Launching a Loyalty Program Is Easy. Closing It Is Suicide.

A retail executive called me last month in a panic.

"Our loyalty program is bleeding money. When we tried to scale it back, customers went nuclear on social media. NPS dropped 30 points overnight."

I wasn't surprised.

Why it matters: This is a structural trap built into how loyalty programs work — and most executives walk into it eyes open.

The pitch always sounds the same

Spend more, shop more often, stay longer. The program pays for itself through lifetime value.

The CFO nods. The CEO approves. Marketing celebrates.

Then the costs arrive — immediately:

• Technology platform: €50,000–500,000 setup
• Point liability growing with every transaction
• Customer service handling redemption disputes
• A full-time team to manage it all

The benefits? Always "coming next quarter."

The real math

Take a standard points mechanic: spend €1,000, earn 1,000 points at 1¢ each, redeem 800 points for an €8 reward.

Add €2–3 in operational overhead per transaction.

You just built a 1% discount with up to 40% operational markup on top.

A straight 1% coupon costs exactly 1% — with zero overhead.

One major European retailer I consulted spent €47 million on their loyalty programme over three years. Incremental revenue: €12 million.

They paid €47M to generate €12M. Call it what it is: financial suicide.

Why you can never stop

Once customers receive loyalty benefits, they treat them as property — property they own, full stop.

Behavioural science calls this the endowment effect. Losing something you already have feels roughly twice as painful as gaining something new feels good.

I learned this during my time transforming loyalty at IKEA.

Every programme change — no matter how minor — triggered the same reaction: "You're stealing from loyal customers."

We were optimising a programme that was haemorrhaging money. Customers see their wallet. The P&L is invisible to them.

The most instructive moment: we reduced bonus point multipliers on a single product category. The saving was marginal. The backlash was disproportionate — customer service volumes spiked, social sentiment tanked, and the story ended up in Swedish retail press.

We reversed the change within two weeks.

Loyalty programmes work like government entitlements: easy to create, political suicide to remove. When JCPenney eliminated coupons and promotional pricing in favour of everyday low prices, full-year revenue fell ~25%. The CEO lost his job. The lesson transfers directly: once customers internalise a benefit — whether points, tiers, or a Wednesday discount — removing it triggers the same psychological response as theft.

Three signs your programme is already failing

The excitement trap. Customers love earning points — and keep spending exactly as they did before. You're giving away margin for unchanged behaviour.

The complexity spiral. You add tiers, bonus categories, partner earn. Complexity replaces the value the programme never delivered.

The metrics mirage. Programme members spend 20% more — because your best customers self-selected into the programme. The programme reflects your audience, it didn't create it.

When programmes actually work

This isn't a universal indictment. Some programmes generate real returns — and they share three characteristics.

The points have external value. Delta's partnership with American Express generated $7.4 billion in 2024. American Airlines collected $6.1 billion from credit card partnerships. United: $2.9 billion from Chase. These airlines are financial services companies that use flights as loss leaders. The programme is the product. Most retail programmes will never reach this model — and shouldn't pretend otherwise.

The programme unlocks cross-sell. Amazon Prime charges upfront and uses membership to drive adjacent revenue — video, grocery, pharmacy. The loyalty mechanic funds itself through category expansion.

The programme owns a behaviour shift. Starbucks moved customers from cash to app — capturing data, reducing transaction costs, enabling personalisation at scale. The result: mobile orders now exceed 30% of all US transactions, loyalty members account for ~60% of sales, and the app holds $1.77 billion in customer deposits. The points were the incentive. The behaviour change was the business case.

If your programme delivers none of these, you're running a discount system with organisational overhead on top. The honest move is to convert it — a straight cashback mechanic, a subscription, or a tiered service model — before the liability on your balance sheet makes the decision for you.

The bottom line

Before you launch, answer one question: what behaviour does this programme change — and what's that change worth?

Points move customers from buying to gaming. The best customers learn to wait for triple-point events. Margins erode. Complexity grows. And the exit costs more than the entry ever did.

The companies with durable customer relationships invest in the product, the service, and the moment of truth. They use loyalty mechanics to reinforce behaviour they've already earned — not to bribe behaviour they haven't.

One question before you sign the contract: are you rewarding loyalty, or manufacturing the illusion of it?