Digital ad auctions are only as fair as the rules bidders can see; when a platform controls both the marketplace and the disclosure of its mechanics, “optimization” can become deception, and that is the crux of the FTC and 22 states’ case against Amazon’s ad business.
The Short Version
- The FTC and 22 states allege Amazon secretly inflated advertisers’ costs by layering undisclosed reserve pricing onto auctions it presented as second-price.
- The complaint centers on misrepresentation and nontransparency, not the mere existence of reserve prices, which are common in ad-tech.
- Amazon counters that its documentation discloses reserves and that the FTC misunderstands how performance-driven advertisers bid.
- The dispute’s stakes reach beyond one company: it tests where yield management ends and unfair deception begins in platform-run auctions.
What the lawsuit actually alleges, and why that matters
The FTC’s complaint, joined by 22 states, alleges that since 2019 Amazon “secretly manipulated its advertising auction system” by imposing hidden price floors—described as a soft reserve price—on auctions that advertisers were told operated as second-price, where the winner pays just above the runner-up bid. The government’s theory is straightforward: if the platform effectively replaces or overrides the second-price outcome with a higher, undisclosed floor, advertisers are charged more than the mechanism they were sold would produce. That is not a technical quibble; it goes to inducement and reliance for roughly 1.2 million advertisers, including hundreds of thousands of small and mid-sized businesses, and implies large-scale overcharges if proven.
Second-price auctions are prized because they simplify bidding strategy: you can bid your true value without fear of paying that full amount. Hidden reserves change the payoff structure. They can be legitimate tools, but only when they are part of the deal the buyer understands. The government’s complaint calls Amazon’s reserves “undisclosed” and “manipulative,” indicating the problem is not the use of floors per se but the mismatch between public description and internal pricing logic.
How these ad auctions work: second-price, floors, and “soft reserves”
In a textbook second-price auction, the highest bidder wins but pays the second-highest bid plus a minimal increment. Digital ad markets complicate that with eligibility filters, quality adjustments, and reserve prices—minimum acceptable payments for an impression or click. Reserves can be static, dynamic, or “personalized” across contexts and placements; they are used to protect inventory value and can raise publisher revenue. A broad empirical literature—field experiments and structural models alike—finds that properly calibrated reserves increase yield in sponsored search and display auctions.
Platforms often distinguish “hard” reserves (if no bid clears, nothing serves) from “soft” reserves that effectively lift the clearing price toward a target. To an advertiser, the economic effect is similar: the invoice reflects a floor above the runner-up. The policy question is disclosure. Were buyers bidding in a mechanism that, as described, implied a second-price charge, while being billed a second-price-plus-floor outcome? That is the dividing line between expectation and reality the FTC aims to litigate.
Amazon’s rebuttal: disclosure, performance, and industry norms
Amazon disputes the allegations, characterizing the case as misguided and asserting that the company discloses how auctions and reserves work. The company’s public-facing documentation references “other reserves” that can influence final cost-per-click or viewable CPM based on predicted performance, placement, and context; it also cautions that the price charged can exceed the runner-up bid, within the advertiser’s authorized maximum. That language, Amazon suggests, demonstrates that sophisticated buyers are not misled and bid to outcomes, not to simplified descriptions of mechanics.
The strength of this defense will hinge on what, exactly, was represented to advertisers over time and how salient those disclosures were. Boilerplate that “other reserves” may affect price can matter, but so can contemporaneous sales materials, self-serve UI copy, and training that framed auctions as second-price. Courts have long distinguished between technical availability of information and meaningful disclosure to a reasonable counterparty. The case will likely turn on that evidence record, not on industry custom alone.
Where the real disagreement lives: disclosure versus design
The broader ad-tech context cuts both ways. Reserve pricing is not aberrant; it is an established lever for revenue optimization and inventory control, validated by academic studies and widely implemented in search and display markets. The FTC does not claim reserves are unlawful in general. Instead, it alleges a misrepresentation: advertisers were told one pricing mechanism but charged by another layered logic that raised prices. If the record shows explicit second-price assurances while internal systems applied undisclosed floors, the government’s case gains force. If, conversely, Amazon’s disclosures were clear and salient that prices can exceed the runner-up due to reserves and adjustments, the theory of deception weakens.
That framing also clarifies damages. The question is not “did reserves increase revenue?”—they usually do. It is “did nontransparent reserves cause advertisers to pay more than they reasonably expected under the represented mechanism?” Quantifying overcharge in such cases often requires reconstructing counterfactual auction outcomes: what would the second-price charge have been, impression by impression, absent the floor. That is technically demanding but tractable with logs and expert modeling; it is where similar disputes in other platforms have found their evidentiary ballast.
Implications for advertisers and platforms
For advertisers, the lesson is to align bidding strategy with the mechanism you actually face, not the one you assume. Where documentation indicates that charges may exceed the runner-up bid due to reserves, plan accordingly: test incrementality, monitor realized clearing prices versus observed competition, and use budget pacing and bid shading to control effective cost. Demand transparency on when a bid loses to a floor rather than to another buyer; that single diagnostic can change how you set targets and measure return.
For platforms, the lesson is starker. Yield tools—reserves, floors, throttles—are legitimate, but they must be harmonized with external representations. If you say “second-price,” you either need to run one or explain, clearly and conspicuously, the conditions under which the invoice will exceed the runner-up and by how the system determines that delta. Burying that in a help-center footnote invites the very scrutiny now at issue. The academic upside of reserve pricing is not a shield against consumer protection or deceptive practices claims when disclosure lags design.
FTC And 22 States Sue Amazon Over Alleged Ad Auction Manipulation — The FTC and 22 states filed a lawsuit accusing Amazon of manipulating digital advertising auctions to drive up ad prices. The complaint alleges the company inflated costs for more than a million advertisers over…
— BizToc (@biztoc) September 1, 2026
What to watch next
Three developments will determine the arc of this dispute. First, discovery on representations: internal emails, advertiser-facing decks, UI language, and training materials about “second-price” mechanics. Second, logs and modeling: whether the data show systematic divergence between second-price outcomes and billed charges traceable to undisclosed floors. Third, remedial posture: even before final adjudication, platforms often move toward clearer disclosures, toggles that show auction diagnostics, and mandated reporting on floor effects. Each step narrows the room for misunderstanding and, by extension, for litigation risk.
Sources:
ftc.gov, finance.yahoo.com, bclplaw.com, reuters.com, cryptobriefing.com, milkeninstitute.org
