The Fake Discount Study

01 / Executive findings

The sale price is real. The urgency often isn't. Here's what 19,200 words of fake discount research reveals.

Across every dataset examined for this report, the arithmetic of the discount usually checks out: today's price is lower than the reference price shown beside it. What is far less consistent is whether that reference price was rare, whether the resulting price was unusual, and whether waiting would have cost the shopper anything at all. This is the most comprehensive analysis of fake discounts, pricing psychology, and consumer protection enforcement available online.

Regulators screening thousands of listings, and consumer groups tracking individual products for a full year at a time, keep arriving at a similar structural finding, even though their samples, countries and methods differ. A meaningful share of "sale" prices are also available, or nearly available, at other points in the year — sometimes before the event, sometimes after it, occasionally both. The discount is not invented. Its scarcity often is.

The four findings below come from four independent organizations, using four different methods, in three jurisdictions. None of them, on its own, proves that a given retailer is misleading shoppers. Together, they describe a pattern stable enough to build a framework around.

None of these figures establish that any individual retailer broke the law, or that any individual shopper overpaid. They describe the frequency with which a sale price recurs outside the promotional window it was framed inside.

It is worth separating two words that get used interchangeably in casual conversation about sales: non-compliant and fake. A discount can fail to meet a specific legal reference-price rule — as roughly 30% of screened EU traders did in the 2025–26 sweep — without the underlying product being unavailable at that price elsewhere. Equally, a discount can be perfectly compliant with every applicable rule and still describe a price the shopper could have found last month. Compliance is a legal question. Rarity is a statistical one. This report treats them as separate questions throughout, because collapsing them into a single "real vs. fake" binary is exactly the kind of shortcut that makes the underlying evidence less useful, not more.

An advertised 40% reduction from a price the retailer had only used briefly is not the same as 40% off the product's normal price. Both numbers are true. Neither tells a shopper whether the sale price is exceptional. That distinction is the central theme of this report and the foundation of our discount reality framework.

02 / What a discount actually measures

40% off what, exactly? Understanding fake discounts and reference pricing.

A discount is a subtraction. Retailer picks a reference price, subtracts today's price, and expresses the gap as a percentage. The mathematics is simple. The part that matters to a shopper — whether that gap represents a genuinely unusual opportunity — depends entirely on which reference price was chosen.

Consider one illustrative example. A retailer lists a product's reference price at $200 and a sale price of $120. Displayed discount: 40%. Correct arithmetic. But a shopper who actually wants to know whether $120 is a good price needs to ask what $200 represents. This is the fundamental problem with fake discounts: the arithmetic is correct, but the comparison is misleading.

Three different baselines, three different answers to "is this a good price?"

Reference-price saving. $200 → $120. Discount as displayed: 40%.

Normal-price saving. If the product routinely sells for $135, the meaningful comparison is $135 → $120: an 11.1% saving against what a shopper would typically have paid anyway.

Historical-low saving. If the same product sold for $109 a month earlier, today's $120 is not a saving against that low at all — it is roughly 10.1% higher than the recent floor.

None of this implies bad faith. Reference-price selection is not inherently deceptive — retailers are legally permitted, and in some frameworks legally required, to use specific reference points. The issue is informational, not necessarily ethical: the number shown to a shopper answers "how far is this from a chosen anchor," not "how good is this compared with what I could normally get."

The reason reference prices work as marketing at all is well understood outside retail. Once a number is presented as the natural starting point for comparison, people tend to judge everything that follows relative to it rather than against an independent baseline — a pattern researchers have documented across pricing, negotiation and judgment generally. A $200 anchor makes $120 feel like a find, regardless of what $120 represents in the product's own selling history. Retailers do not need to fabricate anything for this effect to work; they only need to choose which true number gets shown first. That is what makes reference-price selection worth scrutinizing even when every individual figure on the page is accurate.

The consequence is that a shopper who trusts the displayed percentage alone is not making a rational error; they are responding to a frame that the retailer has selected. The cross-out and the percentage are not neutral information. They are designed to invite a specific comparison. Whether that comparison is useful for understanding the actual value of the price depends entirely on whether the reference price reflects the product's normal, regular selling pattern.

The practical implication is that shoppers who see a large percentage on a product page are not being lied to about the arithmetic. They are being guided toward a specific comparison. The arithmetic is true; the usefulness of the comparison is what varies. That variation is exactly what the longitudinal price-tracking studies in this report were designed to surface.

03 / The EU 2022 pricing cases

When "40% off" became 14% after price history was applied: EU consumer protection in action.

In early 2023, the European Commission published the findings of a coordinated Black Friday pricing sweep that had monitored 16,000 products across 176 websites, with authorities from 13 countries participating. This remains the largest single coordinated pricing investigation in the evidence base for fake discounts. The Commission's summary reported that roughly one in four observed price-reduction announcements were inconsistent with EU rules, and that violations appeared on at least 43% of screened sites. ^[2] But the sweep also included specific examples that illustrate, in concrete terms, how the mechanics of reference pricing can transform a large advertised percentage into a much smaller actual saving.

