We use optional cookies to analyze traffic and personalize content.
Black Friday discount truth test 2026
Are Black Friday deals real? We fixed the test before Black Friday
Short answer
We will know on December 4, 2026, and we fixed how we will judge every deal before Black Friday started. On November 4, 2026 we froze 330,171 products at 80 brands' own online stores. For every Black Friday deal (a price 5% to 80% below the "was" price shown), we will check whether the Black Friday price beats what the same product cost in October and in the 30 days before.
330,171
products frozen on November 4
80
brands' own online stores
16
currencies in the panel, never mixed
4
tests for every deal, fixed in advance
Why we announce the test before the results
A discount study is easy to bend after the fact: pick other products, other dates, another way of counting. So we did the opposite. The products were frozen on November 4, 2026, and the dates, the four tests and the code that will compute and count them were fixed on November 5, 2026, three weeks before Black Friday, and fingerprinted below. By December 4, 2026 we publish the results together with the files, so anyone can check that nothing moved.
Earlier studies, such as Which?'s study of the 2024 sales (175 products at eight retailers), follow a sample of products at large retailers. This one follows brands' own online stores: every live, priced product we checked there in the 7 days before the freeze, each against its own price history.
The four tests, for every Black Friday deal
A deal is a product whose Black Friday price is 5% to 80% below the "was" price shown at that moment (the same rule as our Rival Watch Discount Check; a bigger gap is treated as a data error). The Black Friday price is the lowest price the product had from November 27 to November 30 (Cyber Monday). Differences under 1% count as rounding.
A. Lower than in October
The Black Friday price is lower than the product's price on October 1, before the season started. Products we first saw after October 1 are left out of this test.
B. The lowest price of the month
The Black Friday price is lower than every price the product had from October 28 to November 26. One window for every deal, so a sale that started before Black Friday is compared with its own earlier sale price.
C. The "was" price vs the 30 days before the price cut
The "was" price is no higher than the lowest price of the 30 days before the price cut that set the deal price, even when that cut came before Black Friday. This is the window EU law uses for the "prior price" of a price reduction. Deals whose cut we cannot date with 30 days of history before it are left out.
D. Raised, then cut
The price went up after October 1, and the Black Friday price is not lower than the October 1 price: the deal only undoes the rise. Products we first saw after October 1 are left out of this test.
What we froze on November 4
- Stores: every store RivalSweeper monitors that had a price change recorded between October 5 and November 3, 2026, and 50 or more products whose price we checked in the 7 days before the freeze: 80 of 102 stores.
- Products: every live, priced product we checked at those stores in the 7 days before the freeze, one price per product (sizes and colours share one product), chosen by our internal id, never by price, so the choice cannot favour any result.
- Same store page, same market, same currency: a product whose price source, market or currency changes after October 1 is set aside, never compared.
| Category | Stores | Products |
|---|---|---|
| Apparel & fashion | 34 | 71,985 |
| Home decor | 15 | 160,115 |
| Beauty | 12 | 6,537 |
| Jewellery & watches | 7 | 78,714 |
| Other categories | 12 | 12,820 |
| All | 80 | 330,171 |
Results by category only where a category has 15 or more stores in the frozen panel: Apparel & fashion and Home decor.
How we count
- Each store counts once. The headline for each test is the median store: we take each store's share of its deals that pass, then the middle store, so a store with 60,000 products weighs the same as one with 60. A store counts only with 10 or more deals the test can answer. We also show the share across all deals and the spread between stores.
- Price histories are rebuilt from every price change we recorded since September 1, including changes that reversed within 72 hours: a one-day price is still a price the store charged. A history whose recorded changes from October 1 on do not add up is set aside, never repaired.
- Our limit, stated now: between two checks we take a price as unchanged, and we date a change when we record it. A price that lasted less than our check interval can be missed, which makes the lowest prices we see higher than the true ones. So tests B and C can only be too kind to a deal; tests A and D can err either way.
