For founders and product teams running pricing experiments, that distinction matters more than it looks. A price test can tell you a lot about willingness to pay — or it can tell you almost nothing, because what you’re really measuring is how much confusion, inattention, or lack of alternatives your customers are willing to tolerate. Learning to tell the difference is one of the more underrated skills in early-stage demand validation.
Surveillance Pricing as a Warning Light, Not a Verdict
The phrase "surveillance pricing" has moved from niche trade jargon into mainstream consumer complaints over the past couple of years, and for good reason. It describes pricing built from personal signals — location, browsing history, device type, even mouse movements — used to estimate what a specific person might tolerate paying, rather than what the market broadly supports. One widely cited analysis found that identical grocery items were priced up to 23 percent higher for certain shoppers than others, on the same platform, at the same moment.
It’s worth being precise about what that number does and doesn’t prove. It doesn’t mean every company using customer data to set prices is running a scam, and it doesn’t mean personalized pricing is inherently illegal — regulatory scrutiny is real and growing, but rules vary by market and are still being written. What the figure does illustrate is something founders can use directly: a pricing mechanism can technically "work," in the sense of extracting more revenue, while simultaneously breaking the thing that makes a sale meaningful — the customer’s ability to understand what they agreed to.
That’s a useful reframing of a well-known business idea. Dynamic pricing, in its classic form, responds to shared conditions — an airline seat gets pricier as a flight fills up, a rideshare surges during a downpour. Everyone riding that route faces the same curve. Personalized pricing is a different animal: it responds to who you are, not what’s happening in the market, and it can quietly turn two identical customers into two different transactions with two different truths. Neither approach is automatically unethical — a loyalty discount, a disclosed regional promotion, or a transparent A/B test of price points can be a perfectly legitimate way to learn about demand. The difference is whether the customer could, in principle, find out why they’re paying what they’re paying.
The Question Behind the Question
A recent essay framed this as a simple gut-check for any business: are you creating value people knowingly want to pay for, or are you good at collecting money people didn’t intend to spend? It’s a sharper version of a question every founder should already be asking about their own pricing experiments, because early-stage teams run a lot of pricing tests — different price points to different segments, regional pilots, "founder pricing" for early users, waitlist discounts. All of that is normal, and none of it is inherently suspect.
The trouble starts when a team can’t articulate, even internally, why one customer paid more than another — or when the honest answer is "because they didn’t look closely enough." That’s the moment a pricing experiment stops measuring demand and starts measuring something closer to inattention. It’s a subtle shift, and it’s easy to miss because the metric that matters most to a growth dashboard — conversion, revenue per user — often looks identical either way. A customer who pays because they clearly understand and value the offer, and a customer who pays because they didn’t notice the fine print or couldn’t compare prices easily, generate the exact same line in your spreadsheet. Only one of them is a repeatable, trustworthy signal of product-market fit.
Sorting Signal From Noise
Because the surface metrics look so similar, it helps to have a concrete way to sort a given pricing test before treating its results as evidence.
| Signal | Looks like validation | Looks like a warning sign |
|---|---|---|
| Transparency | Customers can find out why the price is what it is (loyalty tier, region, promo code) | Price varies with no discoverable reason, even on request |
| Customer awareness | Buyer knowingly opts into a segment or offer (early-bird, member pricing) | Buyer has no idea a different price exists elsewhere for the same thing |
| Consistency | Same customer, same conditions, same price across visits | Price shifts session to session with no stated cause |
| Trust impact | Customers who learn the pricing logic still feel it’s fair | Customers who learn the logic feel deceived or spied on |
| Disclosure | Terms and eligibility are stated up front, not buried | Total price or eligibility only becomes clear after commitment |
None of these rows are legal tests — they’re a practical lens for a founder deciding whether last week’s price experiment actually told them something about demand, or just about how much opacity their current customers happen to tolerate.
Reading a Test Before You Believe It
It helps to walk through the judgment in order, rather than jumping straight from "conversion went up" to "we found our price."
flowchart TD
A[Price variation observed] --> B{Can the customer learn the reason?}
B -- Yes --> C{Would most customers call it fair?}
B -- No --> D[Treat as a trust risk, not demand evidence]
C -- Yes --> E[Likely valid segmentation signal]
C -- No --> D
The branch that trips people up is the second one. It’s not enough that a reason for the price difference exists somewhere in a policy document or an internal segmentation model — the test for legitimacy is whether an ordinary customer, on reflection, would call the reason fair. A student discount is disclosed and generally accepted. A price that quietly rises because someone searched for the same flight three times in one afternoon is disclosed nowhere and accepted by almost no one who notices it.
Why a Short-Term Lift Isn’t the Finish Line
This is the part that’s easy to skip when a team is excited about a metric moving in the right direction. A pricing change that increases short-term conversion can still be a bad experiment if it does so by making the offer feel hidden, inconsistent, or unfair once customers notice — and they eventually do. Surveys on marketing trust suggest consumers are unusually alert to this kind of thing right now: a majority of consumers across a recent multi-country survey said they believe brands routinely use misleading discount tactics, and complaints about deceptive marketing have been climbing sharply in the last few years. That climate means a pricing tactic that "worked" in an isolated A/B test can still generate reputational cost the moment it’s noticed by the wrong customer, journalist, or regulator — cost that never shows up in the original experiment’s dashboard.
None of this means every price test is a trap, or that founders should treat all personalization as toxic. Segmenting customers by genuine differences — regional cost of living, verified student or nonprofit status, early-access pricing that’s clearly labeled as temporary — is a normal, often generous way to widen access to a product. The line worth watching isn’t "did you vary the price," it’s "could your customer explain, in one sentence, why theirs is what it is."
The Real Takeaway for Founders
A pricing experiment is genuinely useful only when it clarifies value rather than obscuring it. If a test lifts revenue but the underlying reasoning couldn’t survive being said out loud to the customer paying it, you haven’t validated demand — you’ve measured how much confusion your market will currently absorb. That’s a much shakier foundation to build a business on, and it tends to erode exactly when a company needs trust the most: when it’s trying to grow past its earliest, most forgiving users. None of this is legal advice, and no single test — fair or not — can prove product-market fit on its own. But asking whether a customer could knowingly say yes to your price is a cheap, early way to find out whether you’re building demand or just borrowing it.
