Do Loyalty Programmes Actually Increase Customer Retention?

Short answer: yes, but less than the headline stats suggest, and the reason why is genuinely interesting rather than just a boring caveat. I want to walk through both the strong evidence for loyalty programmes working, and the honest problem with a lot of that evidence, because I think you'll trust the conclusion more if I show you where it's shaky.

The case for loyalty programmes working

Accenture found loyalty members generate 12 to 18% more incremental revenue per year than non-members. McKinsey's "Next in Loyalty" research found top performing programmes lift revenue from actively redeeming members by 15 to 25% annually, and that redeemers spend a full 25% more than members who joined but never engage. Existing customers convert at a 60 to 70% rate, against 5 to 20% for a brand new prospect, per Marketing Metrics, and Bain's long-running research puts the cost of acquiring a new customer at 5 to 25 times higher than retaining an existing one. Paid, premium membership programmes do even better: McKinsey found paid members are 60% more likely to increase their spend, against 30% for free programme members.

That's a genuinely strong case, and if you stopped reading here you'd walk away thinking loyalty programmes are close to a guaranteed win.

The honest problem: self selection

Here's what almost nobody puts in the same article. McKinsey and the loyalty analytics firm Brandmovers have both pointed out that a big chunk of that member-versus-non-member gap is self selection, not programme effect. Your highest spending, most engaged customers were always the most likely to sign up for your loyalty programme in the first place. Comparing their spend to non-members isn't really measuring what the programme did. It's partly just measuring who joined.

McKinsey's own broader research is blunt about the result: roughly two-thirds of established loyalty programmes fail to deliver real value, and some actively destroy it, usually through cost without a matching behaviour change, or through devaluation that erodes trust faster than points get earned.

So what's actually driving retention

If it's not simply "having a loyalty programme," what is it? The evidence points fairly clearly at behaviour change, not enrolment. Programmes that get someone to do something differently, come back more often, engage more deeply, change a habit, show real retention gains. Programmes that just record purchases someone was going to make anyway don't move the needle much beyond the self-selection effect.

This is where the action based and gamified models covered elsewhere in this cluster start to look structurally different from a plain points card. Sweatcoin's link to a roughly 20% increase in users' physical activity, backed by research in the British Journal of Sports Medicine, is retention evidence of a genuinely different kind: it's not "loyal customers spent more," it's "the programme changed what people actually did." That's a much harder thing to fake with self-selection, because the behaviour itself is the measured outcome, not a spend comparison between two different groups of people.

The SAP Emarsys 2025 finding that true loyalty fell from 34% to 29% in a single year fits the same story. Enrolment isn't the problem. Engagement is. Antavo's data backs this up directly: the average member is active in only 8.8 of the 17.4 programmes they've joined (Bond, 2025), and $10 billion in points sit unspent in the US every year. People joined. They just stopped caring.

What this means if you're deciding whether to build one

Build a loyalty programme if you can design it to change a real behaviour, more frequent engagement, healthier or more sustainable choices, actual advocacy, not just to formalise the spending your best customers were already doing. If your programme's success metric is "member spend versus non-member spend," build in a way to account for self selection before you present that number to your board as pure programme impact, because McKinsey and Brandmovers will not be impressed, and neither should you be.

This is genuinely the argument behind everything Zenko does. Rewarding verified action rather than spend gives you a cleaner signal that something real happened, not just a comparison between people who were already going to behave differently. If that's the kind of programme you're trying to build, our brands page has more detail. And if you haven't already, the rest of this cluster covers the full picture: the complete 2026 guide, real examples across every model, how gamification adds a genuine engagement lift, and the build versus buy decision if you're ready to move on platform choice.

FAQ

Do loyalty programmes actually increase customer retention?

Yes, but the effect is smaller than headline statistics suggest, because a meaningful share of the member-versus-non-member spend gap reflects self selection rather than programme impact. McKinsey found roughly two-thirds of established programmes fail to deliver real value.

What's the strongest evidence that loyalty programmes work?

Accenture's finding that members generate 12 to 18% more incremental revenue than non-members, and McKinsey's finding that actively redeeming members spend 25% more than inactive enrolled members, are among the better sourced, more specific results available.

Why do some loyalty programmes fail to retain customers?

Usually because they reward a purchase the customer was already going to make, rather than changing behaviour, and because devaluation (expiring points, rising redemption thresholds) erodes trust faster than the programme builds it.

Is member spend the right way to measure loyalty programme success?

On its own, no. It's contaminated by self selection: your best customers were likely to join regardless of the programme's design. Behaviour change, like the roughly 20% activity increase linked to Sweatcoin in peer reviewed research, is a cleaner signal that the programme itself caused something.