‘Frogs in the boiling water’: The Amazon-FTC case and the myth of ad auction transparency

“Business as usual” is an apt description for Amazon’s auction dynamics under FTC scrutiny. What the FTC has flagged Amazon for seems to be a gross violation of digital advertising’s auction standards, but in reality it appears to be more a latent confirmation of how opaque these auctions have become.

“We’ve been living with this already [in the programmatic display ad market], so maybe that’s why I’m not totally shocked [by Amazon’s auction dynamics],” said another agency executive.

One thing that the FTC has dinged Amazon for is a move away from the second-price auction format – in which the winning advertiser only pays $0.01 more than the runner-up advertiser’s bid – without properly notifying advertisers of the change. That lack of disclosure is not ideal, per agency executives, but it’s not being taken as that big of a deal. And that’s because so much of the digital ad market had already moved away from second-price auctions. Meta and the programmatic display ad industry had done so last decade.

This decade, meanwhile, has seen a further evolution in opaque auction dynamics. Case in point: Oftentimes advertisers don’t even know what type of auctions they’re participating in.

“We have log-level data with all of the independent DSPs we work with. And probably like 70% of the time, there isn’t data passed on the auction type,” said an agency executive.

Subscribe: Apple Podcasts • Spotify

For more on the Amazon-FTC case and the state of ad auction dynamics, check out the Sept. 8 edition of the Digiday Podcast.

Amazon offered another example of opaque auction dynamics with its introduction of “soft reserve” pricing. Per the FTC’s complaint and Amazon’s published response, Amazon implemented what was effectively a secondary floor price. The “soft reserve” price was Amazon’s way of making sure an ad wasn’t sold for so cheap that the company’s profit margin took a hit.

After an advertiser won an impression in an auction against other advertisers — an auction that had an initial “hard reserve” price floor that advertisers had to meet to even participate in the auction — then Amazon would compare the advertiser’s bid amount against Amazon’s own “soft reserve” price, which was the price Amazon set how much an impression was worth.

Here’s Amazon’s description of auction mechanics with soft reserve pricing at play:

“Advertisers bid a maximum price for a placement. When the winning advertiser’s bid exceeds both the hard and soft reserve, they pay the soft reserve, which is less than they were willing to pay. When the winning advertiser’s bid exceeds the hard reserve but doesn’t meet the soft reserve, we still grant the placement to that advertiser and they pay their bid. In no scenario does an advertiser pay more than their bid.”

Given that the advertiser doesn’t pay more than their bid, Amazon’s soft reserve pricing scheme doesn’t seem that objectionable — unless an advertiser thought they were in a pure second-price auction. In a pure second-price auction, the winning advertiser would expect to only pay a penny more than the runner-up bidder. But in this soft reserve model, Amazon would be in position to artificially inflate the price above the runner-up’s bid and without the winning advertiser being aware of the so-called “surcharge,” to use the FTC’s term.

Therein lies the rub. The FTC’s case is based on the idea that Amazon was manipulating its auction mechanics without being sufficiently forthcoming about those auction mechanics with advertisers. But no platform seems sufficiently forthcoming about its auction mechanics with advertisers. Not Amazon, not Meta, not Google, not the long list of ad tech intermediaries.

Instead ad buyers are having to make do with the levers they have available. Which are not exactly numerous.

“The old thing was like ‘maximize, maximize, find efficiency, find some pocket,’ and now the thinking is how do I train this algo,” said an agency executive.

Last month, Google updated its bidding system “to deliver more consistent and predictable performance based on the targets you set,” according to the company. The words “more consistent and predictable performance” aren’t exactly obscenities in advertising circles. Google’s changes are largely welcome, per agency executives. At the same time, the changes do limit the likelihood of ad buyers outsmarting buying platforms by configuring campaigns in ways that result in lower-than-expected costs at higher-than-expected results.

“The benefit of [Google’s bidding update] is it is predictable, and we could probably use automation more consistently to do that. But it is changing a lot of the mindset and how search marketers treated things like [Google’s AI-powered ad buying tool Performance Max] to try to outperform them,” said an agency executive.

Having said that, search advertisers have gotten accustomed to Google catching wise to ways ad buyers can outsmart its platform.

“We’ve been sort of dealing with this for a while. We’re probably the frogs in the boiling water a little bit on this and been looking at other ways of sort of dealing with this. It’s just something that we’ve gotten used to, which is why we’re so focused on, as much as we can, on the conversion signal and the qualitative conversion signal,” said the agency executive.

Meanwhile, as Meta pushed advertisers to adopt its Advantage+ platform that aims to have AI automate much of the decision-making process in ad buying, the parent company of Facebook and Instagram has been driving advertisers to focus on ad creative as an optimization lever.

“That’s looking at things like fatigue rates, being able to identify creative fatigue before that happens, content diversification, and then just creating more outlets for content development,” said an agency executive.

Whether Meta’s emphasis on advertisers providing more creative options is just a Trojan horse to acquire more examples on which to train its generative AI model until the company automates ad creative generation entirely — well, who knows. But ultimately, so long as it doesn’t hinder and actually helps campaign performance — who cares? Probably not advertisers.

“Clients come to us for performance. If there’s auction dynamic manipulation, they’re not going to care as much if the performance is there. That’s what they’re looking for,” said an agency executive. “They’re looking to get in front of the audience. They’re not coming to Meta or Amazon or Google because it’s a pure auction.”