For the past year, we’ve been up to our necks in one of those projects where the word “transformation” loses all its glamour. We’re rebuilding a marketing platform and a loyalty system for an international retail group: first the consulting phase, then implementation, then the first go-live in the first country. Big logo. Big machine. Lots of people. Lots of meetings. The kind of project where the steering committee stops talking about ideas and starts talking about what’s going to break tomorrow - and who’s going to catch it. Which is exactly why the last steering meeting was interesting. Something happened that’s not a given in projects like this: after the first go-live, someone voiced a doubt out loud. Not a technical doubt. A strategic one.
“We’ve got the technology. But… did we go bold enough? Did we just swap the engine, or did we actually change how the program shifts behavior?”
That’s the moment it hits you: you can have the best platform in the world and still end up with the most expensive version of a discount card. Better plumbing. Same old deal. And then someone said something even more uncomfortable—exactly what I’d been trying to explain in the consulting phase, but back then it sounded too “soft” because everyone was in love with the roadmap and the integrations:
“We have a loyalty program. But sometimes it feels like we’re paying for discounts for people who would’ve bought from us anyway.”
That’s it.
If I had to name the most expensive marketing habit of the last decade, it would be this: replacing loyalty with discounts, discounts with “strategy,” and strategy with the comforting feeling that “at least we’re doing something.” Most “loyalty programs” aren’t about loyalty at all. They’re a systematic way of handing margin back to your best customers - in exchange for data and a clean conscience. And then we act surprised when we’re stuck in a price war where the only “creativity” left is whether the discount shows up as a coupon, points, or cashback.
On paper, it’s tidy. Earn points, get a reward. Mathematically fair. The problem is the logic is built for Homo Economicus - an imaginary customer from an Excel spreadsheet who decides purely on price and rational benefit. That customer doesn’t exist.
In the real world, we’re emotional, social, and predictably irrational. We build loyalty with a feeling, not a calculation. If you build loyalty purely on logic, your only lever is price. And when price is your only lever, you’re not building customers into loyalty - you’re training them to wait for your next discount. That’s not loyalty. That’s a bribe. And bribes have a nasty property: someone can always outbid you.
Rory Sutherland has a sharp point here: solve problems only with logic and you end up with solutions anyone can copy. Loyalty programs are a textbook case. Everyone has “points per dollar.” Everyone has “a coupon.” Everyone has “an app.” And leadership teams soothe themselves with membership counts instead of asking the harder question: does this program actually change behavior - or are we paying for behavior we would’ve gotten for free.
My take: no psychology, just expense.
Predictability kills attention. “Spend $100, get $5 back” is fair - but psychologically dead. We don’t respond most strongly to rewards; we respond to anticipation - the chance that something nice might happen. That’s why classic programs get boring fast: they’re built like paperwork. When Chipotle runs “Guac Mode” and adds a “maybe you get something extra today” moment, they’re not selling guac. They’re selling attention. When Pret A Manger built “random acts of kindness” (a coffee on the house at an employee’s discretion), they weren’t handing out discounts. They were building a relationship.
The difference is brutally simple: a discount feels like a right; a gift becomes a story.
The second mechanism is ego - what the literature calls idiosyncratic fit, and in practice means: let me feel smart. People like to believe they’re above average. Not at everything, but at something. If everyone gets the same discount, it signals nothing. But if the program lets someone stand out - because they have a habit, discipline, knowledge, or taste - you get identity. That’s why Strava challenges work even when the “reward” is nothing special. A digital badge - and people still ride through rain and cold. Because the badge isn’t a thing. It’s proof of belonging. Same with Sephora Rouge: the threshold is irrational, but status isn’t a rational category. Status is psychological currency.
My verdict here: a “program for everyone” is usually a program for no one. If you want people to identify, you have to allow different types of customers to shine in different ways - even if that means some people reach rewards more easily than others.
Loyalty isn’t math. Loyalty is a feeling.
The biggest mistake in digital programs is “0 points” and a dead number. The start is psychologically the hardest part. When people feel they’re already in the game, quitting hurts more. And that’s good - not because we want pain, but because it’s how the relationship stops resetting after every transaction.
That well-known car wash study (two “free” stamps on a 10-stamp card) is a perfect illustration: mathematically identical, behaviorally dramatically different. Why? People aren’t motivated by how far they are from the goal. They’re motivated by the feeling that they’ve already started chasing it.
And one more thing companies often miss: after a reward, motivation drops. If your program is built on “reset to zero,” you’ve baked in a moment where the customer says, “Alright, I got mine - moving on.” With a rolling, carryover logic, people stay in the game.
If you run loyalty as a financial instrument, you’ll get a financial outcome: margin erosion and customers who stick around only as long as the discount does. If you run it as a behavioral product, you get something more resilient: habit, belonging, and a reason to return that competitors can’t copy as easily - because it’s not in the discount, it’s in the experience.
Three questions I’d bring to the next check-in are painfully simple:
If your loyalty program only gives back margin, you’re an ATM. An ATM is useful, but it’s not a basis for a relationship. And relationships are what survive price wars.
If you’d like, let’s talk. I’d love to have a short chat - and together we’ll look at whether your program is truly creating incremental behavior, or simply paying for what was already there.