In short: Loyalty programs do not significantly change purchasing behavior. Empirical research — Sharp & Sharp 1997, Ehrenberg-Bass Institute, Romaniuk & Sharp 2022 — shows that retention is explained by the Double Jeopardy Law, not by points programs. A brand's growth depends on mental availability and physical availability, not on retaining already-loyal heavy buyers.

• Sharp (2010), How Brands Grow, Ch. 11: loyalty programs do not alter the empirical laws of purchasing behavior (Ehrenberg-Bass Institute)
• Double Jeopardy Law: brands with lower market share have fewer buyers and less frequent buyers — a pattern replicated across hundreds of categories (Journal of Advertising Research)
• NBD-Dirichlet: the distribution of purchases follows a stochastic law independent of loyalty programs (Ehrenberg, Goodhardt & Barwise, 1990)
• Byron Sharp & Jenni Romaniuk, How Brands Grow Part 2 (Oxford University Press, 2022): growth comes from penetration, not from loyalty

Do loyalty programs actually work?

The short answer — the one that makes loyalty platform vendors wince — is no, or at least far less than commonly believed. The long answer is that for thirty years, marketing science has accumulated empirical evidence against the hypothesis that loyalty programs generate loyalty. And this evidence is systematically ignored by those who sell such programs.

At Deep Marketing we always start from the same question: what does peer-reviewed data say? On loyalty programs, the data says something uncomfortable for an entire industry. Let's take a look.

What is a loyalty program

A loyalty program is a structured system of incentives — points, discounts, VIP tiers, cashback, rewards — that a company offers to customers to reward and stimulate repeat purchases. The stated goal is to increase retention, share-of-wallet and customer lifetime value. The most widespread forms are retail points cards, airline frequent flyer programs, bank cashback and apps with digital stamps.

The implicit assumption of any loyalty program is simple: rewarding repetition increases repetition. It's a reasonable hypothesis, but it's not what the data shows when you control for confounding variables.

What science says: 3 pieces of evidence

The literature on loyalty programs is vast. Here we summarize the three most robust and replicated findings, all attributable to independent peer-reviewed research or to research institutes not funded by loyalty platform vendors.

1. Sharp & Sharp (1997): the FlyBuys study

Byron Sharp and Anne Sharp, publishing in the International Journal of Research in Marketing, analyzed FlyBuys Australia, one of the largest multi-brand loyalty programs in the world. They compared the loyalty curves of participating brands with those predicted by the Dirichlet model — the reference standard for purchasing behavior in competitive categories. The program did not produce significant deviations from the Dirichlet predictions: participants bought as they would have bought without the program.

2. Ehrenberg-Bass Institute: the Double Jeopardy Law

The Double Jeopardy Law, formalized by Andrew Ehrenberg from the 1960s onward and replicated in hundreds of categories by the Ehrenberg-Bass Institute, states that brands with smaller share suffer a double penalty: they have fewer customers and, among those customers, a slightly lower purchase frequency. Retention is therefore a mathematical function of market penetration, not of the loyalty program. Increasing loyalty without increasing penetration is statistically impossible.

3. Romaniuk & Sharp (2022): How Brands Grow Part 2

Jenni Romaniuk and Byron Sharp, in How Brands Grow Part 2 (Oxford University Press, 2nd ed. 2022), extend the analysis with global data across consumer, B2B and services categories. The conclusion is unambiguous: growth derives from the acquisition of new buyers (penetration), not from intensifying purchases by existing heavy buyers. Loyalty programs, by focusing on those who already buy, act on the wrong lever.

The Double Jeopardy Law explained

The Double Jeopardy Law is probably the most robust empirical regularity in marketing science. In simple terms: in a competitive market, small brands are penalized twice. First, they have fewer buyers in absolute terms. Second, each of those buyers buys the brand slightly less often than a leader's customer buys the leader.

A counter-intuitive consequence: retention is an output of penetration, not a substitutable input. If Coca-Cola has higher retention rates than a local cola, it isn't because Coca-Cola "retains better", but because it has more penetration, more physical availability and more mental salience. A loyalty program cannot reverse this relationship: at best, it collects the data of those already bound to come back.

That's why the typical promise of loyalty vendors — "increase your customer retention" — is, according to marketing science, a promise that confuses correlation and causation.

What really works (Mental Availability + Physical Availability)

If loyalty programs aren't what grow brands, what does? The Ehrenberg-Bass Institute's answer boils down to two concepts: mental availability and physical availability.

Mental availability is the probability that the brand will be thought of at a relevant Category Entry Point: when I'm thirsty, when I need to give a gift, when I open the browser to book a flight. It's built with distinctive brand assets (logo, colors, jingle, packaging, iconic endorsers) and with broad, consistent advertising over time.

Physical availability is the ease with which the brand can be purchased when the customer chooses it: extensive distribution, online and offline availability, legible packs, frictionless checkout. Without physical availability, mental availability doesn't convert into purchases.

