Used by businesses for more than 25 years, loyalty programs aim to entice consumers to make repeat purchases by offering them rewards—things like discounts on future purchases or points toward free airline tickets.

Since companies continue to expand them, one would think loyalty programs are powerful tools for boosting market share. Our research indicates many aren’t, at least not as designed.

The biggest problem with loyalty programs, we would argue, is that most retailers adopt a one-size-fits-all approach: They use monetary rewards to encourage repeat purchases. But product discounts won’t change buying behavior in the long run in shoppers who value things like personalized service, convenience or shopping pleasure more. These types of consumers may change their behavior to access the price promotion, but they likely will revert back to their regular brands or buying habits shortly thereafter, resulting in, at best, a temporary change in sales and market share.

Loyalty programs also seem to be mainly of interest to existing customers—the heavier, more frequent, more loyal buyers of the store, who tend to live closer to it. In researching grocery stores, our team found that 88% of loyalty-card holders were clients of the store two years before joining the program. Loyalty programs often fail to attract the competition’s customers because the rewards aren’t compelling enough to make them want to switch brands or stores. Indeed, the proliferation of loyalty programs offering the same kinds of rewards has destroyed a key reason for them in the first place: differentiation.

A more effective way to woo customers and maintain their patronage is to offer them individualized rewards, based on what they value. By offering different types of rewards to different groups of shoppers, companies set themselves apart and give people a reason to keep coming back. Providing access to a speedy checkout lane, for example, would be a more powerful way to win the loyalty of a person who hates grocery shopping than would a discount on a future purchase.

Here is how to build a loyalty program with the best chance of paying off:

Group customers according