Loyalty program examples: compare three reward types

A customer loyalty program offers a reason to return, but it does not automatically create additional purchases. A discount can reward customers who would have bought at full price anyway. Before launching points or a club, compare three reward models and measure the complete cost of each.

Define the behavior you want to change

Choose an action the program might influence: a second order, a return after a long gap, purchase of a complementary category or a genuine referral. “Increase loyalty” is too vague. Record current repeat-purchase rate, typical time between orders, margin and known reasons customers leave.

Segment cautiously. A customer who buys monthly has a different need from a seasonal buyer. Giving the first group a discount may reduce margin without changing behavior. The second group may need a relevant reminder or better product availability.

Check the causes of non-repeat purchases: quality, delivery, service, price, assortment or simply infrequent need. A reward scheme cannot repair a disappointing product. The Small Business Administration's business guidance places marketing decisions within broader operating decisions; measure the commercial outcome, not merely sign-ups.

Example 1: points earned on purchases

Buyers earn points and exchange them for a benefit. This works best when purchases recur often enough for the reward to feel reachable. Define earning rate, redemption threshold, any expiry, excluded products and how returns change the balance.

Test a real order, partial refund and exchange. Are points reversed when an order is cancelled? Can the customer understand the rules before buying? An opaque points balance may create more support work than goodwill.

Estimate the likely cost of redeemed points as well as the software and administration costs. Not every point will be spent, but an arbitrary low redemption assumption should not be used to make the scheme appear profitable.

Example 2: a benefit after a threshold

The store offers a gift, delivery benefit or discount after a number of orders or a cumulative spend. The rule is easy to understand if the threshold is attainable. It can also encourage buyers to split orders or add a low-margin item just to reach it.

Fictional example: after four orders, a customer receives free delivery. If the delivery cost exceeds the margin on the next basket, the reward can destroy value. Test costs against actual order types and state conditions clearly.

A threshold can unlock service instead of a price cut: early access to restocks, help choosing a size or priority repair. Promise such benefits only when the team can deliver them consistently.

Example 3: a club with ongoing benefits

Membership offers a benefit while the buyer remains enrolled: content, support, delivery or another service. A paid tier needs enough continuing value to justify its price. A free tier still requires staff, technology and a clear reason to join.

Describe benefits, limits, renewal, cancellation and data use in plain language. Do not make non-members' normal terms artificially poor merely to make membership look attractive. Test cancellation and the messages sent to the customer.

This model fits a recurring need or real ongoing service better than a one-off purchase. If the offer ends after a single download, membership may add complexity without value.

Compare the models using margin

Use a comparable group or period to estimate additional purchases plausibly linked to the program. Deduct reward cost, software, support and communication costs. Compare with similar non-members where possible. Members may already be the most loyal customers; their higher frequency does not prove the program caused it.

Run a limited pilot

Start with one segment and a defined period. Document the rules, then test joining, earning the benefit, returning an item and leaving. Measure orders, margin, complaints and repeat purchase, and ask a few customers whether the benefit makes sense.

If participants buy at the same pace and merely claim rewards, revise or stop the test. If repeat orders improve but support is overloaded, simplify the rules. Keep a dated change log so different versions are not compared as one experiment.

The best retention improvement may instead be reliable stock, an accurate product and prompt support. Choose a loyalty scheme only when its mechanism solves a real problem and the cost fits the margin.

A simple cost calculation

Imagine a fictional store with 100 participating customers over a month. Members place 120 orders, compared with an expected 100 orders from a similar group without the scheme. Each incremental order contributes $12 after product and variable delivery costs. The twenty extra orders would contribute $240. If rewards used by all members cost $180, the app costs $50 and extra support costs $40, the pilot loses $30 before any other costs. The higher order count sounds encouraging, but the program has not paid for itself under these assumptions.

Now test the assumptions. Were the groups comparable in customer history and season? Were the twenty orders genuinely incremental, or did customers move purchases into the measurement month to earn rewards? Did the program change return rates? A spreadsheet with uncertain inputs should show a range, not one precise forecast. For example, if only ten orders were incremental, the contribution falls to $120 and the loss grows.

This calculation is not a universal formula for loyalty. It is a way to prevent a visible reward from hiding a real cost. Use your own contribution margin, reward redemption and support data, and revisit them after the pilot.

When to stop or redesign

Set stopping conditions before launch: unexpectedly high reward cost, increased refund requests, confusing rules or no credible incremental purchases. A short pilot can still teach you what customers value. If many members ask for easier returns rather than discounts, improve the returns process before building a more elaborate point system. If only heavy buyers use the reward, consider whether a simpler thank-you or targeted service would achieve the same relationship at lower cost.