Learn · Should you run a loyalty program at all
What a loyalty program cannot fix
A loyalty program amplifies an existing reason to return. It cannot manufacture one. Rewards operate at the margin, around one or two percent of spend, so they cannot close a structural price gap or undo a service failure. Reward-driven repeat purchases can also mask the underlying fault for months.
A loyalty program amplifies an existing reason to come back. It cannot manufacture one. The mechanism only ever operates at the margin of a decision the customer was already close to making, because the reward is small relative to the purchase, and that single fact draws the boundary around what the tool can do.
## The reward is one or two percent of spend, and the problem usually is bigger
Work the sizes. A typical program returns between 1 and 2 percent of spend in reward value. A competitor priced 10 percent below you on a 100 dollar basket saves the customer 10 dollars every trip; your program hands back perhaps 1.50. The reward covers 15 percent of the gap, and the customer's own arithmetic covers the rest. Points can hold a customer at rough price parity, where the decision is genuinely marginal. Against a structural price disadvantage, the program is a rounding error with a marketing budget.
## Churn from harm does not respond to points
Two customers stop buying. One drifted away because nothing distinguished you, and a well-timed offer can genuinely reach that person. The other left because an order arrived broken twice, and the same offer reads as an insult, because indifference was never the cause. The two are indistinguishable in a lapsed-member report and unmistakable in their trajectories: harmed customers often show strong engagement right up to an incident, then a cliff. A program addresses the first kind of churn only. Service failures need the service fixed, and no reward balance changes what happened.
## Rewards can hide the fault while it grows
The dangerous case is a program doing its job on top of a product problem. Reward-driven repeat purchases prop up the retention line, so the topline number managers watch stays healthy while underlying preference decays. The program delays the signal. By the time the decay shows through, it carries months more momentum than it needed to, and the instinctive response, richer rewards, buys another quarter of the same disguise. Any retention gain that evaporates whenever promotions pause was rented, and the rent goes up each time.
## Availability sits outside the program's reach
A customer your program has genuinely persuaded still cannot buy what is out of stock, and will not routinely drive past three competitors to reach you. Distribution problems produce churn that gets attributed to fickleness and that points cannot touch. The diagnostic is cheap: if lapsed customers' purchase histories show failed attempts, stockouts and substitutions in their final months, the retention budget belongs in inventory and the program should wait its turn.
## What the tool legitimately does
Inside its boundary the program has real work. It breaks ties at parity, where a small reward is decisive precisely because nothing else is. And it attaches identity to transactions, which makes every other retention effort more precise than it could otherwise be. Both are worth paying for when the module 0 tests pass. The discipline this lesson asks for is diagnostic: before attributing any retention problem to insufficient rewards, rule out the causes above, because points spent against them are points spent hiding them.