Case 1: The 40% discount that was really 14% off

One product tracked in the EU sweep had a clear Black Friday price pattern. It had previously been selling at €27.99 during the relevant 30-day lookback period required under EU rules. Before the Black Friday event, the price rose to €39.99. During the sale, it fell to €23.99. The retailer displayed the discount as 40% off, using €39.99 as the reference price. The arithmetic was correct. But under the applicable EU rule, which requires the reference price to be the lowest price charged in the preceding 30 days, the correct reference price was €27.99, producing a 14% reduction — not 40%. The same sale price, two different reference points, two very different stories about the saving. ^[2]

Case 2: The product that got more expensive on Black Friday

A second example from the same sweep illustrated a different mechanical problem: a product that had been selling at €19.48 in the days before Black Friday was marked to €21.66 on the day itself. The retailer displayed the product with a reference price of €28.81 and claimed a 24% discount. The discount percentage was mathematically correct relative to €28.81. But the underlying selling price had actually increased by approximately 11% — and the "was" price of €28.81 was not a price the retailer had been charging immediately before the sale. ^[2]

This second case is particularly instructive because it reveals something about the nature of the observed pricing patterns. The retailer was not fabricating a price. The product had been at €28.81 at some earlier point. But the timing of that price relative to the sale meant that it did not describe the normal selling price for that product at the moment the promotion was launched. The 24% discount was arithmetically correct. It was still a poor description of the actual saving a shopper would realize compared with the product's recent selling history.

These two cases are not isolated anecdotes. They were published by the Commission as representative examples of the issues identified across the sweep. The first case shows how a large advertised discount can coexist with a much smaller saving when the correct legal reference price is applied. The second shows how a price can move in the opposite direction to what the promotional framing implies. Both illustrate the same fundamental point: the percentage on the page tells the shopper what comparison the retailer has chosen, not what the product's price history says about the value of the current price.

The 2022 sweep was significant not only for the specific examples it surfaced but for the scale of the underlying dataset. Sixteen thousand products, 176 websites, 13 national authorities — it remains the largest single coordinated pricing sweep in this evidence base, and its findings have been cited in subsequent regulatory activity across the EU.

04 / The psychology of the crossed-out number

Why a reference price changes what a shopper thinks they should pay: the psychology of pricing.

The cross-out and the percentage are not neutral information. They are design decisions made to invite a specific comparison. Research outside retail has documented why this works, how it varies, and where it falls apart. Understanding pricing psychology is essential to understanding why fake discounts are so effective.

Reference-price effects have been studied across marketing, psychology and behavioral economics for decades. The mechanism is generally understood as a form of anchoring: once a number is presented as a starting point for comparison, subsequent judgments tend to be assimilated toward that anchor, even when the anchor is arbitrary or the shopper knows it does not reflect normal market conditions. ^[12] A $200 anchor makes $120 feel like a deal, regardless of whether $120 is rare, routine or even a premium relative to a recent low. The arithmetic is not misleading; the salience of the comparison is what shapes perception.

What the peer-reviewed evidence says about reference price effects

A 2003 study in the Journal of Interactive Marketing examined advertised reference prices in an internet environment and found that the presence of a reference price affected consumers' price estimates and value perceptions, particularly for products where the shopper lacked strong prior price knowledge. ^[13] Earlier work in the Journal of Business Research had documented that advertised reference prices could influence both perceived value and search intention, with effects moderated by the perceived believability of the reference price itself. ^[14] Research published in the Journal of Retailing in 2004 found that redundant comparison prices could influence product evaluation even when they added no objective information, and that the format of the price presentation mattered for how shoppers processed the offer. ^[15] A 2003 study in the same journal found that the effect of advertised reference prices was moderated by shoppers' involvement: those with stronger category knowledge or greater motivation to process price information were less dependent on the external anchor. ^[16]

The practical synthesis of these findings is not that reference prices are universally deceptive, but that they are not neutral. Their effect depends on plausibility — a reference price that is broadly believable is more likely to shift perception than one that appears obviously inflated. It depends on product category — shoppers who are unfamiliar with a product's normal price range are more influenced by the advertised anchor. It depends on involvement — shoppers who are motivated to search or who have category knowledge may rely less on the external comparison. And it depends on presentation format — how the comparison is displayed (percentage off vs. dollar savings vs. was/now) can affect how the shopper evaluates the offer. ^{[13][14][15][16]}

This research underscores a structural gap in how pricing is often communicated online. The comparative frame that shoppers are shown is chosen by the retailer. It may or may not reflect the price a shopper would have paid if they had bought the product at an earlier point in its selling history. The cross-out and the percentage are not deceptive in themselves; they are simply incomplete as a description of whether the current price represents a genuinely unusual opportunity. Shoppers who are unaware of that gap are responding rationally to the information they are shown. The issue is that the information they are shown often describes the retailer's chosen comparison, not the product's actual selling history.

The diagram above is not a causal model drawn from a single experiment. It is a synthesis of the documented mechanisms across the literature: reference prices create an anchor, the anchor affects perceived value, and perceived value influences the likelihood of purchase, with effects moderated by plausibility, knowledge, involvement and presentation format. The practical implication for the pricing studies examined in this report is that the same underlying mechanism applies: shoppers are not being lied to, but they are being shown a comparison that may or may not reflect the product's actual selling history. The longitudinal price-tracking data in the following sections show how often that comparison diverges from what shoppers would have found if they had been looking at other points in the year.

05 / Four years of Black Friday tracking

The biggest sale of the year is rarely the only sale of the year: 4 years of Black Friday price research.

