Customer Loyalty Programs: What the Data Doesn’t Tell You

Customer loyalty programs work, but not in the way most brands expect them to. The mechanics are sound: reward repeat behaviour, reduce price sensitivity, increase purchase frequency. The problem is that most programs are designed around the reward, not the relationship, and that distinction matters more than any points multiplier ever will.

Done well, a loyalty program can meaningfully improve retention, increase customer lifetime value, and generate behavioural data that sharpens your entire marketing operation. Done poorly, it trains customers to wait for discounts, attracts deal-seekers with no long-term intent, and costs you margin on purchases that would have happened anyway.

Why Most Loyalty Programs Underdeliver

I have worked with brands across more than 30 industries over two decades, and the loyalty program conversation comes up constantly. It usually starts with the same premise: we need to improve retention, so let us build a points scheme. That logic is not wrong, but it skips several important questions.

The first question is whether the retention problem is actually a loyalty problem. In many cases it is not. Customers are leaving because the product is not delivering, the onboarding is weak, or the pricing has drifted out of alignment with perceived value. A loyalty program layered on top of those problems does not fix them. It just gives you a more expensive version of the same churn.

The second question is who the program is actually designed to attract. Most loyalty schemes, when you look at the redemption data, are dominated by a small segment of highly engaged customers who were already loyal. The vast middle, customers who buy occasionally and could go either way, rarely engage with the mechanics at all. And the discount-driven segment, people who sign up for the welcome offer and disappear, actively costs you money.

If you are working through a broader retention challenge, the Customer Retention hub covers the full picture, from benchmarking and churn analysis to the structural decisions that actually move the needle.

The Difference Between Rewarding Loyalty and Creating It

This is the distinction that most program designers miss. Rewarding loyalty means recognising customers who are already committed. Creating loyalty means changing the behaviour and preference of customers who are not yet committed. These require completely different mechanics.

Points schemes are excellent at the first. They make loyal customers feel valued, they provide a tangible reason to consolidate spend, and they generate data that can be used to personalise future communications. Airlines and hotel chains have built significant commercial advantage from this model, not because the points are inherently valuable, but because the program becomes part of how customers make booking decisions.

Creating loyalty in the uncommitted middle is harder. It requires understanding why those customers are not consolidating their spend with you, and that answer is rarely “they haven’t found the right points scheme yet.” More often it is about experience gaps, product fit, or competitive alternatives that are genuinely better in some dimension. HubSpot’s breakdown of churn drivers is a useful reference here, because it distinguishes between customers who leave due to dissatisfaction and customers who leave due to inertia, and the interventions are very different.

I spent several years running a large performance marketing agency, and we had a client in financial services who had invested heavily in a loyalty program for their credit card customers. The program was well-designed on paper: tiered rewards, partner benefits, a clean app experience. But when we dug into the data, the customers engaging most with the program were already in the top quartile for spend and tenure. The program was rewarding loyalty that existed. It was doing almost nothing to move customers from the second quartile into the first. That is a common pattern, and it has real implications for how you calculate the ROI of the program.

What the Economics Actually Look Like

The financial case for loyalty programs is often built on assumptions that do not survive contact with real data. The most common error is attributing incremental revenue to the program that would have occurred anyway. If a customer who spends £500 a month earns points on that £500, and you count their continued £500 spend as program-driven retention, you have not proven anything. You have just described existing behaviour with extra steps.

Genuine incrementality requires a control group: customers with similar profiles who are not enrolled in the program, tracked over the same period. Without that, you are measuring correlation, not causation. This is the same discipline that applies to any marketing investment, and it is surprising how rarely it is applied to loyalty programs specifically.

The cost side is also frequently underestimated. Points liability accumulates on the balance sheet. Redemption rates vary and can spike unpredictably. Program administration, technology, and communications all carry ongoing costs. And if the program includes discounts or free products, you are reducing margin on every enrolled customer, not just the ones who would have churned.

MarketingProfs has a useful framework for thinking about loyalty and profitability together, which is the right framing. Loyalty that does not improve profitability is not a business outcome. It is a cost centre with a retention story attached.

Where Loyalty Programs Actually Add Value

None of this means loyalty programs are a bad idea. It means they need to be designed with commercial honesty rather than optimism. There are categories and contexts where they consistently deliver.

High-frequency, low-differentiation categories are the natural home for loyalty programs. Coffee, fuel, grocery, pharmacy: places where customers make the same decision repeatedly and the switching cost is low. In these contexts, a well-designed program genuinely changes where customers choose to spend, because the product itself offers little reason to prefer one provider over another. The program becomes the differentiator.

Categories with high lifetime value and long purchase cycles also benefit, but for different reasons. Here the program is less about frequency and more about staying present during the long gap between purchases. Automotive, insurance, and financial services use loyalty mechanics to maintain a relationship that would otherwise go dormant. The goal is to be the brand that comes to mind when the next purchase decision arrives.

The data angle is underrated. A well-run loyalty program is one of the most cost-effective ways to build a first-party data asset. When customers opt into a program, they give you permission to track behaviour, preferences, and purchase patterns in a way that anonymous transactional data cannot match. Forrester’s work on propensity modelling illustrates how that behavioural data can be used to identify customers at risk of churning before they actually leave, which is where retention investment pays off most efficiently.

