Loyalty program design

Loyalty program design is the set of choices made before a rewards scheme launches: which behaviour you are paying for, what the reward costs against what it returns, and who carries the liability for points nobody redeems. It is a retention question.

This entry is about those decisions. A brand loyalty program is the scheme itself, what it is and what it does for a brand; design is what you settle before any scheme exists.

Nothing here is about acquisition. Design starts from people who already buy from you, which is why it belongs beside customer retention rather than beside campaign planning.

Key takeaways

• Start from the behaviour, not the reward — generosity is easy to copy and hard to withdraw.
• Price the scheme on breakage, then stress test it against the redemption rate doubling.
• Points owed are a balance sheet item, so finance signs the earn rate before marketing announces it.
• Rewards data is regulated data; a scheme that identifies buyers inherits obligations about that record.

How it works

Design runs as a sequence of five decisions, each with a cost attached. You choose a target behaviour, an earn rate, an expiry rule, an enrolment route and a funding owner. Get the first one wrong and the other four cannot rescue it.

The target behaviour decision is the one that decides whether the scheme pays back. A reward that fires on any purchase buys nothing from a weekly shopper. A reward that fires on a second category purchase buys something measurable.

Breakage is an assumption, not a fact — and it moves. Price the scheme on the share of rewards you expect never to be claimed, then model what happens if claims rise by half.

Points owed are a liability. The value sitting in customer balances belongs on the balance sheet rather than in last quarter’s marketing spend, which is why finance should approve the earn rate before any launch date is set.

Loyalty data is a liability of a different kind.

In December 2024 the Federal Trade Commission finalised an order requiring Marriott and Starwood to restore loyalty points stolen in the breaches, after failures that affected more than 344 million customers worldwide.

That order also required a way for customers to delete personal information tied to a rewards account number. A scheme that creates a named purchase history inherits that obligation from day one.

Paid membership is the structure with the most law attached. Under 15 U.S. Code section 8403, charging a consumer online through a negative option feature requires express informed consent before the card is charged and a simple way to stop recurring charges.

Design the measurement alongside the mechanics. A scheme judged on enrolment numbers will always look successful — enrolment is the one metric a discount reliably moves.

Set the control group before launch. Without one, you cannot separate the behaviour you bought from the behaviour you would have had anyway.

Examples

The useful comparison is not between schemes but between the decisions behind them. Three common design briefs show how differently the same mechanics get configured once the target behaviour is fixed.

A specialty retailer wants a shorter gap between orders, so the earn rate is banded: the reward only becomes worth claiming on the third purchase inside a quarter. The design buys frequency, and accepts low enrolment.

A business services provider wants a shift of channel, so the reward attaches to self-service ordering rather than to spend. Every order moved off the phone pays for itself twice, in reward cost and in handling cost avoided.

A subscription brand runs paid membership and therefore budgets for cancellations from the start. Its measurement sits on customer churn among members against non members, not on member satisfaction scores.

Each design needs a baseline. Measure the customer churn rate and the customer retention rate for the twelve months before launch, or the scheme will be credited with whatever the market was doing anyway.

Related terms

• Brand loyalty program: the scheme itself once it exists; design is the set of decisions taken before it does.
• Customer retention: the outcome a scheme is built to move, and the only fair test of the design.
• Customer churn rate: the baseline measure a design is judged against, before and after launch.
• Brand affinity: preference that exists without a reward attached, which a scheme can support but cannot manufacture.
• Brand ambassador program: pays for advocacy rather than repeat purchase, so it answers a different design brief entirely.

FAQ

What is the first decision in loyalty program design?

The behaviour you intend to change. Reward structure, earn rate and expiry all follow from it, and none of them can be set sensibly until it is written down.

How do tiers differ from cashback at the design stage?

Tiers reward status; cashback rewards spend — the two attract different customers. Tiers create a threshold people work toward, while cashback functions as a plain discount with simpler maths and fewer complaints.

What is breakage and why does it decide the budget?

Breakage is the share of earned rewards never claimed. Schemes are priced on it, so a rise in redemption can turn a profitable design into a loss making one quickly.

Should points expire?

It depends on which cost you prefer. Expiry reduces the liability carried on the balance sheet and raises complaints, so the decision belongs to finance and customer service together, not to marketing alone.

Who should own the programme after launch?

One named owner with access to both the reward cost and the retention data. Split ownership between marketing and finance is the usual reason a scheme runs for years without anyone able to say whether it works.

Shortlist vetted customer retention and loyalty partners in the Outsource Accelerator directory before you commit to a reward structure.