Hidden Floors? Amazon’s Price Trap Exposed

Photo: Silver Wings / Shutterstock

In digital advertising, the line between acceptable auction design and unlawful deception is disclosure: the same pricing tools that economists treat as routine become a legal liability when a platform’s promises to buyers do not match how its auctions actually charge them.

The Short Version

The FTC and 22 states allege Amazon secretly used “reserve pricing” and related levers that inflated ad costs while telling advertisers they were in a standard second‑price auction.
The case turns on misrepresentation and transparency, not the mere existence of reserve prices; reserves are common in ad-tech when properly disclosed.
Amazon disputes the claims, arguing the FTC misunderstands how advertisers bid and that its systems and documentation already address reserves and pricing dynamics.
Outcome will shape disclosure norms across retail media and search-style auctions, with consequences for small advertisers and the design of pricing algorithms industry‑wide.

What the lawsuit actually alleges

The Federal Trade Commission, joined by 22 states, filed a detailed complaint asserting that Amazon told more than a million advertisers—over half a million of them small and midsize businesses—that they were participating in a conventional second‑price auction, while the company quietly imposed additional price floors and adjustments that caused winners to pay more than a true second price would yield. The core mechanism at issue is an undisclosed “reserve” or price floor, allegedly introduced in or after 2019, that functioned like an extra, noncompetitive bid layered onto auction outcomes. The complaint frames this not as a novel technology, but as a deceptive mismatch between platform representations and the actual price‑setting logic advertisers faced.

Second‑price auctions are familiar: the highest bidder wins but pays just above the second‑highest bid. Reserve prices are also familiar: they set a minimum acceptable payment for an impression or click. The allegation is that Amazon represented the former while quietly tilting outcomes with the latter—raising effective prices on ads beyond the true next‑highest bid without forthright disclosure. That delta, the FTC argues, inflated advertiser spend at scale.

Why reserve prices are not the issue—disclosure is

In online advertising, reserves—and more sophisticated variants such as dynamic floors, contextual floors, and performance‑contingent minimums—are standard revenue management tools. Academic and industry research has repeatedly shown that well‑calibrated reserves increase publisher revenue and can be justified by auction theory; field experiments in sponsored search and display have documented substantial revenue lifts after adjusting reserve logic. In other words, the presence of a reserve price says little by itself about legality or fairness.

The legal hinge is alignment between what a platform tells buyers and how the system actually charges them. If you promise a second‑price auction but systematically introduce undisclosed floors that cause winners to pay more than the second bid, you have potentially crossed from optimization into deception. That is precisely the factual question the FTC has teed up for the court to resolve, supported by a long record of ad‑tech cases where the dispute is not “do reserves exist?” but “were they disclosed and applied as represented?”.

Amazon’s response and how to read it

Amazon rejects the allegations, calling the suit misguided and arguing that the agency misunderstands how advertisers operate. The company’s public documents and help center materials describe a mixed auction environment with variables that influence final cost‑per‑click or vCPM—including predicted performance, placement, and contextual signals—and state that advertisers never pay more than their authorized bids. In that framing, reserve‑like features help value inventory and guide quality, not deceive buyers; sophisticated advertisers, Amazon argues, bid to performance targets and care about return on ad spend, not the fine points of auction plumbing.

Those points are commercially plausible: performance‑based bidding is how many large advertisers behave, and retail media platforms do use quality and conversion predictions to shape both ranking and pricing. But plausibility is not the question the court will decide. The question is whether Amazon’s specific statements about auction type and pricing mechanics, made over years to more than a million paying customers, were accurate given the internal controls and floors actually applied. If the FTC can show clear contradictions—say, documents promising a second‑price rule while the system reliably charged above second price via hidden floors—the company’s general appeals to performance may not carry the day.

Mechanics: how floors and “soft reserves” alter prices

To understand the stakes, it helps to distinguish mechanisms. A hard reserve is a minimum: bids below it simply do not clear. A soft reserve, by contrast, can act as an internal price uplift—if the second‑highest competitive price is $0.65 and the soft reserve is $0.80, the platform can charge the winner $0.80 even though no competitor bid that amount. In a true second‑price auction, the clearing price is pinned to the next‑highest bid (plus a tick); with a soft reserve, the clearing price is pinned to an internal floor whenever the competitive price is lower. The effect is systematic: where demand is thin or heterogeneous, the floor bites frequently, raising average prices without any rival actually bidding higher.

Ad‑tech systems commonly vary these floors by placement, time, query category, and predicted conversion. Properly disclosed, this is legitimate yield management rooted in auction theory. Undisclosed, it creates an expectations gap: buyers think competition sets the price; in reality, the platform’s floor often does. That expectations gap is what regulators treat as deceptive if the platform has represented the auction as second‑price and left no room, contractually or in disclosures, for noncompetitive pricing uplifts.

Consequences for advertisers, especially smaller ones

Large advertisers insulate themselves from auction idiosyncrasies with disciplined testing, target CPA or ROAS frameworks, and portfolio bidding. Even so, if floors are undisclosed, their incremental cost is hard to isolate; optimizers treat it as “market price.” For small and midsize advertisers—the stated majority of Amazon’s ad customers—hidden reserves can be costlier. They lack the data volume and tooling to infer floors, and they may take platform statements about auction type at face value. If the FTC prevails, expect remedies emphasizing plain‑English disclosures, clearer pricing documentation, and possibly monetary relief calibrated to the impact period alleged in the complaint.

The industry spillover would be immediate. Retail media networks and search‑style platforms would revisit how they describe auction rules, when they employ dynamic floors, and how they express exceptions to pure second‑price clearing. Many have already migrated to first‑price or hybrid formats precisely to reduce the gap between mechanism and buyer expectations; when they do retain reserves or quality‑adjusted pricing, they publish guidance that those reserves can influence final CPCs or CPMs.

Amazon shares fall 2.5% as the FTC and 22 states file a massive antitrust lawsuit alleging over $20 billion in hidden advertising surcharges. 📊📉

Regulators accuse Amazon of secretly manipulating ad auction floors, triggering regulatory pressure across big tech digital… pic.twitter.com/yHDn1BMVvK

— 🛡️Homeboy (@JoPoYo67) September 1, 2026

What to watch next

Two signals will tell you where this goes. First, the evidentiary record: internal emails, product docs, and pricing dashboards showing how and when any floors were turned on, and what sales or marketing told advertisers at the time. Second, the remedy posture: cases centered on misrepresentation often settle with forward‑looking disclosure and monitoring obligations; a sharper remedy would suggest the court found broader unfairness. Either way, the lasting precedent is likely to be about truth‑in‑mechanism: if you advertise second‑price, you must actually clear on the second price—or say, in unambiguous terms, when and how you do not.