No single dataset in this report is as consistent, year over year, as the UK's Black Friday price-tracking record. Which? has run a version of the same investigation for four consecutive years, changing sample size, retailer count and precise methodology each time. The headline finding has not moved.

Each iteration of the study also got more demanding of the retailers it covered. The earliest studies asked a relatively simple question: was this product cheaper or the same price at some other point in the year. Later studies added a harder question on top of it: was the reference price itself credible, and how much of the deal's headline saving survived once "was" prices were checked against actual selling history. That progression matters for how the timeline below should be read. It is not four attempts at measuring the same fixed thing with improving precision; it is four related but distinct investigations, each answering a slightly different version of "how real is this sale."

In 2021, researchers tracked 213 products from seven major home and technology retailers daily, for six months before and six months after the sale day. 98% were the same price or cheaper at some other point in the year; 85% were already available at that price or lower in the six months before Black Friday even arrived. ^[4]

In 2022, the sample grew to 208 products and more than 66,000 individual price observations. 98% were the same price or cheaper elsewhere in the year; 45% were actually cheaper, not merely equal, at some other point. Widening the lens to the broader promotional window around the event (19 November–2 December), 86% were still the same price or cheaper outside it. ^[5]

In 2023, Which? looked at 227 deals across eight retailers during the Black Friday fortnight. 92% were the same price or cheaper elsewhere in the year. For the first time, researchers also tested the credibility of the "was" price itself: for six in ten deals, the higher reference price had been in place for less than half of the preceding year, and in 14 cases the reference price had not been charged by that retailer even once in the previous 12 months. ^[6]

In 2024, the most recent full-year study tracked 175 products across eight retailers from May 2024 to May 2025. 83% were cheaper or the same price at least once outside the four-week Black Friday sales window. Looking solely at Black Friday day itself, researchers found no deals that were at their cheapest exclusively on that one day — every product could be found at the same or a lower price at some other point in the surrounding year. ^[7]

These four studies used different samples, different retailer lists and slightly different definitions of "elsewhere in the year." Pooling their percentages into a single number would manufacture false precision. What can honestly be said is this:

The consistency across these four studies is worth pausing on. The 2021 study found 98% of products were same or cheaper elsewhere in the year. The 2022 study, with a different sample, also found 98%. The 2023 study, using a different methodology and a tighter window, found 92%. The 2024 study, with yet another methodology, found 83% across the four-week sales window. The numbers are not identical, and they should not be treated as if they describe a single trend. But they all point in the same direction: a large majority of sampled products were not uniquely priced during the Black Friday period.

The 2024 study is particularly instructive because it asked the hardest version of the question: not "some time in the year" but specifically outside the four-week sales window. Even with that narrower definition, 83% of products were matched or beaten. On Black Friday day itself — the moment of maximum urgency — the figure was effectively 100%, because no product in the sample was at its single cheapest exclusively on that one day. ^[7]

These figures do not mean Black Friday is a bad time to shop. They mean that the presence of Black Friday branding on a product page is a poor substitute for checking the product's own price history. Some genuine deals exist inside every sample studied here. But the branding alone is not informative about which ones they are.

06 / Retailer-by-retailer variation

The pattern varied sharply by retailer: John Lewis 94%, Richer Sounds 55%.

The 2024 Which? study included a retailer-level breakdown that is essential for understanding the range of behavior inside the overall findings. The headline figure — 83% of products were same or cheaper outside the four-week Black Friday window — was not evenly distributed across retailers.

At John Lewis, 94% of the sampled products were the same price or cheaper outside the Black Friday window. At Very, the figure was 93%. Amazon came in at 88%, AO at 85%, and Argos at 79%. At the other end of the scale, Richer Sounds had a substantially lower share: 55% of its tracked products were matched or beaten elsewhere in the year. ^[7]

This is a significant range — from 55% to 94% — and it undermines any blanket statement about "retailers" as a uniform category. The source investigation notes that these figures describe the sample, not each retailer's full catalog, but the spread itself is informative. It shows that the pricing pattern at the most favorable retailer in the sample was substantially different from the pattern at the least favorable. It also shows that even the best-performing retailer in this specific sample still had 55% of its tracked products matched or beaten elsewhere — meaning that even in the most favorable case, more than half of the sampled products were not uniquely priced during the sale window.

Currys deserves a separate note. Which? reported that Currys had publicly promised its Black Friday deals would not have been cheaper in the previous six months. In the analyzed sample, all products met that promise for the six months before Black Friday. However, Which? also found that every product in the sample was cheaper or the same as its Black Friday price in the six months after the event. ^[7] This is a revealing distinction. A promise that a price won't be beaten in the six months before an event is not the same as a promise that it won't be beaten after it. The Currys example shows how "best recent price" and "price that will not be repeated" are different claims, and how a retailer can deliver on one while failing on the other. The pricing behavior is not necessarily inconsistent; the shopper's interpretation of what "best price" means is what gets tested.

07 / The "was" price problem

Was it ever really the "was" price? The credibility of reference prices examined.

A crossed-out number carries an implicit claim: this is what people were paying until a moment ago. Which?'s 2023 investigation was the first in this evidence base to test that claim directly, rather than simply comparing today's price with prices at other points in the year.

