On this page · 11 sections
- What the FTC and 22 states filed, and when
- How the alleged surcharge works, and how Amazon describes it
- Amazon's response, in its own numbers
- Why peak shopping days sit at the centre of the case
- What you can and cannot see in your own Amazon Ads data
- Worked example: one US brand's Cyber Week maths
- Pre-sale audit checklist with dates
- Decision table: what to change before 27 November
- What the case does not settle, and the legal caveats
- How eCorpIT can help
- References
Summary. On 5 November 2026 the FTC published newly unredacted parts of the complaint it and 22 state attorneys general brought against Amazon in federal court in the Western District of Washington. The updated filing, Case 2:26-cv-03097-JHC, Document 113, alleges that a hidden soft reserve price raised Sponsored Products click prices by as much as 50% above the auction result on ordinary days and by as much as 250% on peak days such as Prime Day and Black Friday. The FTC says the unsealed sections also allege a 20% rise in average cost per click (CPC) in the 2021 holiday season, $87 million in extra worldwide revenue from 2022 holiday surcharges, and price rises above 100% for one food and beverage advertiser on Prime Day 2023. These are allegations in a pending case, not court findings. Amazon calls the suit misguided, says inflation-adjusted Sponsored Products CPCs were flat from 2019 to 2024, and says no advertiser ever pays more than its bid. That last point is the practical lesson: on either account, your maximum bid is the price you may pay. With Black Friday on 27 November and Cyber Monday on 30 November, US advertisers should baseline CPC-to-bid ratios now, reset bids to what a click is worth, cap multipliers, monitor hourly and keep every report.
About this article. For US retailers, brands and agencies running Amazon Ads through Black Friday and Cyber Monday 2026. We checked the allegations, figures and dates against the FTC's two press releases, its case page, both versions of the complaint, the New York Attorney General's release, Amazon's published response and Amazon Ads' own bidding guides on 21 November 2026. The case is pending, no court has ruled on any claim, and this is not legal advice. Worked examples use illustrative numbers, and no vendor paid for inclusion. Researched with AI assistance and checked against the primary sources listed in the references.
What the FTC and 22 states filed, and when
Three dates matter.
- 31 August 2026. The FTC and the attorneys general of Alaska, Arizona, California, Colorado, Florida, Idaho, Illinois, Indiana, Iowa, Kentucky, Louisiana, Maryland, Nebraska, New Jersey, New York, North Carolina, Oklahoma, Pennsylvania, Rhode Island, South Carolina, Vermont and Washington sued Amazon. The Commission voted 2-0 to file. The complaint alleges breaches of Section 5(a) of the FTC Act, which bars unfair or deceptive acts in commerce, and of state consumer protection laws, and asks for a permanent injunction, a monetary judgment and other relief. In that first version the surcharge percentages were blacked out.
- 28 October 2026. The updated complaint was filed with fewer redactions, according to the court header on the FTC's copy (Document 113, 181 pages).
- 5 November 2026. The FTC issued a release on the new details and added the document to its case page, which lists the matter as pending.
New York's attorney general says the coalition wants a court order to stop the practice plus penalties, restitution and other damages, and puts the alleged overcharge at more than $20 billion across more than 1.2 million advertisers. The FTC's own wording is tens of billions of dollars from more than one million brands and sellers, including over 500,000 small and medium-sized businesses.
The headline allegations, drawn from paragraphs 50 and 77 of the updated complaint and from the FTC's August and November releases:
| Allegation | Figure | Where it appears |
|---|---|---|
| Hidden surcharge above the auction price, ordinary days | Up to 50% | Complaint paragraph 50; November release |
| Hidden surcharge on days such as Prime Day or Black Friday | Up to 250% | Complaint paragraph 50; November release |
| Average CPC, 2021 holiday season | +20%, some advertisers above +200% | November release |
| Extra revenue from 2022 holiday surcharges | $87 million worldwide | November release |
| One food and beverage advertiser, Prime Day 2023 | Price rises above 100% | November release |
| Sponsored Products clicks charged at the winner's own bid | 30% to 40% (2021), 70% (2022), about 80% (2024) | August release |
| Sponsored Brands, 2024 | Amazon sets the price on 70% of clicks; winner pays its bid 50% of the time | Complaint paragraph 77 |
| Total alleged overcharge | Tens of billions (FTC); more than $20 billion (New York) | August releases |
The November release also describes a 2023 email in which the head of Amazon Ads, writing to a senior vice president, asked for every effort to close a gap to plan and suggested reserves could ideally supply at least $40 million. It also describes an account manager telling an advertiser who asked about the second-price auction that there was no change in auction style and that a more crowded marketplace explained the CPC spike. Amazon disputes how such documents are being read, as set out below.
