Most business owners do not need to become experts in digital-ad auction theory this morning. They do, however, need to understand one uncomfortable fact: the company running the auction may also control the rules, the optimization system and much of the scorecard used to tell you whether the auction worked.
That tension moved from an advertising nerd problem to a federal lawsuit on August 31. The Federal Trade Commission and a coalition of states sued Amazon, alleging the company secretly increased the prices advertisers paid in its sponsored-ad auctions for years. Amazon has denied the allegations. As of September 17, the FTC lists the case as pending.
The legal case will have to answer what Amazon actually did and whether it violated the law. Businesses have a more immediate question: how much trust should you place in a marketplace that both runs the auction and sells you the advice for winning it?
What the government alleges
The FTC says Amazon represented its ad auctions as a form of second-price auction. In plain English, an advertiser might bid up to $1 for a click, but if the next competitive price were 51 cents, the winner would normally pay around that lower competitive amount rather than the full $1.
According to the complaint, Amazon began inserting an undisclosed “soft reserve price” in 2019. The FTC says internal documents described an “invented auction participant” and a calculated proxy second price that pushed winning prices above what real advertiser competition alone would have produced.
Those are allegations, not judicial findings. The case is still pending, and Amazon is entitled to contest both the facts and the government’s interpretation.
The awkward incentive structure
Amazon is not merely selling ad space. It operates the marketplace where many of those advertisers sell products, controls the ranking environment around those products, runs the auction that prices sponsored visibility, provides automated bidding and targeting tools, and reports much of the conversion data advertisers use to judge results.
That does not prove wrongdoing. It does create an incentive structure businesses should understand.
If a platform earns more when advertisers bid more, it has an obvious financial interest in expanding advertiser spend. If the same platform can also improve targeting enough to create more value, higher spend may still be rational. The operator’s job is not to assume the platform is cheating or to assume the platform is benevolent.
The job is to measure whether the extra dollar created enough extra business.
The platform can optimize the auction. It should not be the only party allowed to grade the outcome.
What I would do if I advertised on Amazon
I would not panic, slash budgets or try to reverse-engineer seven years of auction mechanics from a lawsuit filing. That is satisfying theater, not campaign management.
I would make sure I know my own economics outside Amazon’s dashboard: contribution margin after fees and ad spend, new-to-brand customer value where it can be measured, repeat purchase behavior, organic sales dependence, total advertising cost of sales, and what happens when spend is intentionally moved up or down.
I would also run controlled tests rather than permanently accepting recommended bid increases. If a platform suggests spending 20% more, the useful question is not “did reported sales rise?” It is “did profit or incremental sales rise enough to justify the additional 20%?”
That distinction sounds obvious. Advertising dashboards have spent years proving that obvious things still need to be written down.
This lesson extends beyond Amazon
Google, Meta, Microsoft and other advertising platforms continue to move advertisers toward more automated campaign systems. That does not mean they behave like the FTC alleges Amazon behaved. It does mean advertisers are handing more operational control to the same companies that sell them media.
That makes independent measurement more important, not less.
When the system chooses more of the targeting, placement, creative combination and bidding behavior, the advertiser has fewer knobs to inspect. The remaining job becomes sharper: define the business outcome, supply reliable conversion data, preserve baselines, test incrementality where practical, and judge customer quality instead of merely accepting platform efficiency metrics.
The case may change. The incentive problem will not.
The FTC still has to prove its claims. Amazon gets to contest the evidence, the interpretation of its auction design and the claimed advertiser harm. Important parts of the complaint may be narrowed, rejected or understood differently as the case develops.
That uncertainty is exactly why the broader business lesson should not depend on the verdict.
Platforms are extraordinarily good at making complicated advertising systems easier to use. That is valuable. They are also businesses with their own revenue targets.
You do not need to distrust every recommendation. You do need a way to check it.
Because when the auctioneer also sells the paddles, recommends your bid and hands you the report card afterward, asking for your own calculator is not paranoia.
It is management.
Sources
Ready to scale your brand?
Let's talk about what growth looks like for your business.
Book a Free Strategy Call