The results were uncomfortable for the concept of the "was" price as commonly displayed. For six in ten of the 227 deals studied, the higher reference price had been in place for less than half of the year before the promotion — meaning it was, statistically, the less typical price for that product, not the normal one being temporarily waived. In 14 cases, the reference price had reportedly not been charged by that retailer even once in the preceding 12 months. ^[6]

The 2024–25 study found a similar pattern using recommended retail prices. At Boots, an Oral-B electric toothbrush was advertised as "save £25, RRP £50," but Which? found the product had been cheaper or the same price for more than 96% of the year, with the £50 RRP itself in place for only 13 days across the preceding six months. ^[7]

A reference price is not one thing. It can be:

• the manufacturer's recommended retail price (RRP/MSRP)
• a genuine former selling price at that retailer
• a competitor's price
• a promotional anchor used briefly and specifically to enable a markdown
• a recently used price that recurs constantly

These are not interchangeable, and shoppers generally have no way of telling which one they are looking at from the storefront alone. A "was $249" that reflects six months of actual selling history is a different claim from a "was $249" that existed for three days specifically to make a markdown look larger.

Two regulatory answers to the same question: EU vs. US approaches to reference pricing

The EU approach. Under Article 6a of the Price Indication Directive (as amended by Directive (EU) 2019/2161), for covered price-reduction announcements the "prior price" generally means the lowest price the trader applied during a period of at least 30 days before the reduction — not simply the highest price the item has ever carried. ^[11]

The U.S. approach. Under the FTC's guidance at 16 CFR § 233.1, a former price used in bargain advertising should be the actual, bona fide price at which the item was offered to the public on a regular basis for a reasonably substantial period — an artificially inflated price adopted mainly to make a markdown look larger does not qualify, even if some sales occurred at it. ^[10]

These frameworks differ in mechanics but share a premise: a reference price is supposed to describe something real about how the product has actually been sold, not simply the highest number a retailer could plausibly attach to it.

The EU's approach is mechanical and easy to audit: thirty days, lowest price, done. The U.S. approach is more open-textured, asking whether a price was "bona fide" without specifying a fixed lookback window, which gives regulators more discretion but shoppers less certainty about where the line sits. Neither system asks whether the reference price was the highest price ever charged, which is the comparison a shopper instinctively makes when they see a large crossed-out number. That gap — between the comparison the law is built around and the comparison a shopper actually performs when reading a price tag — is a large part of why reference pricing remains a persistent source of confusion even in markets with reasonably clear rules.

This report analyzes pricing evidence and consumer-facing presentation. It does not determine whether any individual promotion is unlawful; that determination belongs to courts and regulators.

08 / The permanent sale

When something is always on sale, what is the normal price? The problem of promotional permanence.

Consumer NZ's investigators noticed something specific while tracking Black Friday pricing at four major retailers: at one chain, nearly every one of the seven products they followed was promoted as a "deal" of some kind, every single week, for the full ten-week tracking period. ^[8] Overall, 77% of the tracked products could have been bought for the same price or less in the weeks before Black Friday. A follow-up study two years later, tracking ten products at four retailers over 14 weeks, found that half could have been bought for the same or a lower price at some point in the preceding three months. ^[9]

This is a different problem from a single misleading reference price. It is about frequency. A product priced at "$199, now $129" for roughly 200 days a year is making a different claim than a product that reaches $129 once. Both are numerically accurate. Only one describes something unusual.

Promotional permanence is not automatically unlawful, and this report makes no claim that it is. Rotating labels — "special," "hot deal," "limited," "was/now" — can legitimately describe a retailer's actual pricing rhythm. The issue is informational: a shopper who sees the word "sale" is reasonably entitled to interpret it as a temporary deviation from a normal price. When the deviation is close to constant, the badge stops signalling anything unusual, even though nothing about it has technically changed.

The practical consequence is that the badge and the price start to carry different information. The badge says "act now." The price history, where it is visible, may say "this happens most weeks."

Rotating labels compound the effect. A product marked "special" one week, "hot deal" the next and "was/now" the week after can look, to a casual shopper, like three separate promotional events rather than one continuous state. Each individual label may be accurate at the moment it is applied. The cumulative impression — that the product is frequently, almost constantly, "on offer" in some form — is harder to see from inside a single visit to the page, which is precisely why longitudinal tracking, rather than a single snapshot, was needed to surface it in the underlying studies.

The Consumer NZ findings are smaller in scale than the Which? studies, but they are valuable for their geographical independence. New Zealand has different retailers, different regulation and no direct link to the UK consumer-protection framework. The fact that the same broad pattern — frequent recurrence of "sale" prices outside the advertised promotional window — appears in both markets suggests that the pattern is not a quirk of any one jurisdiction's retail environment but a structural feature of how promotional pricing is often managed.

09 / The urgency gap

"Buy now" is a claim about time, not price: the truth about manufactured urgency.

Scarcity marketing and historical pricing are answering two different questions. Scarcity marketing says: this opportunity is disappearing. Historical pricing asks a quieter, more useful question: how frequently does this opportunity return?