How the alleged surcharge works, and how Amazon describes it
Amazon sells Sponsored Products, Sponsored Brands and Display placements next to search results through auctions. According to the complaint, Amazon for years described these as generalised second-price (GSP) auctions, where the winner pays one cent more than the next highest bidder. In a second-price auction, advertisers tend to bid close to what a click is worth, because the auction, not their bid, sets the price. In a first-price auction, where the winner pays its own bid, bidders lower their bids to find the minimum that still wins. Economists call that bid shading.
The complaint says that in late 2018 and early 2019 Amazon began overriding GSP prices, starting with Sponsored Brands, moving to Sponsored Products in 2019 and adding reserves to Display auctions by 2023. Internal documents quoted in it describe a two-stage calculation: first the GSP price from the auction, then a transformation to a final CPC using a soft reserve price. The FTC's August release says Amazon documents call this a proxy second price that Amazon calculates and an invented auction participant, which the FTC likens to shill bidding. The complaint adds that Amazon removed second-price wording from one of its guide pages only after learning of the FTC investigation in October 2024.
Amazon's 31 August response describes the same machinery in different terms. A hard reserve is the minimum a bid must clear to enter an auction; a soft reserve is a real-time minimum meant to reflect what a placement is worth. If the winning bid clears both, the winner pays the soft reserve. If it clears the hard reserve but not the soft one, the winner still gets the placement and pays its bid. Amazon says no advertiser pays more than its bid in any case, that its campaign console has said since 2018 that a bid is the most an advertiser could be charged, and that it updated Amazon Ads Help content to explain reserve prices after the FTC raised concerns.
So both sides agree that reserves exist and that a reserve can lift the price up to, but not past, the bid. The dispute is whether Amazon disclosed this clearly, whether advertisers bid differently because of what they were told, and whether anyone was harmed. Those are questions for the court.
Amazon's response, in its own numbers
Amazon rejects both consumer harm and advertiser harm. Its main figures, all its own and not independently verified:
- Average Sponsored Products search CPC was flat after inflation from 2019 through 2024.
- Conversion rates for individual Sponsored Products advertisers rose 24% or more from 2021 through 2025.
- Average winning bids on Sponsored Products search ads fell 50% from 2019 (the post gives 2024 as the end year in one place and 2025 in another).
- In 2024 about 92% of selected Sponsored Products ads were not the highest bid, and the mean winning bid was typically around the 128th bid by amount.
- Even if advertisers never adjusted bids, Amazon estimates they saved over $8 billion from 2021 to 2025 because relevance, not bid alone, picks the winner.
- From 2019 to 2024, 80% of bid changes on clicked Sponsored Products search ads came within one day of a prior change, which Amazon cites as evidence that advertisers bid on outcomes, not on auction descriptions.
Amazon also says the FTC picked a few low-reach training materials, that one stray email does not show a team's intent, and that it expects to make its case in court.
Why peak shopping days sit at the centre of the case
The 250% figure is the one that matters between now and Cyber Monday. The complaint's account of how peak-day pricing allegedly developed also explains why your own data cannot settle the question. According to the updated complaint:
- A 2018 internal reserve price experiment found that price rises were hidden by holiday increases in advertiser demand.
- Around Christmas and other high velocity events, Amazon often raised its maximum surcharge levels. The size of that increase is still redacted.
- In early December 2021, price rises were so sudden and large that Amazon received nearly two dozen complaints, including from big customers and agencies, in what it called internally the CPC Spike incident. The complaint says Amazon replied with an explanation blaming higher demand.
- In 2022 Amazon allegedly tested phasing surcharges in gradually before peak days, and an internal note credited this with avoiding advertiser escalations.
- After complaints that surcharges more than doubled CPCs on Prime Day 2023, Amazon put surcharge alarms and caps in place.
- The head of Sponsored Products auctions allegedly wrote that surcharge limits could be relaxed when seasonality and other confounders made bid shading chances hard to spot.
- The complaint says Amazon still phases surcharge increases in gradually and has standard scripts for escalated advertisers.
If these allegations are accurate, a rising CPC curve through Thanksgiving week (Thanksgiving is 26 November this year) would look the same in your account whether it came from more competitors or from a reserve price. If Amazon's account is accurate, it is competition plus reserves that reflect placement value. Either way, you cannot see the runner-up bid or the reserve, and Amazon itself says reserves are set in real time and cannot be predicted in advance by anyone. What you control is how high your bid ceiling sits. The same week brings the AI-referral traffic surge covered in our Black Friday 2026 AI traffic checklist.
