In brief: The global loyalty program market is worth over 10 billion dollars, but three decades of independent research — Sharp & Sharp (1997), Meyer-Waarden & Benavent (2006), Leenheer et al. (2007), Verhoef (2003) — show that the effect on behavioral loyalty is minimal or statistically non-significant. The real problem is the selection effect: those who sign up are already heavy buyers, and the program takes credit for pre-existing loyalty. Light buyers — 80%+ of the customer base and the engine of growth — do not even notice the program. Customer satisfaction predicts only about 10% of long-term business performance (Gruca & Rego, 2005).
Let us say it right away, without mincing words: your loyalty card is almost certainly a waste of money. And we are not saying this for the sake of provocation. Science says it. Thirty years of empirical studies say it. Common sense says it, if only someone would stop to examine the numbers instead of repeating like a parrot that "retaining customers costs less than acquiring them."
And yet, in 2026, companies continue to pour billions of euros into loyalty programs. Point cards, apps with digital stamps, VIP tiers, cashback, gamification of spending. An entire ecosystem built on a flawed premise: that you can significantly change purchasing behaviors by rewarding those who already buy.
At Deep Marketing, we have a certain habit: checking whether the things "everyone knows" are actually true. Spoiler: they usually are not. And loyalty programs are perhaps the most glaring case of disconnect between widely held belief and empirical reality in all of modern marketing.
Brace yourselves. This article will be painful for anyone who has just renewed their contract with their loyalty platform vendor.
How much loyalty programs cost businesses
Before dismantling the myth, let us put the numbers on the table. According to industry estimates, the global loyalty program market is worth over 10 billion dollars and continues to grow. In Europe, the average retailer spending on their loyalty program ranges between 2% and 5% of revenue. That is not pocket change.
Consider what a serious loyalty program entails: the technology platform, integration with CRM and POS systems, points and rewards management, dedicated staff, specific communications, the cost of the rewards themselves, customer support for program-related issues. An expensive infrastructure that must justify its own existence.
And here is the interesting part. Because when you go to measure the real impact of all this machinery on customers' behavioral loyalty, the results are, to use an academic euphemism, "disappointing."
What the research says: three decades of ignored evidence
The scientific literature on loyalty programs is vast and, for those who promote them, rather uncomfortable. Let us look at the main studies.
Sharp and Sharp (1997): the FlyBuys study
Byron Sharp and Anne Sharp, of the Ehrenberg-Bass Institute, conducted one of the most rigorous studies ever carried out on a loyalty program: FlyBuys Australia, one of the largest loyalty programs in the world. Their conclusion? The effect on behavioral loyalty was so small as to be inconsistent across participating brands. In other words: the program was not measurably making anyone loyal. Customers who participated bought exactly as they would have bought without it. The study, published in the Journal of Advertising Research, remains a milestone that loyalty platform vendors prefer not to cite.
Meyer-Waarden and Benavent (2006): the French confirmation
In France, Lars Meyer-Waarden and Christophe Benavent analyzed loyalty programs in the retail sector. Result: minimal effects on behavioral loyalty. When the effects were corrected for selection bias (more on this shortly), most of the supposed impact evaporated. The "loyal" program customers were simply customers who were already loyal before.
Leenheer et al. (2007): the Dutch data
In the Netherlands, Jorna Leenheer and colleagues examined loyalty programs in grocery. Once again: minimal impact on participants' share-of-wallet. The study showed that most of the variance in purchasing behavior was explainable by factors that preceded enrollment in the program, not by the program itself.
Verhoef (2003): financial services
Peter Verhoef studied loyalty programs in the financial services sector, analyzing the impact on retention and cross-buying. The verdict: minimal and often non-significant effects. Loyalty programs did not produce a detectable increase in the duration of the customer relationship nor in the propensity to purchase additional products.
Sharp (2010): the definitive verdict
In 2010, Byron Sharp dedicated the entire Chapter 11 of How Brands Grow to loyalty programs, titling it "Why Loyalty Programs Don't Work." His analysis is merciless: loyalty programs do not alter the loyalty structure of the market, do not change the empirical laws of purchasing behavior (such as Double Jeopardy), and their apparent successes are almost entirely explainable by the selection effect.
