The Discount That Gives Away Margin to Customers Who Were Buying Anyway: How to Segment Before You Discount

Almost every retailer has run this campaign at some point. An email, a push notification, a coupon offering the same discount to the entire customer base. It goes out on Monday, gets opened on Tuesday, and by Thursday there’s already a report showing open rates and redemption numbers. It looks like it worked.

What almost nobody measures is the one thing that actually matters: how many of those customers would have bought exactly the same thing, coupon or no coupon. Without RFM segmentation for retail, without looking at when a customer last bought, how often they buy, and how much they spend, that question has no answer.

It’s one of the most common mistakes in the standard retail loyalty program in Spain: a discount that should never have gone out to the entire customer base in the first place.

The symptom: the campaign that goes out to the entire customer base

The pattern repeats itself with few variations. A date arrives with commercial pressure attached: end of season, a sales target that isn’t being hit, a competitor running an aggressive promotion. And the response tends to be the same one: the same discount for everyone in the database.

No filter for purchase behavior, no look at frequency, no distinction between what a given customer usually buys or when they last bought it. There’s a date, a percentage, and a send button.

And it works, in the sense that it generates sales that week. The problem isn’t whether the campaign converts, it’s who it converts. Within that base there are at least three types of customer who will react very differently to the same coupon, and the campaign treats them all the same.

 

Why it happens

The cause is almost never negligence. It’s a structural limitation that starts long before anyone hits “send.”

What RFM segmentation for retail actually means

RFM segmentation for retail means classifying each customer according to three pieces of real purchase behavior: how recently they bought (recency), how often they buy (frequency), and how much they spend (monetary value). With those three data points, a campaign can tell the difference between a customer who was already going to buy and one who genuinely needs a nudge, instead of treating them the same way.

In many marketing teams, building a real segment depends on IT. Not a list exported once, but a group that updates itself as behavior changes. Marketing requests the segment, and someone in IT has to pull together data scattered across the POS, the ecommerce platform, and messaging channels, cross-reference it by hand, and hand back a file.

That file is a snapshot. The day it’s generated, it’s already starting to go stale, because customers keep buying, stop buying, or change their behavior while the segment stays frozen.

It’s the same problem seen from another angle: whoever runs the loyalty program day to day knows the in-store customer and the online customer are the same person, but their platform still treats them as two strangers. If the system doesn’t recognize it’s dealing with a single customer, it can’t make an informed call on whether that customer needs an incentive to buy or not.

When getting a segment costs another team days or weeks of work, not segmenting becomes the reasonable choice, at least for whoever has to get the campaign out this week. It’s faster, and it doesn’t depend on IT finding a gap between ERP maintenance and POS support.

So the laziness isn’t marketing’s, it’s built into a system that makes doing segmentation properly more expensive than not doing it at all.

The cost of not segmenting doesn’t show up on the P&L until it’s too late

That cost doesn’t appear as an obvious negative line item. It dissolves into the margin of every campaign, quarter after quarter, until someone, usually the CFO, asks why promotional spend keeps climbing without repeat purchase rates keeping pace. By then it’s an established habit, not a one-off decision that can be undone with a quick fix.

 

What’s actually happening to your margin

Within any customer base that receives a blanket discount, there are, at minimum, three groups that will react very differently to the same coupon:

• The customer who was going to buy that week anyway, discount or not, because they already had purchase intent or simply buy on that schedule.

• The customer the discount genuinely tips over the edge, because they were deciding between buying now or waiting, or between this brand and another.

• The customer who won’t buy even with a discount, because they’re simply not in a buying cycle right now, for reasons that have nothing to do with price.

A “same discount for everyone” campaign can’t tell these three groups apart, because it doesn’t have the segmentation to do so. That means the margin given up on the first group, the one that was buying anyway, is pure margin handed away for nothing in return. There’s no new sale there. It’s the same sale that was always going to happen, just at a lower margin.

Take a simple example (this isn’t real data, it’s just to visualize the mechanism): a chain sends the same-value coupon to its entire base on the weekend of a seasonal changeover. Some of those customers were going to walk into a store that same week anyway, simply because it’s their regular shopping routine.

