Ad Tech’s Trust Problem: What the Amazon FTC Case Teaches

Three stories landed in the first week of September, and together they say more about buyer psychology than about code. Amazon was charged by regulators over how its ad auctions actually price bids. Google dodged a breakup in its ad tech antitrust case. And The Trade Desk, the poster child of independent ad tech, cut 15% of its workforce.

Digiday’s Ad Tech Briefing frames these as evidence of an increasingly uneven playing field. We’d frame it slightly differently: this is what happens when the mental model buyers hold about a marketplace stops matching how that marketplace behaves.

What actually happened

• Amazon: The FTC and 22 state attorneys general allege that from 2019, Amazon introduced a “soft reserved price” — effectively a phantom competing bid — that pushed winning advertisers closer to their maximum bid while the system was still presented as a second-price auction. Regulators say the practice grew from 30–40% of relevant auctions in 2021 to roughly 80% by 2024, touching over a million advertisers and generating tens of billions in extra charges. Amazon calls the suit “misguided” and says inflation-adjusted Sponsored Products costs stayed broadly flat between 2019 and 2024.
• Google: Judge Leonie Brinkema rejected divestiture in the ad tech remedies ruling, opting for behavioural remedies drawn largely from Google’s own proposals — limits on self-preferencing, publisher data-sharing rules, non-discriminatory treatment of rivals. Her full reasoning stays sealed until mid-September.
• The Trade Desk: CEO Jeff Green told staff the roughly 15% reduction — around 570 roles based on its 3,843 headcount at the end of last year, at a cost of $51 million — is not about financial distress. The company cites about $1.5 billion in cash, no debt, and revenue growth from $202 million in 2016 to over $2.9 billion last year.

The psychology: expectation is the product

A second-price auction is not just a pricing mechanism. It’s a promise. Its entire appeal to advertisers is behavioural: bid your true maximum, because you’ll never actually pay it. That promise removes the anxiety of the winner’s curse and makes bidding feel safe.

Break that promise quietly and you don’t just change a number in a spreadsheet — you invalidate the buyer’s decision-making frame. Every bid they set for years was calibrated against a rule they believed applied. That’s why the allegation stings more than a straightforward price rise would. Behavioural economists call it betrayal aversion: people punish a party that violates trust far more harshly than one that simply charges more.

The same logic runs through the Google ruling. Behavioural remedies depend on continuous verification. Structural ones don’t. When you can’t verify, you have to trust — and trust in a market where the referee also owns a team is expensive to maintain.

Why this matters for brand and performance marketers

Most Indian and global marketers reading this aren’t going to litigate anything. But you are allocating budget inside systems whose rules you can’t inspect. Increasingly, as Digiday notes, AI-driven buying is the default rather than a selling point — and automated systems optimise for the platform’s economics as well as yours.

Practical moves:

• Instrument your own truth. Track win rate, effective CPM and paid-versus-bid gap yourself. If your realised price creeps toward your cap, that’s a signal, not a coincidence.
• Run holdouts. Geo or audience holdouts are the only honest arbiter of incrementality inside a black box.
• Price transparency as a vendor criterion. Ask any platform to state, in writing, its auction type and reserve logic. The willingness to answer is itself data.
• Diversify the dependency, not just the spend. Retail media, CTV and open web each carry different conflict-of-interest profiles.

The positioning lesson

Green reiterated The Trade Desk’s line: it buys media, it doesn’t own media. That is a positioning claim built entirely on the absence of a conflict of interest — the same insight that made “we don’t own inventory” a category-defining stance. In a week where the two largest players faced questions about self-dealing, that differentiation is worth more, not less. The open question is whether independence at scale can survive against ecosystems that subsidise advertising with commerce, search and devices.

For marketers, the takeaway is simple: in opaque markets, trust is a purchasing criterion. Treat it like one.

Source: Digiday