In the salerooms of the old auction houses there is a practice so familiar that it has a name. When a lot opens slowly, the auctioneer looks past the paddles toward the back of the room and calls a bid that no one in the room has made. The trade calls it bidding off the chandelier. It is not quite a trick, because the large houses still print, in the conditions of sale at the back of the catalogue, that the auctioneer may open and advance the bidding on the seller's behalf up to the reserve. It is a reminder of something every serious bidder eventually learns: the person holding the gavel works for the seller, and the price you pay is partly a matter of how well the house has read you.
Every marketer who buys search advertising is sitting in that room now, and most of them believe the chandelier was taken down years ago.
The Promise in Second Price
The reason they believe it is an elegant piece of economics. In 1961 William Vickrey showed that in a sealed auction where the winner pays the second-highest bid rather than her own, the best strategy is simply to bid what the item is truly worth to you. Bid less and you risk losing something you valued; bid more and you risk paying more than it was worth. The rule removes the incentive to guess, shade and bluff, because the price is set by your nearest competitor, not by you. Vickrey shared the Nobel prize for that insight in 1996, and versions of it became the foundation of the search ad business: tell the system what a click is worth, and competition, not the house, will decide what you pay.
That promise is not a technicality. It is the reason a marketing team can set a maximum bid at the honest value of a customer and walk away. The arrangement asks the bidder for the truth and in return protects her from paying for it.
What the Complaint Describes
On November 5, the Federal Trade Commission and 22 states made public previously redacted portions of their lawsuit against Amazon over its search advertising auction. The unredacted complaint alleges that Amazon used undisclosed surcharges, built on reserve prices, to raise what advertisers paid by as much as 50% above the auction result on ordinary days and as much as 250% on peak days such as Prime Day and Black Friday. It alleges that after a 2024 increase, Sponsored Products advertisers were charged their own winning bid close to 80% of the time, which the regulators describe as quietly turning a second-price auction into a first-price one. According to Search Engine Journal's reading of the filing, when one advertiser asked whether something had changed in the auction, an account manager replied that the marketplace was simply more crowded than ever.
The most instructive allegation is about intent. The complaint says Amazon concluded that advertisers were enjoying a higher return on ad spend during peak shopping events, decided they would be content with the lower return of an ordinary day, and calibrated its peak surcharges to bring the two into line, expecting the surge in shopping to "obfuscate" the change.
Amazon disputes all of it. It says no advertiser ever pays more than its bid, that the average cost per click for its sponsored search ads stayed flat after inflation from 2019 through 2024 while conversion rates rose 24% from 2021 to 2025, and that ranking ads by relevance rather than bid alone has made advertisers more money, not less. These are allegations in a case that will take years to resolve, and nothing here assumes the court will side with the regulators.
The Price of Telling the Truth
The marketing lesson does not wait for the verdict, because it is structural rather than legal. A platform that runs the auction, sells the shelf, sees every bid and measures the return on each one is not a neutral auctioneer; it is the seller with a view of your wallet. Whether or not any particular surcharge was improper, a house in that position will always be tempted to price toward the most you have said you would pay, and the most you said you would pay is exactly what Vickrey's bargain asked you to reveal.
Read that way, the peak-day allegation is a lesson in where margin goes. The extra return a brand earns on Black Friday, when shoppers arrive ready to buy, looks like the reward for being in the right place at the right moment. To a house that can set reserves, it looks like surplus left on the table. The bidder thinks the good day is hers; the auctioneer thinks it is inventory. A model that holds return on ad spend steady across the calendar is, in effect, a model that has decided your best days belong to the platform.
None of this argues for leaving a channel that still sells. It argues for bidding as though the chandelier is lit. That means setting the maximum at what a click is worth in incremental margin, measured in your own order data, and accepting that you may pay it, because a ceiling you would regret paying is not a ceiling. It means watching peak-day costs against your own conversion rates and revenue rather than against the platform's report of its own performance, since that report shows what you paid but not how the price was made. It means holding back a small, deliberate slice of spend each season so you can see what the ads add, and treating any partner that both sells you the placement and grades its results as a counterparty rather than an adviser. Black Friday is eighteen days away; this is the year to know before it arrives what your best day is worth to you, not after it is gone.
In a room where the chandelier can bid, the only number you truly control is your own. Write it down as if you will pay it, because increasingly you will.