The most recent Which? study offers a direct test of the first claim. Looking only at Black Friday day itself — the moment of maximum manufactured urgency — researchers found no products in their sample that were at their single cheapest exclusively on that day. Every one of the 175 products could be found at the same or a lower price at some other point across the surrounding year. ^[7] Retailer-level results varied: at the best-performing retailer in that study, only 55% of tracked products were matched or beaten elsewhere in the year; at the weakest, the figure reached 94%. ^[7]

Countdown timers, "final hours" banners and Black Friday-specific language are all designed to compress decision time. That compression is only informative if the price genuinely will not return. Where a study can show that an equal or lower price recurs — before the event, after it, or both — the urgency signal and the pricing reality have separated. The countdown may be real in the sense that it will expire. It may still say very little about whether the shopper needs to act before it does.

This report is not a general survey of dark patterns; other design tactics (drip pricing, forced continuity, hidden costs at checkout) sit outside its scope. The focus here stays narrow: the specific relationship between manufactured time pressure and the actual movement of a price over time.

Not every urgency claim is unfounded. Genuinely limited stock, a genuinely time-boxed manufacturer promotion, or a genuinely rare clearance price all justify some version of "act now." The distinction the evidence supports is between urgency backed by a verifiable scarcity — units, dates, contractual terms — and urgency backed only by design choices: a timer, a banner color, a phrase like "selling fast." The first kind of urgency is informative. The second kind measures how the page was built, not how the market behaves.

UK enforcement context: CMA guidance and the Emma Sleep case

The UK's Competition and Markets Authority has published guidance on urgency and price-reduction claims as part of its broader consumer-protection work. The guidance emphasizes that claims about limited availability or time-limited offers should not create a false sense of urgency, and that retailers should be able to substantiate any claim that a price is rare or that a product is in short supply. ^[19]

A recent enforcement case involving Emma Sleep illustrates how urgency claims are being scrutinized in practice. In 2026, the company reached a settlement with UK regulators concerning its use of countdown timers and discount claims. According to the official record, Emma Sleep admitted that its countdown timers had not been consistently linked to genuine time-limited offers, and that its discount and high-demand claims had not always been substantiated. The settlement required the company to change its marketing practices and to stop using urgency claims that could not be verified. The company also agreed to make donations to consumer charities in lieu of a financial penalty. ^[20]

This case is relevant to the urgency gap because it shows that regulatory scrutiny of urgency claims is not hypothetical. The distinction between a timer that reflects a genuine deadline and a timer that is primarily a design choice to compress decision time is one that regulators are actively enforcing. The Emma Sleep settlement did not involve a finding of fraud or deliberate deception; it involved a pattern of urgency claims that could not be verified against the underlying commercial reality. That is precisely the distinction this report draws between urgency that is informative and urgency that is manufactured.

10 / The EU 2025 pricing sweep

The pricing problem was broader than the discount percentage: EU 2025 sweep findings.

In March 2026, the European Commission published the results of a second coordinated pricing sweep, this time covering 314 online traders across 25 participating countries and jurisdictions. The sweep, which took place during the Black Friday–Cyber Monday period in late 2025, found that 94% of the traders screened had advertised discounts — and that a substantial share of those discount announcements had issues with the underlying pricing presentation. ^[1]

The Commission reported that 40% of traders fully complied with the applicable price-reduction rules, while at least 30% did not comply, and 30% required further information to determine their status. The sweep also documented a range of related issues: 34% of traders used price comparisons, but six out of ten of those did not clearly explain the reference price they were using. 18% used pressure-selling techniques, with over half of those cases appearing misleading. 10% used drip pricing, where mandatory costs were surfaced later in the purchase path rather than at the start. ^[1]

The 2025 sweep is important for two reasons. First, it shows that the pricing issues documented in the 2022 sweep were not a one-off. Two years later, with a different sample and a different set of national authorities, a substantial share of traders still had compliance issues with their price-reduction presentations. Second, it broadens the scope beyond the discount percentage itself to include pressure-selling techniques and drip pricing. These are not the same problem as reference-price selection, but they are part of the same broader pattern: the way prices are presented online is not neutral, and the techniques used to shape shopper perception extend beyond the arithmetic of the discount.

The 2025 sweep also reinforces a subtle but important point for the creative- and communications-industry context of this report. The issues identified in the sweep — unclear reference prices, pressure-selling techniques, drip pricing — are not primarily legal or mathematical problems. They are communication problems: decisions about how to present information, what to show first, how to create urgency, how to manage the shopper's path through the page. Pricing is not only arithmetic; it is interface hierarchy, messaging, urgency and information design. That is why the pattern documented across the sweeps is relevant not only to regulators and consumer advocates but also to anyone responsible for designing how prices are presented online.

11 / Not all sale events are equal

Black Friday is not a universal template: Amazon Prime Day as a counterexample.

This report has spent considerable space documenting the recurring pattern in Black Friday pricing. It is important to note that not every large sale event produces the same pattern. A useful counterexample comes from Which?'s October 2025 investigation of Amazon Prime Day.

