FTC Sues Amazon for Secretly Inflating Ad Auction Prices Seven Years

Full analysis

The FTC just sued Amazon for allegedly sticking its own hidden bid into its own ad auctions for seven years, jacking up what advertisers paid without telling them. The bigger news is buried at the bottom: the FTC opened a parallel investigation into Google on the same theory. If "the auction operator secretly bid against you" becomes an unfair trade practice, every closed-loop ad auction in the business is exposed, and that's most of them.

What's actually being decided: not Amazon's fate. Whether hidden reserve prices in operator-run second-price auctions become legally toxic across retail media, walled gardens, and the open web. Type 1 for the industry, hard to reverse once a court blesses the theory. The forcing function isn't the Amazon docket, which will grind for years. It's the Google investigation and the compliance scramble it triggers now.

The Market Analyst buys side won't wait for a verdict. A retail media network is a store's own ad system, where the store runs the auction and sees every bid. That's the exact setup the FTC just called illegal. In plain terms: the referee was also playing, and only the referee knew. Amazon Ads went from about $4.6B in 2018 to roughly $68B last year, so nobody's leaving over a lawsuit. But the parallel Google probe is what moves money. If auction opacity gets reclassified as an unfair practice, the independent buy-side path (The Trade Desk, independent DSPs) gets a real safety story to sell. Magnite and PubMatic sit closer to the blast radius, since their economics assume the operator can price what it can see.

The Skeptic files a lot it doesn't win. To hurt Amazon Ads it must prove intent, survive expert challenges on the damages math, and get a judge fluent in auction microstructure. Amazon's defense writes itself: a soft reserve is a floor price, floor prices are standard yield management, and the terms disclosed it at some level of abstraction. The "70 to 80% of auctions" and "50% Prime Day spike" are FTC assertions, not findings. And here's the hole in the harm story: advertisers grew Amazon spend roughly 15x during the alleged overcharging. If buyers were being fleeced, the rational move was to shift budget. They poured more in. That's a hard fact for a "consumers were harmed" narrative to walk past.

The Operator says every retail media and marketplace team should be on the phone with legal this week. The theory isn't Amazon-specific, which is the point. It hits any closed auction where the operator sees all bids before clearing. Tuesday morning the real work is boring: can you produce a document that says what your reserve price does and whether buyers were told? Most teams can't, cleanly. Agency programmatic leads running Amazon DSP should demand auction transparency reports now. Without bid-level data, reconciliation is impossible, and "we couldn't see it either" is not a fiduciary defense when a client's CFO asks where the money went. In plain terms: if you run the auction, you now need to prove you weren't the other bidder.

The CFO says model the CPC normalizing down. The settlement is a rounding error against $68B and years away. The margin question is structural. If disclosure forces reserve prices into the open, clearing prices drop toward the real runner-up bid, and effective yield per auction falls. That's the number that funds the retail media gold rush everyone's building toward. In plain terms: a chunk of retail media margin may have been a quiet tax on buyers not knowing the mechanics, and that tax gets harder to collect once the mechanics are on the table. Every RMN business case pitched at Amazon-like take rates now carries a discount nobody priced in.

Where the council splits:

1. Does the lawsuit matter, or does the Google probe? The Skeptic says the Amazon case may go nowhere and the growth numbers gut the harm story. The Market Analyst and CFO say the verdict is irrelevant, the reclassification of auction opacity is the event, and Google is where it bites.

2. Is a soft reserve fraud or standard yield management? The Operator says undisclosed operator bidding is indefensible once a client asks for the data. The Skeptic says floor prices are routine and Amazon disclosed enough abstractly to survive.

3. Will buyers actually change behavior? Nobody left Amazon over seven years of this. The Operator argues fiduciary pressure changes that now that it's public and litigated. The Skeptic argues revealed preference already answered the question: buyers care about ROAS, not mechanics.

What it hinges on: whether "operator secretly bids in its own auction" gets treated as a disclosure problem (fixable with a terms update) or a structural conflict (fixable only by opening the books). If it's the former, this is a paperwork cycle and margins hold. If it's the latter, and the Google probe suggests the FTC wants the latter, then the information asymmetry that priced closed auctions above their open-market clearing level starts eroding across the whole category.

The council leans toward "the Amazon case is slow and contestable, but the theory is contagious." The thing to de-risk is your own exposure: audit whether your auction mechanics are documented and disclosed before a client or a regulator asks. The defensible position isn't a good lawyer. It's a transparency report you can hand over without flinching.

Prediction: By December 31, 2027, the US Federal Trade Commission will file a complaint against Google alleging that it placed undisclosed bids or hidden reserve prices in ad auctions it ran.

Confidence: Medium

Why: In September 2026 the FTC sued Amazon for allegedly inserting its own secret bid into its ad auctions for seven years, and at the same time opened an investigation into Google on the identical theory. Google's exchange has the structure the complaint attacks: the operator sees every bid before setting the clearing price, and its past floor-price programs are already documented in the record of the US ad tech antitrust case. Having written the first complaint, the agency has the template, the economists and the theory of harm, so a second filing costs it far less than the first. The obvious counterargument is that the Amazon case is weak on harm, since advertisers grew Amazon ad spend from about $4.6 billion in 2018 to roughly $68 billion last year while the alleged overcharging ran, but waiting for a ruling on that would take years and the FTC's interest is in outlawing the practice rather than punishing one company, which argues for moving while the theory is live.

Revisit by 2027-12-31: We're right if The FTC files a complaint in court or an administrative action against Google on or before December 31, 2027 that alleges undisclosed bidding or undisclosed reserve pricing by Google in auctions Google operated. We're wrong if No such FTC complaint or administrative action against Google exists by December 31, 2027.