- Hand check: before the results go out, 100 deals drawn at random are re-read by hand on the live product page, and we publish how many match.
Fingerprints
SHA-256 of each file on the day it was made (data files before compression). The files are published with the results; if a single price, product, rule or line of code in them changed, its fingerprint would not match.
Panel (330,171 products, November 4, 2026)
cab669fed494f5c648a08bb472310131cfc64295540e6ba814620adf0650122eThe method (README.md),
148aea838cd8676c6b761168c33e44638f433acc6735ea8fe4bbac4a9a356e0fThe rules for the four tests (lib.js),
094acaa4fee7659c81b2bae06eb6bd43f8c0265fdb70e03ce2bd313ec2a5a0caThe code that reads the prices (snapshot.js),
c774d9ee928e51e9cf6712440179b62b26a7215717571098608ddba2fad80019The code that counts the results (results.js),
30ab28159b857ab7733a3483f4a348cac4bbd0f0688cda210baf04c5df1765e8Baseline 1:
260,209 products with a usable price history; 69,732 set aside (price source, market or currency changed since October 1, so never compared); 230 set aside (recorded price changes that do not add up).
3fe78c7f2528cf97d7c7f548da984a112eec90800095affc77b8dcce207c7039Baseline 2:
Not made yet.
Baseline 3:
Not made yet.
Final baseline (after 00:00 UTC, clipped before any Black Friday price):
Not made yet.
Questions
Are Black Friday deals real?
We will know on December 4, 2026, and we fixed how we will judge every deal before Black Friday started. On November 4, 2026 we froze 330,171 products at 80 brands' own online stores. For every Black Friday deal (a price 5% to 80% below the "was" price shown), we will check whether the Black Friday price beats what the same product cost in October and in the 30 days before.
Are Black Friday deals really cheaper?
Often not, in studies of large retailers. In its study of the 2024 sales, the UK consumer group Which? found that 83% of the 175 products it tracked at eight retailers were the same price or cheaper at least once outside the Black Friday sale period. This test asks the question of brands' own online stores, product by product, from each product's own price history: 330,171 products at 80 stores. The results come out on December 4, 2026.
What is the 30-day lowest price rule?
In the EU, a shop that announces a price reduction must show the prior price: the lowest price it charged in the 30 days before the reduction (Price Indication Directive 98/6/EC, Article 6a, added by Directive 2019/2161). In September 2024 the EU Court of Justice ruled (case C-330/23, Aldi Süd) that the percentage off must also be calculated from that 30-day lowest price. Our test C uses the same window: the 30 days before each deal's own price cut. We report what the prices did; we do not judge any store's compliance, and the rule applies only to traders selling in the EU.
Source: Court of Justice of the EU, press release 152/24 (C-330/23)
How can I tell if a Black Friday discount is real?
Check the product's price before the sale, not the "was" price on the label. If you can, look at the same product a few weeks before Black Friday: a discount is real only if the sale price is lower than what you could have paid in October or early November. In the EU, the reference price must be the lowest price of the 30 days before the reduction.
Will you name the stores?
No. Stores appear as codes (S001 to S080) and products as codes too; brands are never named. Results are shown for all stores together and for a category only when it has 15 or more stores in the frozen panel. We never call a price "fake": we report what it did.
How do we know you did not change the method after seeing the prices?
Because the method, the code and the data were fingerprinted before Black Friday. The SHA-256 values on this page were first published on November 5, 2026, and the page is saved to the Wayback Machine each time a fingerprint is added. With the results we publish the files themselves; anyone can check that they produce exactly these fingerprints. Any change to the method is dated on this page.
Related
- Black Friday 2026 Rival Watch: when brands start Black Friday, checked weekly
- Price history and discount verification: check your competitors' discounts against their own price history
- All RivalSweeper research
Changes to the method: none since it was fixed on November 5, 2026 (the panel was frozen on November 4, 2026). Page last updated .