To summarize: five independent studies, four different countries, different sectors, same result. When science speaks with this clarity, we must have the courage to listen.
The selection effect: the Achilles heel that nobody wants to see
If loyalty programs do not work, why are companies convinced otherwise? The answer lies in a statistical error as common as it is devastating: the selection effect (or selection bias).
The mechanism is simple and, once understood, impossible to ignore.
Who signs up for a loyalty program? Not a random sample of all customers. Predominantly the customers who already buy a lot and often: the heavy buyers. They are the ones who notice the sign at the checkout, who have the time and motivation to fill out the form, who have a reasonable expectation of accumulating enough points to earn a reward.
When the company then compares the behavior of "program members" with "non-members," it discovers (surprise!) that members spend more, visit more often, and have a higher lifetime value. And it triumphantly concludes that the program works.
But this is a completely fallacious conclusion. Those customers would have spent those amounts anyway. It is not the program making them loyal: they are loyal a priori, and the program has simply cataloged them. As Sharp wrote: "The loyalty program attracts already loyal customers, then takes credit for their loyalty."
Think about it for a moment. If you created a "joggers' club" in a park and then discovered that club members run more than the average citizen, would you conclude that the club makes people run more? No. The club simply attracted people who already ran.
And here we arrive at a crucial point: increasing the size of a biased sample does not make it more accurate. Having a million loyalty program members does not improve the quality of the insight if those members are systematically different from your actual market.
Why light buyers matter more than heavy buyers
This is the part that truly topples the house of cards of loyalty programs. And to understand it, you need to understand how the customer base of any brand actually works.
The distribution of purchases follows a well-known law in marketing science: a small percentage of customers (the heavy buyers) generates a disproportionate share of sales. But there is another side of the coin that is systematically ignored: light buyers make up 80% or more of the customer base of any brand.
And here is the key point: a brand's growth depends primarily on acquisition and penetration among light buyers, not on intensifying purchases from heavy buyers. This is one of the most robust and replicated findings in all of marketing science, documented by decades of research at the Ehrenberg-Bass Institute.
Loyalty programs do exactly the opposite of what is needed. They concentrate resources on those who already buy a lot (and who, by definition, have little room for growth), completely ignoring the enormous mass of occasional customers who represent the brand's true untapped potential.
Light buyers:
- Do not notice the loyalty program (they buy too infrequently to pay attention)
- Do not sign up (the entry barrier, however low, exceeds their motivation)
- Do not accumulate points (their few transactions do not generate reachable rewards)
- Do not receive program communications (not being enrolled, they are invisible)
The result? The loyalty program creates a self-referential bubble where the company speaks only to those who already know and choose it, convincing itself that the market consists of those customers. While the real market, the one where you win or lose, remains outside the door.
As Dekimpe et al. (1997) highlighted, customer loyalty is partial and directed at an entire repertoire of brands. No consumer is truly "loyal" to a single brand. They buy from a set of known brands, choosing based on availability, context, price, and mental salience. The loyalty program does not change this structure: at best, it ignores it.
Regression to the mean: why "loyal customers" do not stay that way
There is another statistical phenomenon that loyalty program advocates prefer not to mention: regression to the mean.
What is it in simple terms? Customers who in one period spent much more than the average will tend, in the following period, to return to more normal spending levels. And vice versa: customers who spent little will tend to spend slightly more.
This phenomenon is purely statistical; it has nothing to do with customer satisfaction, product quality, or the effectiveness of the loyalty program. A heavy buyer in 2025 might be a medium buyer in 2026 simply because their extreme behavior was partly due to temporary circumstances (a move, a birth, a job promotion).
What does this have to do with loyalty programs? Everything. Because when a company identifies its "best customers" and places them in the VIP tier of the program, it then observes that these customers "decline" slightly in the following period. And concludes that the program did not work hard enough and that more needs to be invested to "retain" them. In reality, the decline was inevitable. It is statistics, not disloyalty.
