Do loyalty programs really create loyalty?

Today’s article is “Loyalty Programs and Their Impact on Repeat Purchase Loyalty Patterns” by Byron Sharp, published in 1997 – a classic that remains surprisingly relevant.

The takeaway for marketing practitioners?

1. Loyalty programs do not create “excessive loyalty”: The main conclusion – that loyalty programs do not generate significant loyalty beyond what is expected for a brand’s market share – remains empirically supported. You cannot “buy” loyalty using points alone.

2. Focus on penetration, not retention: To get more loyal customers, you first need more customers. The law of “Double Jeopardy” applies: smaller brands have fewer buyers, and these buyers are slightly less loyal. This is not a problem you solve with a loyalty program.

3. Consumers are polygamous: Even Gold status travelers will fly competitors when the price or times are better. Loyalty cards do not fundamentally change purchasing habits: people buy from a directory of brands.

4. Modern loyalty programs have a different value: If the key behavioral findings still hold, why do brands continue to invest? The value proposition has shifted from discounts on plastic cards to digital data collection. First-party data, personalization and keeping brands “available” on customers’ phones may be the real benefits, not the points themselves.

Long version:

I recently revisited Byron Sharp’s 1997 article from International Journal of Marketing Research. It asks a simple question: Do loyalty programs really generate “excessive loyalty” – that is, do customers buy more from a brand than would be expected given its market share?

The answer, then and now, is no.

Sharp applied “Double Jeopardy” to loyalty programs. This law, well established in marketing science, states that small brands face a double penalty: they have fewer buyers and those buyers are slightly less loyal. The paper examines whether loyalty programs could help smaller brands break this trend.

They couldn’t.

The data showed that loyalty program members did not deviate from expected purchasing patterns based on their brand’s market share. A brand with a 5% market share had the loyalty profile expected of a brand with a 5% market share, whether or not it had a loyalty program.

This discovery has held up remarkably well. The Ehrenberg-Bass Institute continues to publish research demonstrating that loyalty is largely a function of penetration. You get loyal customers by attracting more customers, not by bribing existing ones with points.

But if the science is clear, why is every brand launching a loyalty app?

This is where the conversation has evolved since 1997. The paper evaluated loyalty programs in their original form: plastic cards offering delayed rewards. The mechanism was simple: buy more, earn points, eventually exchange them for something. The goal was retention.

Today’s loyalty programs are fundamentally different. Starbucks isn’t really trying to get you to drink more coffee through points – they’re creating a direct channel to you. Nike’s app isn’t about shoe discounts, it’s about first-party data.

There are two camps in current research:

Camp A (The scientific point of view): Sharp’s conclusions still hold. Meta-analyses show only weak correlations between loyalty programs and repeat purchase behavior. People remain polygamous buyers regardless of their loyalty status. A loyalty card on your keychain (or an app on your phone) doesn’t fundamentally change the way you make purchasing decisions.

Camp B (The relationship marketing point of view): It is true that behavioral loyalty is difficult to change. But loyalty programs are now stimulating mental availability (notifications keep brands in mind) and physical availability (which makes purchasing easier). The value lies in data-driven personalization and targeted media spend, not in the points themselves.

Here’s my take: Both sides are right.

If you launch a loyalty program hoping it will fundamentally change customer purchasing behavior, the evidence suggests you will be disappointed. The patterns of buyer behavior described by Sharp in 1997 – the predictability of loyalty given market share and the polygamous nature of consumers – remain true.

But if you’re building a loyalty program as a data acquisition and customer engagement channel – to understand your customers, personalize their experience, and remain accessible when they’re ready to buy – that’s a whole different matter. Don’t confuse this with “building loyalty.”

The article remains important because it challenges the fundamental assumption that loyalty programs generate loyalty. This is not the case. They can drive data, engagement, and convenience. But if you want more loyal customers, you’re better off focusing on growth.

Reference: Sharp, B. (1997). Loyalty programs and their impact on repeat purchase loyalty patterns. International Journal of Marketing Research, 14(5), 473-486.

Note: AI-supported content.

Related