Released:
The FTC says Amazon put a bidder that does not exist into its ad auction — for seven years — so the CPCs you paid on Amazon Ads were never the price the auction produced.
Mike Ryan and Chris Scharmüller break down the FTC’s case against Amazon Ads: what Amazon is accused of doing to its own auction, what it cost advertisers, and why Amazon’s defense is not a denial.
Then the story Amazon would rather talk about — its demand-side platform plugging into ChatGPT ads, and what that does to targeting on a brand-new surface.
For in-house PPC and ecommerce teams buying on Amazon, Google and now AI surfaces.
00:00 - Amazon, the FTC and a ChatGPT deal in one week 02:00 - What Amazon is accused of doing to its ad auction 07:00 - What it cost advertisers, and why nobody will leave 13:20 - Amazon’s defense: “these ads were too efficient” 17:54 - Amazon Ads lands on ChatGPT
Key takeaways:
A second-price auction only delivers its benefit if the auctioneer is trusted. The design makes honest bidding safe; a dishonest auctioneer turns that honesty into exposure. The alleged mechanism is blunt: a bidder that does not exist, placed between the real second bid and the winner. The FTC says it ran from 2019 and extracted tens of billions. The damage per click was small by design — calibrated to stay under advertisers’ minimum acceptable return, which is why it went unnoticed for seven years. Amazon’s defense is not “we didn’t.” It is “no one was harmed, because targeting gains were pushing CPCs down anyway.” Precedent says spend does not move: nobody left Google after the DOJ hearings. Meanwhile, Amazon’s DSP brings real first-party targeting to ChatGPT ads — plugging that inventory’s biggest gap and giving OpenAI a way to monetise intent it currently leaks to Google.
Amazon Allegedly Manipulated Second-Price Auctions
The FTC has presented evidence alleging that Amazon manipulated its second-price ad auctions by inserting fake proxy bidders to artificially drive up costs for advertisers. Instead of running a transparent auction environment, Amazon systematically calibrated these fake bids to get as close as possible to the top bid without pushing advertisers beyond their break-even limits. This tactic allegedly generated billions in extra revenue for Amazon at the expense of millions of advertisers who trusted the platform’s pricing integrity. For ecommerce leaders, this revelation highlights the critical need to closely audit ad spend efficiency and maintain healthy skepticism when committing budgets to dominant walled gardens.
00:00:00
Mike:
Welcome to another episode of Growing Ecommerce. I’m one of your hosts, Mike Ryan, and with me is Chris. As always, sponsorship obligation fulfilled! Yeah, fulfilled. We will keep grinding even if we get auto-flagged as sponsored content or something, and then we’ve achieved it. Great.
00:00:24
Mike:
So, this week we are going to be talking about Amazon. We’ve got our sights set on Amazon, and so does the FTC with allegations of bid manipulation. On a positive note for Amazon, though, is an integration with everyone’s darling ad platform, ChatGPT ads. So, great topic. We’re talking finally about Amazon. It feels like it’s been a while.
00:00:50
Chris:
It has been. Yeah, we let Jeff off the hook for quite some time, but now it’s time for him. It’s a serious topic, but for sure we haven’t seen something like this since the DOJ investigations with Google.
00:01:15
Chris:
With Google, yeah. And it hurts me because I have a lot of trust in these big platforms for whatever reason—shame on me. Once again, it shows that you should have a close look at their business tactics. Can you shed some light on it, and then we can have a great discussion?
00:01:41
Mike:
I’m so glad that you have trust in the company whose motto is “your margin is our opportunity.” This is the physical instantiation of that. This is like that manifested into reality.
Chris: Don’t be too harsh on me! Already stated, shame on me. I don’t know why this trust is here.
Mike: It’s because you would never violate any trust. You’re such a trustworthy person, you know.
Chris: Depending on how much money is on the table! No, kidding, just kidding. Everyone has a price.
Mike: Everybody.
00:02:00
Chris:
Mike, what has been going on here? Probably it was not the decision of Jeff.