Similarly, if a loyalty program "recruits" new members among those who have just made a large purchase, the subsequent decline will be attributed to the program's lack of effectiveness. Leading the company to spend even more on incentives. A costly and futile vicious cycle.
The only scenario where it (maybe) makes sense
Let us be honest: there is one scenario in which a loyalty program can have marginal value. Not the one they sell you, but a real, tangible, and measurable value.
That value lies in data collection and the creation of a direct communication channel. A well-managed loyalty program gives you the email address, phone number, and purchase history of a portion of your customers. And this undoubtedly has value.
But — and this "but" is enormous — the data you collect through a loyalty program is inherently biased. It predominantly represents heavy buyers, not your market. Using it to make strategic decisions is like studying a country's population by only interviewing people who go to the gym every day: you will get a completely distorted picture.
Furthermore, that direct communication channel reaches only those already positively predisposed toward your brand. It is preaching to the choir. It may have tactical value (targeted promotions, new product launches to a receptive base), but it does not change the competitive structure of the market.
If you already have a loyalty program, use it for what it is: a tactical communication tool. But do not delude yourself that it is "retaining" anyone.
Customer satisfaction is not the answer either
At this point, someone will object: "Fine, the loyalty program does not work, but at least customer satisfaction generates loyalty, right?" Unfortunately, here too science is less enthusiastic than NPS survey vendors would like.
Gruca and Rego (2005) demonstrated that customer satisfaction predicts only about 10% of long-term business performance. Ten percent. Which means that 90% of performance depends on other factors, many of which relate to market penetration and the brand's mental salience.
This does not mean that customer satisfaction does not matter. It means it is a necessary but absolutely insufficient condition. A poor product will sink you, but an excellent product will not save you if nobody knows you exist.
Where to invest that money: brand building, reach, and penetration
Here we are at the constructive part. If loyalty programs are a low-return investment, where should that money go?
The answer comes from a fundamental study: Krasnikov and Jayachandran (2008), published in the Journal of Marketing, demonstrated that a company's performance correlates 35% with marketing quality, more than with R&D (28%) and operations (21%). Marketing, when done well, is the most powerful driver of business performance. But "done well" means something very different from "managing a points program."
Here are the three areas where you should reallocate your loyalty budget.
1. Brand building and mental salience
Brand building is the investment that generates the highest returns in the long run. Building distinctive brand assets (recognizable logo, iconic colors, memorable jingle, distinctive packaging), creating advertising that activates relevant mental associations for the category, being present in moments when the consumer is not yet looking for the product.
Brand building works because it acts on all potential buyers, not just existing customers. It increases the probability that your brand will be considered at the moment of purchase, by light buyers, medium buyers, and heavy buyers alike.
2. Reach and mental distribution
Reach — reaching the largest possible number of potential category buyers — is the single most important factor for growth. Not frequency, not hyper-specific targeting, not obsessive retargeting. Reach.
Every euro shifted from the loyalty program to broad-coverage media (TV, digital video, OOH, sponsorship) reaches potential customers that the loyalty program would never see. And those are exactly the ones you need to grow.
3. Market penetration
Growth happens primarily through increasing penetration: more people buying your brand at least once. Not through increasing purchase frequency among existing customers (which is exactly what loyalty programs attempt to do, and fail).
Invest in trial, sampling, first-purchase promotions, widespread distribution, ease of access. Make it easy for a new customer to try your product. This generates more growth than a thousand point cards.
Loyalty programs vs brand building: the comparison
Where to reallocate the budget: a practical guide
If you are thinking about dismantling your loyalty program (or not creating one), here is a concrete table on how to redistribute that budget more effectively.
FAQ: the questions we get asked (and the uncomfortable answers)
"But is Amazon Prime not a successful loyalty program?"
Amazon Prime is not a loyalty program in the traditional sense. It is a paid subscription service that eliminates purchase barriers (free shipping, speed). It works because it reduces friction, not because it rewards loyalty. Moreover, Amazon is a unicorn case: it has market penetration and physical availability (catalog) that no other retailer can replicate. Using Amazon Prime as justification for your local chain's points program is like using SpaceX's success to justify building a rocket in your garage.