For them, the coupon changes nothing about their decision to buy, it only lowers what they pay for something they were already going to get. The margin given up there hasn’t bought a single extra sale; at best, it’s bought the feeling of having “done something” that week.

The more this mechanic repeats, the more it trains the customer base to wait for the next promotion instead of buying at full price. The discount stops being an occasional lever and becomes the default expectation. That’s where margin erodes structurally, not just in the campaign that started it.

 

Consequences: margin, profitability, and, the surprising part, no more loyalty either

And here’s what tends to get overlooked: this kind of campaign doesn’t just cost margin, it doesn’t build real loyalty either. A customer who buys because there’s an active discount isn’t developing any distinct relationship with the Brand, they’re responding to price, the same way they’d respond to any other brand’s price. Once the discount disappears, there’s no extra reason for them to come back before going to a competitor.

The loyalty that generates real repeat business comes from somewhere else. It comes from a customer receiving something they recognize as relevant to them specifically, not to “all customers”, and from that benefit arriving at a moment that actually makes sense, rather than whenever the campaign calendar dictates.

A program that only knows how to run blanket discounts has neither of those things: no individual relevance, and no sync with actual customer behavior.

This is what separates a points card, or a generic coupon, from a loyalty program that genuinely works. The difference isn’t in the name or the technology behind it, it’s in whether the benefit is decided for each customer individually, or for the entire base at once.

How the Wapping Networks approach tackles this

The starting point isn’t “discount less.” It’s changing the question from “what discount are we sending out this week” to “what does this specific customer need for it to be worth giving them margin.”

Dynamic segmentation without depending on IT

The first piece is that the segment stops being a static file someone in IT generates by hand. A dynamic segmentupdates itself as each customer’s behavior changes, without marketing having to request it again every time. That changes the economics of the decision: if maintaining a segment doesn’t cost another team days of work, giving up on segmentation stops being the reasonable choice. Segmentation goes from being an occasional technical project to the default behavior of every campaign.

Applying the benefit at checkout, in real time

The second shift is about where and when the benefit gets applied. Instead of a coupon sent out in advance to the entire base, one that anyone can use whether they need it or not, the benefit is calculated and applied at the moment of payment, checking that customer’s data in real time. This is what Wapping Networks calls a transactional gateway: the platform resolves at checkout whether that specific customer should receive a benefit and what kind, instead of applying the same rule to everyone in advance.

 

Mechanics that build repeat business without giving away margin

Not all loyalty-building work has to happen through price. Some mechanics drive repeat business without touching margin the way a discount does: tier acceleration within a leveled program, well-designed gamification, or a returns wallet, which gives the customer an incentive to come back without giving up margin on the purchase they were already going to make. They reward future behavior instead of subsidizing behavior that was already going to happen.

The question you need to ask yourself

Would your program hold up to this kind of scrutiny? If the honest answer is that your last discount campaign went out to the entire base without distinguishing who needed convincing from who didn’t, the problem isn’t that particular discount. It’s that the underlying structure still doesn’t allow you to do it any other way.

The diagnosis is also the roadmap: the margin you’re giving away today to customers who were buying anyway is budget you already have — it’s just pointed in the wrong direction. It can go to the customers who genuinely need that nudge to decide. You don’t need to spend more on promotions. You need to decide, more precisely, who each one reaches.

That shift doesn’t require rebuilding the loyalty program from scratch, or pausing campaigns while you fix it. It requires segmentation to stop being a project requested once a quarter and become a permanent capability, available every time someone in marketing decides to launch something. At its core, it’s RFM segmentation for retail made operational: everything else — which mechanic to use, how much margin to give up, and to whom — becomes much easier to decide once you no longer have to guess across the entire base at once.

If you’re also interested in understanding who your best customers really are before deciding who to discount, here’s RFM matrices explained. And if you’d like to receive future analysis as it’s published, you can subscribe to the Wapping Networks newsletter.