The table below compares popular loyalty program claims with the peer-reviewed empirical evidence.

Popular claim vs empirical evidence

When a loyalty program can make sense

It would be dishonest to dismiss every loyalty program as pure loss. There are scenarios where it can make sense, but not for the reasons usually given.

First: as a data collection and direct communication channel. A well-run program provides emails, purchase history, segmentation. The data is skewed (it represents heavy buyers, not the whole market), but it remains useful for tactical operations: product launches, service communications, inventory management.

Second: when it is itself a distinctive brand asset. Nintendo, Starbucks, Amazon Prime: in these cases the program is so integrated into the brand experience that it becomes part of mental availability. Amazon Prime, in particular, is not a loyalty program in the classic sense: it's a subscription that reduces purchase friction (shipping, content). It works because it increases physical availability, not because it rewards loyalty.

Third: in high-frequency, low-involvement categories (coffee shops, supermarkets, aviation), where a small recurring incentive can reduce competitor-switching friction. Even here, the effect is marginal and must always be weighed against the opportunity cost of investing the same budget in brand building.

Evidence-based alternatives

If the goal is growth, marketing science points to four priority levers, all documented in peer-reviewed literature.

1. Brand building. Invest in creativity and distinctive brand assets that build mental availability in relevant Category Entry Points. Binet & Field (IPA, 2013-2023) have documented that an optimal mix is roughly 60% brand building and 40% activation.

2. Reach. Reach the largest number of category buyers, not just existing customers. Reach — not excessive frequency on those already converted — is the primary driver of share growth.

3. Penetration. Trial incentives, sampling, extensive distribution, ease of first purchase. Penetration is the variable from which everything else derives, including retention.

4. Physical availability. Presence on the right channels, distinctive packs, zero purchase friction. A brand that's invisible on the shelf or impossible to find online won't grow even with the best loyalty program in the world.

Need help with your brand?

At Deep Marketing we work with Italian brands by building mental availability and physical availability based on the evidence of marketing science, not on the slogans of loyalty vendors. Request a free audit of your positioning, or discover our branding and visual identity consulting to work on the distinctive brand assets that truly matter.

Frequently Asked Questions

Why don't loyalty programs work?

Loyalty programs don't work as promised because they don't change the mathematical structure of the market described by the Double Jeopardy Law: retention is an output of penetration, not an independent input. The apparent success of programs is due to the selection effect: already-loyal heavy buyers sign up, and the program takes credit for pre-existing loyalty (Sharp & Sharp, 1997; Sharp, 2010).

What is the Double Jeopardy Law?

The Double Jeopardy Law is an empirical regularity in marketing science: brands with smaller market share have fewer buyers and, among those buyers, a slightly lower purchase frequency compared to leading brands. Formalized by Andrew Ehrenberg and replicated in hundreds of categories by the Ehrenberg-Bass Institute, it implies that retention and penetration are mathematically linked: you cannot increase the former without acting on the latter.

Can a loyalty program ever be useful?

Yes, but not as a loyalty engine. A loyalty program is useful as a data collection and direct communication channel, or when integrated as a distinctive brand asset (like Amazon Prime or Starbucks Rewards). In both cases it must be weighed against the opportunity cost: the same budget invested in brand building and reach typically generates a higher return on market share (IPA, 2013-2023).

What is the evidence-based alternative to the loyalty program?

The evidence-based alternative is to work on mental availability and physical availability: build distinctive brand assets, ensure broad reach across potential category buyers, facilitate penetration through trial and extensive distribution, and maintain a 60/40 ratio between brand building and activation as suggested by Binet & Field for the IPA. It's the strategy documented in How Brands Grow (Sharp, 2010) and in the decades of work by the Ehrenberg-Bass Institute.

Who is Byron Sharp?

Byron Sharp is professor of marketing science at the University of South Australia and director of the Ehrenberg-Bass Institute for Marketing Science, the largest independent research institute on purchasing behavior. He is author of How Brands Grow (Oxford University Press, 2010) and co-author with Jenni Romaniuk of How Brands Grow Part 2 (2nd ed. 2022). His work has redefined the evidence-based approach to marketing globally.

Sources and References

• Ehrenberg-Bass Institute — About Loyalty (decade-long synthesis of evidence on loyalty programs)
• Sharp, B. — How Brands Grow, Oxford University Press (2010)
• Romaniuk, J. & Sharp, B. — How Brands Grow Part 2 (Revised Edition), Oxford University Press (2022)
• Journal of Empirical Generalisations in Marketing Science — open access archive on empirical marketing laws
• Marketing Science Institute — research priorities and peer-reviewed working papers
• IPA — Binet & Field, The Long and the Short of It (2013-2023 updates)
• International Journal of Research in Marketing — Sharp & Sharp (1997) and related studies
• Reinartz & Kumar — The Mismanagement of Customer Loyalty, Harvard Business Review (2002)