"Our data says program customers spend 40% more"
Of course they do. And this is exactly the selection effect we discussed. Those customers were already spending 40% more before they enrolled. It is not the program making them spend more: the program simply intercepted those who already spent a lot. To demonstrate a real effect, you would need to compare the behavior of the same customers before and after enrollment, controlling for regression to the mean. When studies do this (Sharp & Sharp, 1997; Meyer-Waarden & Benavent, 2006), the effect disappears.
"But our sector is different"
Everyone says that. Yet the cited studies cover retail (Australia, France, Netherlands), financial services (Verhoef, 2003), aviation, hospitality. The underlying principle — the selection effect and the structure of market loyalty — is universal. Double Jeopardy is observed in every product category studied, from detergents to cars. The empirical laws of marketing have no sectoral exceptions.
"We cannot eliminate the program, customers would complain"
This is indeed a valid practical consideration. Discontinuing a loyalty program can generate short-term discontent, especially among heavy buyers who have accumulated points. The solution is not necessarily an abrupt shutdown, but a gradual transition: progressively reduce investment in the program, shifting resources toward brand building, and stop acquiring new members. In the meantime, honor the accumulated points. It is less traumatic than you think: most loyalty points are never redeemed.
"But loyalty programs collect valuable data"
They collect data, yes. But data systematically biased toward heavy buyers. If you use that data to make strategic decisions, you are piloting the company looking at a mirror that shows only 20% of the road. For representative data about your actual market, invest in panel research, brand tracking, and Media Mix Modeling. They cost less than a loyalty program and provide incomparably more useful information.
"Does customer satisfaction really not matter?"
It matters, but much less than you think. Gruca and Rego (2005) quantified it: about 10% of the variance in long-term business performance is explainable by customer satisfaction. The remaining 90% depends on factors such as market penetration, the brand's physical and mental availability, and the overall quality of the marketing strategy. A satisfied customer who does not remember your brand at the moment of purchase is a lost customer just as much as a dissatisfied one.
"Our competitor has a loyalty program, we cannot stay out"
If your competitor is throwing money out the window, you are not obligated to imitate them. In fact, the fact that your competitor is investing in loyalty gives you a competitive advantage: the resources they waste on rewards for already-loyal heavy buyers, you can invest in reach and brand building to win new customers. While they talk to those who already buy from them, you talk to the entire market.
The big picture: marketing that actually works
Let us close with a broader reflection. Loyalty programs are a symptom of a deeper strategic error: the belief that growth comes from "retaining" existing customers rather than winning new buyers.
Marketing science tells us the opposite. Krasnikov and Jayachandran (2008) demonstrated that marketing quality accounts for 35% of business performance, more than R&D (28%) and operations (21%). But "marketing quality" means reach, brand building, penetration, not point cards.
Marketing that works in 2026 — the kind based on evidence, not hype — is marketing that:
- Reaches the broadest possible market, not just existing customers
- Builds mental salience through distinctive brand assets and memorable creativity
- Facilitates purchase through widespread distribution and physical availability
- Measures what matters: penetration, share of voice, mental availability
- Invests for the long term: brand building pays increasing dividends over time
If you are investing 3% of revenue in a loyalty program and 0% in brand building, you are doing the exact opposite of what science suggests. And you are probably losing market share to competitors who have understood where to put the money.
Our advice? Take the loyalty program budget, shift it to reach and brand building, and compare results in 18 months. Science is on your side. The loyalty gurus, not so much.
Sources and References
- Sharp & Sharp — Loyalty Programs and Their Impact on Repeat-Purchase Loyalty Patterns, International Journal of Research in Marketing (1997)
- Ehrenberg-Bass Institute — About Loyalty
- LinkedIn B2B Institute — The Loyalty Lie
- Ad Age — Byron Sharp: Evidence-Based Marketing — Why Reach Beats Loyalty
- Krasnikov & Jayachandran — Marketing Capability and Firm Performance, Journal of Marketing (2008)
- Sharp, B. — How Brands Grow, Ch. 11: Why Loyalty Programs Don't Work (2010)
- R-Bloggers — Do Loyalty Programs Actually Create Loyalty? (2026)