What you can and cannot see in your own Amazon Ads data
You cannot see the GSP price, the reserve or the runner-up bid. You can see your bid, what you paid, where you paid it and what it earned. The most useful single number is the CPC-to-bid ratio: average CPC divided by the maximum bid that applied.
A ratio of 0.65 means you paid about two-thirds of your ceiling. A ratio of 0.95 or higher across many keywords means you are paying close to your bid, which is first-price behaviour in practice. The ratio does not prove a surcharge; real competition can produce the same number. It tells you that your bid is now your price, and you should set it as if it were.
Two cautions. Dynamic bidding and placement adjustments change the effective maximum, so compute the ratio against the adjusted bid where you can. And keep a control group of fixed-bid, exact-match campaigns with no placement adjustments; their ratios are the cleanest read.
Amazon offers placement reporting split by top of search, rest of search and product pages, a top-of-search impression share report and search term reports. Its response also points to Amazon Marketing Stream, which pushes hourly keyword and placement data into advertisers' own systems.
| Report or tool | What to pull | Pattern that matters |
|---|---|---|
| Sponsored Products targeting report | Daily bid, clicks, spend, CPC, orders, sales by keyword | CPC-to-bid ratio by day, 1 to 26 November versus 27 to 30 November |
| Placement report | CPC and return on ad spend (ROAS) for top of search, rest of search, product pages | Which placement's CPC jumps first and furthest on peak days |
| Search term report | Spend on terms with zero orders | Holiday queries that absorb clicks without sales |
| Top-of-search impression share | Share by keyword | Whether multipliers buy share above break-even |
| Amazon Marketing Stream | Hourly keyword and placement data | Intraday CPC steps on Black Friday and Cyber Monday |
| Billing and invoices | Billed totals by period | Reconcile billed spend to report spend and keep both |
If an agency or bid tool runs your account, ask it to export the hourly series for 20 to 30 November, not only daily totals. If you also run Google Ads, the same week is a good time for the click-quality baseline in our Google Ads invalid traffic audit.
Worked example: one US brand's Cyber Week maths
Illustrative numbers for a US home goods brand that plans 30,000 Sponsored Products clicks across the four days from Black Friday, 27 November, to Cyber Monday, 30 November.
Early November baseline (1 to 21 November). Average maximum bid on exact-match keywords: $1.60. Average CPC: $1.12, a CPC-to-bid ratio of 0.70 ($1.12 / $1.60). Conversion rate: 10%. Average order value: $48. Sales per click are $4.80 (0.10 x $48), so advertising cost of sales (ACOS) is 23.3% ($1.12 / $4.80) and ROAS is 4.29 ($4.80 / $1.12).
Cyber Week scenario. Conversion rate rises to 12% and the discount cuts average order value to $42, so sales per click are $5.04 (0.12 x $42). Bids stay at $1.60 and CPC climbs to $1.56, a ratio of 0.975. ACOS becomes 31.0% ($1.56 / $5.04) and ROAS falls to 3.23 ($5.04 / $1.56).
Across 30,000 clicks, spend is $46,800 (30,000 x $1.56) instead of $33,600 (30,000 x $1.12) at the baseline CPC, a difference of $13,200. Sales are $151,200 (30,000 x $5.04). The brand still made money, but its marginal clicks were bought at a price it never chose: it left a $1.60 ceiling in place expecting to pay about $1.10.
The fix: bid what a click is worth. If the brand's target ACOS for the event is 28%, the most it should pay per click is $1.41 ($5.04 x 0.28 = $1.4112). Setting maximum bids at about $1.41 before 27 November means that even if every click clears at the bid, ACOS stays at target. This is ordinary first-price bid discipline, and it holds whatever the court decides.
The multiplier trap. Amazon's dynamic bidding guide says the up and down strategy can raise a bid by up to 100% for all placements, so a $1.60 bid can become $3.20, and it tells advertisers to be ready to spend double the bid they enter. Placement adjustments go up to 900%, so a +200% top-of-search adjustment turns a $1.00 bid into $3.00 for that placement. Amazon's holiday guide suggests dedicated exact-match campaigns with dynamic bids up and down and a top-of-search adjustment of +200% or more. That can work, but it raises the ceiling any price can reach. A click that clears at $3.20 against $5.04 of sales is a 63.5% ACOS ($3.20 / $5.04). Use high multipliers only where conversion data supports the full adjusted bid.