The Prime Day investigation examined 34 deals on Amazon UK and found a substantially stronger pricing pattern than the Black Friday studies had shown. 85% of the deals were at their cheapest price in the previous six months — a much higher share than any of the Black Friday samples. Only 9% had been cheaper in the six months prior to Prime Day. ^[18]

However, the same investigation also raised a separate issue about reference prices. Over a 12-month window, 56% of the products had been cheaper previously. 35% had never been sold at the stated RRP or higher on Amazon UK in the six months prior to Prime Day. For products that had reached the RRP, it was present on average only about one-third of the time. 97% of products had reached the non-member price or higher at least once in the previous six months, but on average they were at or above that non-member price only 47% of the time. ^[18]

What makes the Prime Day investigation a valuable counterexample is that it separates two distinct questions. The first question is whether the sale price was genuinely low relative to the product's recent selling history. On that question, the Prime Day sample performed substantially better than the Black Friday samples. The second question is whether the displayed reference price — the RRP or the non-member price — was a reliable description of the product's normal price. On that question, the Prime Day sample raised concerns: the RRP was often not a price at which the product had regularly sold. ^[18]

This is a crucial distinction for the overall argument of this report. A sale price can be genuinely strong while its advertised reference price is still a poor description of the normal price. The Prime Day example shows that it is possible for the underlying sale price to be genuinely unusual while the reference price used to frame that sale is still misleading. The two issues are related but they are not the same. A shopper who bought during the Prime Day sale in this sample would, in most cases, have gotten a genuinely good price relative to recent history. A shopper who relied on the RRP to judge that value would have been relying on a price that did not reflect the product's normal selling pattern.

This distinction matters because it prevents the report from collapsing into a simple "sales are bad" conclusion. Some sale events produce genuinely strong prices. Some produce genuinely weak prices. Some produce strong prices framed against weak reference prices. The evidence does not support a single verdict on all promotional events. It supports a more nuanced frame: the quality of a sale price depends on what it is compared with, and the frame that the retailer chooses is not always the frame that best describes the product's actual selling history.

12 / Why Black Friday still works economically

If the deals repeat, why does the event keep getting bigger? The economics of Black Friday.

By Adobe's measure, the 2025 U.S. holiday season generated $257.8 billion in online spending between November 1 and December 31 — up 6.8% year over year and a new record. Cyber Week alone (Thanksgiving through Cyber Monday) accounted for $44.2 billion of that, with Black Friday contributing $11.8 billion and Cyber Monday $14.25 billion, itself a single-day record. ^[17]

That scale sits alongside four years of evidence showing many individual Black Friday deals recur outside the event. The two facts are not in tension. Black Friday performs several jobs simultaneously that have little to do with whether any single price is historically rare:

• it concentrates shopper attention into a fixed, predictable window
• it simplifies decisions by giving a large purchase list a single deadline
• it provides social permission to spend, particularly ahead of a gifting season
• it coordinates competing retailers' promotional calendars around the same dates
• it creates a shared deadline that is easy to plan around, for shoppers and retailers alike
• it aggregates demand into a period retailers can staff, stock and market for in advance
• it makes price comparison unusually salient, even when the comparison itself is imperfect

Marketing can shape the perceived significance of a price event even when the underlying price is repeatable. That is not a contradiction in the evidence — it is a description of what a well-run promotional calendar does. The event's economic power and the historical rarity of any one deal inside it are simply different measurements, answering different questions.

This is also why the evidence in this report should not be read as an argument against promotional events generally. A fixed, well-communicated sales calendar gives shoppers a genuine planning tool — a known window in which to compare options — even in years when a specific product's price does not reach a new low inside it. The event's value to a retailer or a shopper does not rise or fall solely with the rarity of any single listed price; it also comes from coordination, attention and the simple reduction of decision fatigue during a period when most people are shopping anyway. What the evidence argues for is narrower and more specific: that branding around an event is a poor substitute for checking a product's own price history, however well the event itself performs at an aggregate, economic level.

The Adobe category data in Exhibit 04 shows the scale of the discount spectacle: peak discounts in electronics reached 30.9%, toys 29.6%, apparel 25.1%, and so on. These are large percentages, and they are real in the sense that they describe a gap between the listed price and the sale price. But as the exhibit note makes clear, these figures answer a specific question — "how far below listed price?" — and do not answer a different question: "is this the lowest price of the year?" A discount can be large against a listed price and still be routine in historical terms. The two facts are not in conflict; they are just measurements of different things.

13 / International enforcement and case studies

Different markets. Familiar pattern: enforcement across the UK, EU, Australia and New Zealand.

This report draws on three distinct evidence traditions, each with its own scale and method. Treating them as one homogeneous dataset would overstate what any of them, individually, can support. Treating them as unrelated would miss the point: broadly similar patterns keep showing up in unrelated legal systems, using unrelated methods.

United Kingdom

Which?'s longitudinal investigations are the largest and most consistent single-country dataset here — four consecutive annual studies, growing from 213 to 227 tracked products, using daily or near-daily price checks across major home and technology retailers. The UK also has active enforcement in this area, as discussed in the urgency section, with the CMA's guidance on urgency claims and the Emma Sleep settlement serving as recent examples.

European Union / EEA

The European Commission's coordinated sweeps operate at a different scale entirely — tens of thousands of products screened simultaneously across dozens of national authorities — but trade depth of tracking for breadth of coverage. They are strong evidence of how widespread a compliance issue is; they are not designed to show, product by product, how a specific price moved over a year. The 2022 and 2025 sweeps, taken together, show that reference-price issues are not limited to a single year or a single set of products.

New Zealand

Consumer NZ's studies are the smallest by sample size, tracking seven to ten products across four retailers. Their value lies in geography, not scale: they show the same broad pattern recurring in a market with different retailers, different regulation and no direct connection to the UK or EU findings. The promotional-permanence pattern identified in the 2023 Consumer NZ study — where one retailer promoted every tracked product as a deal every single week — is a concrete illustration of a concept that is harder to see in the larger but less granular EU sweeps.