When I was growing the agency from around 20 people to close to 100, one of the things we learned early was that client retention was not primarily about price or even output quality. It was about whether clients felt understood. The brands that build genuine loyalty, in any category, tend to be the ones that use what they know about their customers to make the experience feel more relevant. A loyalty program can be the infrastructure for that, if the data is actually used.

The Design Decisions That Determine Whether a Program Works

Assuming you have decided a loyalty program is the right tool for your situation, the design choices that follow have a large impact on whether it delivers. A few of them matter more than most.

Simplicity of value proposition. If a customer cannot articulate the benefit of your program in one sentence, engagement will be low. Tiered structures with multiple earn rates, partner networks, expiry conditions, and bonus categories are appealing to program designers and confusing to customers. The programs with the highest engagement rates tend to be the simplest: spend X, get Y, with a clear and immediate reward.

Reward timing. Behavioural economics is clear on this: immediate rewards are more motivating than deferred ones. A program that makes customers wait six months to accumulate enough points for a meaningful reward will see low engagement in the accumulation phase. Smaller, more frequent rewards, or progress indicators that make customers feel they are getting closer to something, outperform large deferred payouts in most categories.

Emotional versus transactional rewards. The most durable loyalty programs combine transactional benefits, the points, the discounts, with something that feels less like a financial transaction. Early access to products, exclusive experiences, recognition status, and personalised communication all contribute to a sense of being valued rather than simply incentivised. Moz’s analysis of local brand loyalty touches on this distinction well, noting that the emotional connection to a brand often matters more than the rational benefit of any specific offer.

Personalisation at scale. Generic loyalty communications are worse than no communication at all, because they signal that you have the data but are not using it. If a customer has bought the same product category three times and you send them a generic “earn double points this weekend” email, you have wasted the relationship signal. Testing personalised loyalty communications against generic equivalents consistently shows meaningful differences in engagement and conversion rates.

Exit mechanics. Most programs focus entirely on acquisition and engagement, and almost none think carefully about what happens when a customer becomes inactive. Points expiry handled badly is one of the fastest ways to generate negative sentiment from customers who were previously advocates. A grace period, a re-engagement offer, or a clear notification before expiry costs very little and preserves goodwill that took months to build.

The Measurement Problem

Loyalty programs are notoriously difficult to measure accurately, and most organisations are measuring them in ways that overstate their impact. The three most common measurement errors are: counting all enrolled customers as program-driven, attributing all retained customers’ revenue to the program, and comparing program members to non-members without controlling for the fact that program members self-select based on higher engagement.

That last point is particularly important. Customers who join a loyalty program are, by definition, more engaged with your brand than customers who do not. If you compare their retention rate to non-members and conclude the program is driving the difference, you are almost certainly wrong. The difference existed before the program. The program may be reinforcing it, but it did not create it.

Proper measurement requires either a randomised control trial, which is operationally complex but genuinely rigorous, or a matched cohort analysis that controls for pre-program engagement levels. Hotjar’s approach to understanding churn behaviour is a useful model for the kind of qualitative layer that should sit alongside quantitative measurement, because the numbers tell you what is happening and the qualitative research tells you why.

I judged the Effie Awards for several years, and one of the things that became clear across hundreds of submissions was how rarely brands could demonstrate genuine incrementality from their loyalty investment. The submissions that stood out were the ones that had built measurement frameworks before the program launched, not after. If you design the measurement at the end, you will find a way to make the numbers work. If you design it at the beginning, you will find out whether the program is working.

When a Loyalty Program Is the Wrong Answer

There are situations where a loyalty program will not solve the problem you are trying to solve, and building one anyway is an expensive distraction.

If your churn is driven by product gaps, a loyalty program will slow the bleed slightly while the underlying problem continues. Customers who leave because a competitor’s product is genuinely better are not going to stay for points. Forrester’s research on cross-sell and upsell dynamics makes the point that retention through value expansion, giving customers more reasons to stay rather than more rewards for staying, tends to outperform pure loyalty mechanics in categories where the product experience is the primary driver of preference.

If your margins are thin, a loyalty program that discounts existing revenue is likely to make the economics worse before they get better. The payback period on loyalty investment is typically long, and if the business cannot sustain the cost during that period, the program becomes a liability.

If your customer base is too small or too infrequent to generate meaningful data, the data advantage of a loyalty program disappears. A B2B business with 50 enterprise clients does not need a points scheme. It needs account management, relationship investment, and a clear expansion strategy.

The honest question to ask before building a loyalty program is: what is the actual mechanism by which this program will change customer behaviour, and is there evidence that mechanism works in our category? If you cannot answer that clearly, the program is probably not the right tool.

There is a lot more on the structural side of retention, what drives it, how to measure it, and where to invest, in the Customer Retention hub, which covers the full range of levers available beyond loyalty mechanics alone.

The Verdict

Customer loyalty programs work when they are designed around a genuine understanding of why customers stay or leave, built with commercial discipline, measured with intellectual honesty, and used as infrastructure for a better customer experience rather than a substitute for one.

They do not work when they are a response to a retention problem that is actually a product problem, when the economics are built on attributed rather than incremental revenue, or when the program is designed to impress a boardroom rather than change customer behaviour.

The brands that get the most from loyalty programs tend to be the ones that were already doing the fundamentals well. Good product, good experience, good communication. The program amplifies what is already working. It does not rescue what is not.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions