The Five Pitfalls of Loyalty Program Design

Loyalty program design is the set of decisions, made before launch, about what a program rewards, what it costs, how members earn and redeem, how it stands apart from competitors, and how it will be measured. Those decisions are hard to change once members are enrolled, which is why design mistakes are among the most expensive a program can make.

The stakes are high. McKinsey reported in 2021 that around two-thirds of established loyalty programs fail to deliver value, with many actually eroding value. Some of those causes are set before launch: in how the program was structured, what it rewarded, and what it asked of members. Fixing them later usually means a costly redesign and a difficult conversation with members who liked the old rules.

Good design pays. Workplace-incentive research published by the Incentive Research Foundation (Stolovitch, Clark and Condly, 2002) found that properly structured incentive programs lift performance by an average of 22%, and by 44% for programs that run a year or longer. The studies measured employee performance, so treat the figures as directional for customer and channel programs, but the point holds: structure matters.

This guide covers five common and costly design pitfalls, and how to avoid each. Two caveats. These are not the only five; they are five that recur across program designs. And they interact: the fixes trade off against one another, so good design is less about avoiding each in isolation than about balancing them. Design pitfalls are also distinct from the operational failures that sink a well-built program after launch, which the guide to why loyalty programs fail covers.

The first design decision, and the easiest to get wrong, is what the program rewards. A program is a machine for producing whatever behavior it pays for: reward the easy-to-measure action and members will optimize for it, sometimes by working the rules rather than changing behavior. Reward pure transactions and the program may simply subsidize purchases customers would have made anyway. Offer a rich sign-up bonus and it attracts deal-seekers who claim it and leave.

Channel programs have their own version. Rewarding sell-in, what a distributor buys, rather than sell-through, what it actually sells, can pay distributors to stock up at quarter-end without selling any more product. The business case for smarter B2B incentives covers how revenue-only goals invite this kind of gaming.

The fix is to start from the business outcome, decide which member behaviors actually drive it, such as engagement, cross-category purchasing, advocacy, or retention, and design the earning structure to reward those, not just the easy-to-measure ones.

A large nutritional CPG brand built its program on BLOYL™ around this principle, turning an influencer rewards program into a full loyalty program that rewards missions and activities, not just purchases. It recorded more than 35,000 transactions in its first six months and more than 16,600 missions completed since launch, with a 62% engagement rate among members and a 3+ increase in average transactions per user. (Metrics disclosed by Brandmovers.) The case reports engagement rather than incremental sales.

Warning signs: most points are earned on purchases members would have made anyway; sign-ups spike around the join bonus and fade; partners or members find ways to earn without the behavior the program was built for.

Every loyalty program is an economic exchange: the member gives engagement and spend, the brand gives rewards, and the design has to make that trade work for both sides. Get the economics wrong in either direction and the program fails. Too stingy, and there is no reason to participate. Too generous, and the program erodes margin or builds a balance-sheet liability of outstanding points it cannot afford.

Weak economics is one route to the value erosion McKinsey describes. Programs can give away value without driving profitable behavior, and McKinsey cautions against programs that depend on breakage, the points members never redeem, "to make their economics look successful." Some breakage is normal and has to be forecast for the points liability; the danger is economics that depend on it, because high breakage is better read as a sign of disengagement than as a saving. The fix is to model the economics before launch, concentrate the richest rewards on the behaviors and members worth the most, and design earning and redemption so the program pays for itself through the behavior it changes.

As an illustration of the math: a program that awards 5 points per dollar, with each point worth 0.2 cents, gives back 1% of spend. If it expects 80% of points to be redeemed, the reward cost is 0.8% of member sales, or $8,000 for every $1 million those members spend. At a 30% gross margin, the program needs about $26,700 in incremental sales per $1 million just to cover its rewards, before platform, fulfillment, and marketing costs. Running that calculation before launch shows whether the earn rate is affordable and how much behavior change the program has to produce.

Warning signs: no one can state the program's cost as a share of sales; the reward budget is set by matching a competitor's earn rate; the business case only works if breakage stays high.

A loyalty program can be undone by its own cleverness. When earning rules are convoluted, tiers are too many, or redemption is a maze, members disengage, not because the value is not there, but because they cannot find it. Complexity is a design choice, and usually the wrong one.

The instinct to add mechanics, more tiers, currencies, and conditions, is strong, and each addition feels reasonable on its own. Together they produce a program only its designers understand. The fix is discipline: a member should grasp how to earn and how to redeem in seconds, and every added layer should have to justify itself against the confusion it creates. Simple does not mean unsophisticated; it means legible. For tiered programs specifically, the guide to tiered loyalty programs covers how to build a tier hierarchy members can follow, including the mistake of so many levels that no one knows where they stand.

This is also where the pitfalls trade off. Simplify too aggressively and a program can slide into the next pitfall, becoming generic and indistinguishable from every other points scheme. The craft is to keep the program easy to understand while still giving it a reason to exist. Scale changes the trade-off. A lean team is usually better served by one earn rule and one or two tiers it can run well; an enterprise program spanning brands or channels can carry more structure, but only with the analytics and staff to manage it.

As an illustration: a program with three currencies (points, stars, and bonus credits), five tiers, and different multipliers by category asks a member to track a dozen rules before they know what a purchase is worth. A program with one currency, one earn rate, and two tiers can be explained in a sentence: earn 5 points per dollar, and reach the second tier at a set annual spend. The second design gives up some precision, but members can see their value, and the team can test one change at a time.

Warning signs: the program rules need more than a short paragraph to explain; customer service fields the same "how do I earn" questions repeatedly; members hold balances they do not know how to use.

Many programs are designed by looking at what competitors do and copying it: earn a point per dollar, redeem for a discount, done. The result is a program indistinguishable from a dozen others, which gives members no particular reason to choose it or stay. A me-too design competes only on the size of the discount, the one axis everyone can match.

Deloitte's 2026 loyalty research found that up to 40% of a brand's perceived value is driven by factors other than price. That is the opening a me-too program leaves on the table. The same research found that US loyalty program members rank "overall value" as the top reason to join a new program, ahead of attractive ongoing benefits and immediate sign-up incentives, so value has to be built, not just discounted. Matching category norms can be table stakes; the pitfall is stopping there. The fix is to design for differentiation and brand fit: build the program around value only the brand can offer, whether experiential rewards, community, exclusive access, or a structure that reflects how its customers actually engage.

A leading Canadian regional distributor took that route with its Culture Club program on BENGAGED™. Rather than a transaction-only program, Culture Club was designed to focus on community and brand advocacy, rewarding purchases along with activity such as consuming content and taking quizzes and surveys. Sales among enrolled customers grew by an average of 25%, compared with 5% among non-enrolled customers, and the case reports a 2x increase in customer acquisition after launch. Because the comparison is between enrolled and non-enrolled customers rather than a controlled test, that gap may reflect who joined as well as the program. (Metrics disclosed by Brandmovers.)

Warning signs: the program description could be swapped onto a competitor's site without anyone noticing; the only lever anyone proposes is a bigger discount.

The most expensive loyalty program mistakes are the ones designed in and discovered later. Reward the wrong behavior, misjudge the economics, overcomplicate the structure, copy a competitor, or skip the data foundation, and no amount of clever marketing afterward will fully fix what the blueprint got wrong. These pitfalls are avoidable, and avoiding them is mostly a matter of deciding, deliberately and early, what the program is for: what it rewards, what it costs, how simple it is, how it differs, and how it will be measured. Good design does not guarantee results; it removes the failures that execution alone cannot fix. On BLOYL, earning and redemption rules can be set by product SKU, purchase channel, customer segment, or behavioral action, and its analytics include A/B testing against a control group, so the behavior, the economics, and the measurement can be designed into the same platform from the start.

The case, in numbers