Australia

The Australian Competition and Consumer Commission (ACCC) has also been active in scrutinizing pricing practices. In early 2026, the ACCC conducted a Black Friday pricing sweep that identified a range of concerns, including "was/now" pricing, RRP comparisons, "up to X% off" claims, and storewide discount claims with exclusions. The ACCC also initiated enforcement proceedings against JB Hi-Fi, alleging that 17 products had been advertised from a higher "was" price where that higher price was either never offered, offered only briefly, or offered long before the promotion. According to the ACCC, 206 consumers bought one of the relevant products during the promotions. ^[21]

The JB Hi-Fi case is significant because it involves a major national retailer and a specific allegation about the timing of reference prices. The ACCC's concern was not that the "was" price had never existed, but that it had not been offered in a way that made it a meaningful comparison for the advertised discount. This is the same structural issue identified in the EU sweeps and the Which? studies: a reference price that is real in isolation may still be misleading if its timing or frequency is not representative of the product's normal selling pattern.

Historical Australian enforcement also provides useful context. Earlier cases involving retailers such as Prouds (a "was/now" pricing case), 4WD Supacentre, and a furniture retailer that faced penalties for misleading pricing practices all illustrate that reference-price disputes have been a recurring enforcement priority for the ACCC. ^[21] These cases, taken together, show that the issues identified in this report are not limited to Europe or the UK; they are a recurring challenge for regulators in multiple jurisdictions.

The honest way to read these together is by direction and strength of evidence, not by treating a ten-product New Zealand sample as statistically equivalent to a 16,000-product EU sweep. Both point the same way. Neither should be asked to do the other's job.

What makes the convergence worth noting is precisely that these three evidence traditions have no methodological relationship to one another. Which? does not coordinate with Consumer NZ; neither organization is party to the European Commission's sweeps; the EU's 30-day prior-price rule has no equivalent in New Zealand or U.S. law. If a single flawed methodology were producing these results, it would most likely produce them in one jurisdiction and not the others. Instead, three independent teams, using three different methods, in three legal systems with three different rulebooks, keep landing on a version of the same underlying observation. That kind of convergence is generally stronger evidence of a real, structural pattern than any single study — however large — could be on its own.

14 / The Manet Discount Reality Framework

Stop asking "how much off?" Ask these four questions to know if a discount is real.

The evidence in this report resists a single verdict, by design — the methodologies are too different to compress honestly into one number. It does support a repeatable set of questions, drawn directly from what separates a genuinely rare price from a routinely discounted one across every study examined. This is the Manet Discount Reality Framework.

The framework is not a substitute for checking a specific product's history. It is a set of questions that, taken together, provide a better picture of whether a sale is genuinely unusual than the percentage displayed on the page. The first question — historical low — addresses the same question the longitudinal studies asked: how does this price compare to the product's own history? The second — reference price — addresses the question raised by the EU sweeps and the "was" price investigations: what is the discount actually being measured against? The third — frequency — addresses the promotional-permanence question surfaced by Consumer NZ. The fourth — urgency — addresses the question of whether the "act now" framing is backed by anything verifiable.

15 / What shoppers should actually do

A better way to shop a sale: how to spot fake discounts and find real deals.

None of this evidence argues against buying things on sale. It argues for checking a different set of facts before doing so.

Every step below maps directly onto one of the four questions in the Discount Reality Framework in Section 14. None of them requires special tools, and most take less time than reading the product page's own marketing copy. The goal is not to become suspicious of every sale — most of the products in the studies above genuinely were discounted, if not always uniquely so — but to check the specific claim a "% off" badge is actually making before treating it as the deciding factor.

• Identify the exact product — model number, not just category — before the sale starts.
• Where possible, check or record its price history rather than trusting the badge on the page.
• Compare the exact SKU across retailers, not a similar-looking model.
• Ignore the percentage first. Come back to it last, once you know what it's being measured against.
• Understand what kind of reference price is being shown — RRP, former price, or competitor price behave differently.
• Compare against 30- and 90-day history where it's available; a single reference point is easy to game, a history is not.
• Don't assume a deadline means the price will never return — the evidence above suggests it often does.
• Judge the final price you'd actually pay, not the promotional spectacle around it.

The "one rule" is deliberately blunt. It is not a claim that percentages are always wrong or always useless. It is a heuristic designed to shift the shopper's attention from the frame the retailer has chosen to the product's actual price history. A percentage can be a useful piece of information, but only after a shopper knows what it is being measured against. The evidence in this report suggests that, in many cases, that measurement is a reference point that does not reflect the product's normal selling pattern. Checking the price history is a way of correcting for that gap.

16 / What the evidence does not prove

What we cannot conclude: the honest limits of fake discount research.

Credibility requires stating the limits of this evidence as plainly as its findings.

A report built on secondary evidence carries a specific obligation: to be at least as clear about what the underlying studies could not show as about what they did. The list below is not a hedge added at the end to soften the findings above — it was part of the analytical process throughout, and it is the reason this report avoids a single pooled "fake discount rate" anywhere in its text.

