What Surveillance Pricing Gets Wrong — And How Loyalty Marketing Can Do Better
For most of retail history, pricing was public by default. A product had one price, and it was printed underneath it for everyone to see.
Dynamic pricing has always existed at the margins (think airline seats, hotel rooms and holiday surge fares), but it responded to conditions everyone could observe: High demand, limited supply or a holiday weekend.
Then online shopping changed that. All of the consumer's actions became trackable. That shift has a name: Surveillance pricing.
Most people don't even think to ask about different prices for the same item because, for most of retail history, they didn't have to. That's the world surveillance pricing is replacing. And the industry I work in sits uncomfortably close to it.
This isn't someone else's ethics problem. It's ours as well. Loyalty programs sit on exactly the kind of shopper data that makes this possible. So before we get defensive about the comparison, I'd rather look straight at it.
How Far the Data Goes
The scope is bigger than most shoppers realize. In 2025, the Federal Trade Commission found that companies supplying pricing technology to retailers could fold remarkably granular consumer signals into pricing and promotion decisions. Researchers at the Groundwork Collaborative went further, documenting how these systems can generate inferred traits a shopper never disclosed: Emotional state, purchase intent, willingness to pay and even likely eligibility for food assistance.
A class-action lawsuit filed this year alleges The Washington Post didn't just price renewals based on tenure or subscription status, but built individual profiles from readers' opened articles, which headlines drew attention, and, for subscribers connected through Amazon, layered shopping and demographic data on top. One subscriber cited in the suit said their renewal jumped from $170 to $260 in a year, while another paid $60 for the same subscription.
Is It Always Bad? Not Necessarily
I understand why brands are drawn to personalization. Done well, it can make offers more relevant, reduce waste and create real value for shoppers.
Personalized pricing genuinely benefits some shoppers. One widely cited study from the University of Chicago Booth School of Business found that more than 60% of shoppers received a lower price under personalized pricing than they would have under a single uniform price.
The technology itself isn't the villain. The problem is that shoppers have no way to know which side of that split they landed on, whether the technology helped or worked against them, or why. Even when it works in someone's favor, they don't get to see the mechanism or participate in the decision. That opacity is the problem. And it's exactly where loyalty marketers need to draw a line, because we're building programs from the same raw material.
Give Versus Take
The distinction I keep coming back to is simple to state and genuinely hard to hold onto in practice: Are you using what you know about someone to give them something, or to take something from them?
Take a straightforward example. If I know a shopper buys dog food every three weeks, I can use that to remind them before they run out. Or I can use it to decide they'll pay whatever price I set, because they have a dog that needs feeding regardless. It's the same data point. What changes is who it benefits.
That's the uncomfortable overlap between the two disciplines: Surveillance pricing and good loyalty marketing can start from the same observation that this person is likely to buy. One instinct says charge them more for the certainty. The other says give them something new, since their basics are already covered. Same fact, completely different response.
I'm not arguing that brands should know less. I'm arguing that they should do better with what they know. Challenges built on purchase history, genuinely fun game mechanics, and surfacing something a shopper would love but hasn't found yet are all part of good loyalty marketing. The data should tell a brand what to offer next, not whether to charge more for what someone's already committed to buying.
What This Means for Building Loyalty Programs
I'll say the obvious thing plainly: Loyalty programs are exactly the kind of system people are right to be wary of in this conversation. Most loyalty marketers aren't trying to use data against shoppers. They're trying to create something useful. I know that because this isn't theoretical for me; it's the work. But good intentions don't eliminate the responsibility to account for every piece of data we collect.
The standard I hold our own work to is to build almost entirely from zero- and first-party data: What shoppers tell us on purpose, and what we or a brand partner observe through the actual relationship, rather than what shoppers infer or buy. It's not just a more defensible position; it's a more accurate one. If a shopper tells you about their dog, you never have to guess whether one exists.
In practice, that means preference centers, signup quizzes, surveys attached to specific rewards, receipt uploads, sweepstakes, polls and profiles that fill in gradually across visits. None of that happens by accident. Nobody uploads a receipt for the thrill of it.
It also shapes what "promotion" means to us. We build primarily around multipliers and bonus points: Double points for a weekend, triple points on a category a shopper hasn't tried, or a bonus tied to completing a challenge. All of it is additive. Nobody's baseline price moves. The shopper's data determines which multiplier shows up in front of them. In the worst case, they ignore it and lose nothing. This is the kind of personalization I can get behind.
A Test Worth Applying to Any Brand
Whatever the data source, there are three questions worth asking about any company using shopper information to set a price or a personalized offer:
- Can the company say plainly what data it used?
- Did the consumer knowingly provide it?
- Can the consumer say "no" to that data being collected or used, without being penalized for it?
Uber has said individual rider behavior does not determine fares. Instacart has described a pricing experiment as randomized rather than personalized. The Washington Post's algorithmic-pricing disclosure only became public after New York passed a law requiring it. In each case, the price was visible. The reasoning behind it wasn't, and in most cases still isn't, unless regulation forces disclosure.
That's the standard loyalty marketers should hold themselves to before anyone else asks us to. Loyalty data is given, not found. If a shopper chooses to tell a brand something about themselves, the least that brand owes them is clarity about what they're getting in return, and the ability to walk away without that information ever being turned against them.
Any brand running a loyalty program should be able to answer those three questions clearly and confidently. That's the bar. If it feels high, good. Loyalty asks shoppers to trust us with a lot. We should earn it.
About the Author
Jen Kunz is director of loyalty program strategy at Brandmovers. She has spent more than a decade helping Fortune 500 brands build loyalty programs and customer experiences around zero- and first-party data.
