it, adjust one variable, or end it with clear communication.
The right reward is neither the most generous nor the cheapest. Customers understand and want it, staff can deliver it, and its cost remains acceptable as participation grows.