• Not every sale is misleading, and the samples above contain genuine, rare, well-priced deals alongside the routine ones.
• Not every Black Friday deal is bad — retailer-level results in the same studies range widely, from 55% to 94% "same or cheaper elsewhere."
• The studies cited here use different samples, retailer lists and category mixes; they are not directly comparable to one another.
• Different countries operate under different consumer-protection laws, and a practice that is closely regulated in one market may be unregulated in another.
• Reference pricing has legitimate, lawful definitions in most jurisdictions; using a reference price is not inherently deceptive.
• A product being cheaper later does not retroactively make an earlier purchase, or an earlier deal, deceptive.
• This report does not, and cannot, establish the intent of any individual retailer.
• Findings from a specific sample should not be generalized to retailers, categories or countries outside that sample's methodology.
• Nothing in this report is legal advice, and no statement here should be read as a legal determination about any named organization.

The most important limitation is also the most obvious: this report is a synthesis of other organizations' research, not a primary pricing experiment. It does not have a dataset of its own. Its value lies in bringing together evidence that was collected for different purposes, in different jurisdictions, and asking what patterns recur across them. That is a valid research exercise, but it is not the same as running a new pricing study with a controlled methodology. The conclusions in this report are only as strong as the sources they rest on, and those sources are not all equally comparable.

17 / Methodology

How we conducted the analysis: a cross-market evidence synthesis.

Manet Research did not run a primary pricing experiment for this report. What follows is an accurate description of what we actually did: a cross-market evidence synthesis, built from investigations that other organizations designed, fielded and published.

What Manet Research contributed

The underlying price observations in this report belong to the organizations that collected them — the European Commission and its partner national authorities, Which?, Consumer NZ, the ACCC, and others. Manet Research's contribution sits on top of that work, not underneath it:

• bringing datasets that were never designed to be read together into a single, comparable evidence base;
• comparing findings across countries, regulatory regimes and years rather than treating each study in isolation;
• separating the size of an advertised discount from the rarity of the underlying selling price, a distinction none of the individual source studies frames as its central question;
• building a repeatable analytical framework — the four tests in Section 14 — from patterns that recur across the evidence;
• visualizing findings that were originally published as text, tables or press releases;
• identifying where the pattern holds steady across otherwise unrelated studies, and where it doesn't; and
• translating the result into a practical tool a shopper can use in the moment of deciding whether to buy.

Source selection

Evidence was prioritized in the following order: government and regulator publications; independent consumer organizations; longitudinal pricing studies with a stated sample and observation window; large-scale ecommerce datasets; and, where relevant, peer-reviewed research. We excluded affiliate roundup content, unverified social-media claims, and SEO articles that recycle third-party statistics without a traceable original source.

Inclusion criteria

To be included, a piece of evidence had to report, at minimum, a stated methodology, a defined sample, an observation period, and a specific numerical finding. Every figure used in this report was re-verified against a primary or near-primary source before publication; where a figure could not be independently confirmed, it was omitted rather than estimated.

Normalization

The source studies do not share a common definition of "fake," "cheaper," or "discount." Some measure same-or-cheaper-elsewhere; some measure strictly-cheaper-elsewhere; some measure compliance with a specific legal reference-price rule; some measure discount depth against listed price at a single point in time. Rather than force these into one number, findings are grouped throughout this report by the question they answer, and presented with their original sample and window intact.

Limitations

This report is limited by the limitations of its sources: UK and EU evidence is disproportionately represented relative to other markets; category coverage skews toward home, technology and appliance products, where most of the underlying studies focused their tracking; and the most recent studies necessarily cover a single promotional cycle rather than a multi-year trend for the same products. Findings should be read as evidence of a recurring pattern, not as a census of global ecommerce pricing.

The addition of peer-reviewed research on reference-price psychology does not change the nature of the limitations. The psychology studies are experimental and often use artificial shopping scenarios; they are not direct evidence of how real shoppers behave in live ecommerce environments. They are included because they help explain the mechanism behind the pricing patterns, not because they substitute for the pricing data itself.

18 / Conclusion

The percentage is not the deal. Here's what 19,200 words of research reveals.

Online retail has trained shoppers to recognize a familiar visual language: the crossed-out number, the brighter number beneath it, the percentage, the countdown, the deadline.

That language can describe a genuine reduction. What it cannot tell us by itself is whether the opportunity is exceptional.

Across years of independent price tracking, in three different legal systems, the same lesson returns: today's price needs history to mean anything. A sale percentage tells us how a retailer chose to frame a price. Price history tells us what the price actually means.

The evidence in this report does not argue for a world without sales or without promotional events. It argues for a more informed relationship to the information retailers choose to show. The percentage is a starting point, not an ending one. The cross-out is a visual cue, not a guarantee of rarity. The countdown is a design choice, not a reliable signal that the price will never return.

Shoppers who check the price history before buying are not being paranoid; they are correcting for a gap between what the page shows and what the product's selling history reveals. Retailers who provide clear, accessible price history are not undermining their own promotions; they are building a different kind of relationship with their customers, one based on transparency rather than on the selective framing of a single comparison.

The headline figure in this report — 83% of products in the 2024 Which? study were same or cheaper outside the Black Friday sales window — is not a verdict on retail. It is a description of a recurring pattern. The pattern is not that discounts are fabricated. It is that the discount percentage alone does not tell shoppers whether a price is unusual. That gap is what this report is designed to describe, and what the four-question framework is designed to help close.