Pre-sale audit checklist with dates
By Monday 23 November
- Export 60 days of daily Sponsored Products targeting, placement and search term reports, plus Sponsored Brands campaign data. The complaint's Sponsored Brands allegation (winning bid charged 50% of the time in 2024) makes that worth checking too.
- Compute CPC-to-bid ratios for your top 200 keywords by spend. Flag any already above 0.90 in November.
- For each flagged keyword, calculate a value-based maximum bid: expected conversion rate x expected average order value x target ACOS, starting from your 2025 Cyber Week data.
- Write down, in one shared document, every bidding strategy and placement adjustment on deal campaigns.
By Wednesday 25 November
- Reset maximum bids on flagged keywords to the value-based figure. Leave room for volume by keeping budgets open rather than raising bids.
- On campaigns using dynamic bids up and down, check whether a doubled bid is still profitable. If not, move them to dynamic bids down only, or fixed bids, for the four days.
- Cap top-of-search adjustments at the level your conversion data supports. Do not copy a generic +200% across campaigns.
- Set up hourly monitoring through Amazon Marketing Stream, the Amazon Ads API or your tool, with an alert when the CPC-to-bid ratio exceeds 0.95 for two hours in a row while ROAS is below target.
- Ask your Amazon account team, in writing, how reserve prices apply on Black Friday and Cyber Monday, and keep the reply.
Thursday 26 to Monday 30 November
- Review hourly. Lower bids on targets whose marginal ACOS exceeds target; raise budgets, not bids, on targets that convert below target ACOS.
- Do not chase impression share for its own sake. A lost top-of-search slot at a loss-making price is a saving.
- Save hourly exports daily so you keep the version of the data you acted on.
Tuesday 1 to Monday 7 December
- Reconcile report spend to invoices for 27 to 30 November.
- Compare event CPC-to-bid ratios with your November baseline and with 2025.
- Archive everything: reports, invoices, account-team correspondence and bid change logs.
Decision table: what to change before 27 November
| If you see this | Do this before Black Friday | Avoid |
|---|---|---|
| CPC-to-bid ratio above 0.90 in early November | Lower maximum bids to value-based levels | Raising bids to stay competitive without checking margin |
| Dynamic bids up and down on deal campaigns | Model profit at double the bid; switch to down only if it fails | Assuming the algorithm will never use the full 100% |
| Top-of-search adjustment above 100% | Keep it only on keywords with proven conversion at the adjusted bid | Copying a blanket multiplier across all campaigns |
| Budgets ran out by midday in 2025 | Raise budgets in steps from 23 November and use budget rules | Raising bids and budgets at once, which hides what drove spend |
| Agency reports only daily totals | Request hourly keyword and placement exports | Judging peak-day pricing from a weekly report |
| Sponsored Brands is a large share of spend | Run the same ratio check on Sponsored Brands | Treating the audit as Sponsored Products only |
| You plan to cut Amazon spend because of the case | Decide on marginal ROAS, not headlines | Pausing a profitable channel in its biggest week |
What the case does not settle, and the legal caveats
No court has found that Amazon did anything unlawful. The case is pending, Amazon disputes the claims, and the figures above are allegations or Amazon's own estimates. The FTC's case page lists only the two complaint filings and no trial date, and the outcome could be a dismissal, a settlement, a judgment or something in between.
There is no claims process or refund programme today. The plaintiffs seek restitution and damages among other relief, and Amazon's response notes that the proposed redress would go to advertisers, not shoppers. Whether any advertiser eventually receives money, and how it would be calculated, is unknown. Keeping your own records costs little and may help if a redress process is ever created. If you think your business was affected, or you have questions about your Amazon Ads agreement, speak to your own lawyer. This article does not tell you whether you have a claim.
The case is brought under US federal and state law. The 2022 figure in the FTC release is worldwide, but this article does not cover any action outside the United States. A separate FTC matter, on platform ad practices that help impersonation scams, takes public comments until 30 November; see our guide to the FTC impersonation rule comments.
How eCorpIT can help
eCorpIT is a Gurugram-based technology organisation, founded in 2021, assessed at CMMI Level 5 and MSME certified, with senior-led engineering teams working across AWS, Microsoft and Google platforms. Our performance marketing services team can pull your Amazon Ads keyword and placement history, compute CPC-to-bid ratios, set value-based bid ceilings and build hourly monitoring before the sale weekend. Talk to us at /contact-us/.
Last updated: 21 November 2026.