Mike: You never know.
Chris: Yeah, you never know. But what’s up for Amazon right now?
Mike: So the FTC alleges, and has already provided some pretty compelling documentary evidence to support, this idea that Amazon was manipulating second-price auctions.
Chris: Wow. We were talking about second-price auctions not that long ago.
00:02:29
Mike:
Yes. By the way, it’s a great methodology. It’s a great concept for an auction.
Chris: Yes, yes.
Mike: And it seems like it leaves a lot of room open to manipulate it. The problem is there’s a concept of truthful auctions where you want to elicit people’s actual valuations in their bids. You want the true bid.
Chris: Yeah.
00:02:53
Mike:
And if you have a first-price auction… But this depends on a trusting environment. What we run into is this temptation for digital ad platforms—this goes back to Google and all the way back to Yahoo—and it seems like there’s this huge temptation to just break that trust.
00:03:11
Chris:
Let’s jump into the details. What’s going on there?
Mike: Well, as we were just saying, in a second-price auction… Let me compare it to a first-price auction, which is basically how eBay works (or at least one of their auction models). Highest bid wins.
00:03:34
Mike:
The problem there is that you’re incentivized to try to read the minds of other bidders. It’s like that classic auction house scene in every movie or TV drama, where people are actually trying not to reveal how much they would bid. So in a second-price environment, you can safely volunteer your maximum bid because you will typically pay only a penny more—some nominal sum more—than the second bid.
00:04:10
Mike:
So it incentivizes you to be truthful and be more aggressive on your bids. I think that this is a win-win for the whole system. I believe that it is good for the auction participants when, in terms of maximizing their value, they have a better chance at maximizing their value when they’re honest about their valuations.
00:04:34
Mike:
And at an aggregate level, it should lift bids and be good for the platform or auctioneer as well.
Chris: Because as far as I see it, in a first-bid auction, you constantly have to calculate the tradeoff between the highest bid and your margin, right? That probably leads to a more defensive and not-so-truthful bid. On paper, it makes perfect sense. So if the environment is trustful, it’s the right thing to do.
00:05:02
Chris:
I think there’s not a lot of room for discussion here, for sure. But the question is: is the environment trustful? Because otherwise, it’s dangerous to give your true valuation to someone who can’t be trusted.
00:05:28
Mike:
So that’s the framework Amazon is operating in, and all the advertisers are operating in. And now there is a case to be made that some wacky shit was going on.
Chris: Exactly.
Mike: So the FTC continues here. Amazon represented this to advertisers as a second-price auction, and maybe there was some fine print in there about surcharges or something like that that could cause your bid to be a little bit higher.
00:05:58
Mike:
But as it turns out, they were actually putting in fake proxy bidders that were inserted between the second price and the first price.
Chris: That’s quite straightforward, I’d say.
Mike: Exactly. Because with the DOJ investigation of Google, they at least tried to hide it, right? There were these complicated systems like compressing the dynamic predicted click-through rate range, modification of ad rank… It was not that straightforward.
00:06:36
Mike:
But here, it’s like, “We’re just going to put in a fake bidder.”
Chris: Often, the bidder’s name is Jeff, by the way! Pure coincidence, but the bidder’s name is Jeff. Sorry to laugh about this, because actually it’s quite serious.
Mike: It is serious because at the end of the day, I mean, I don’t know if there are any numbers out there yet, but let’s face it: this probably led to additional costs for millions of advertisers, right?
00:07:00
Mike:
It’s actually crazy. It’s literally the third largest ad platform.
Chris: Let’s go through this on a conceptual level. Sure, the both of us bid for a running shoe. We want to sell on Amazon, and there might be situations where the first-place bid versus the second-place bid will make a substantial difference, right?
00:07:28
Chris:
So what Amazon has probably been doing is bidding very closely to the first one.
Mike: That’s right.
Chris: So you as the first-place bidder, you probably paid for hundreds of clicks for that shoe $0.15 or $0.20 more. You never know, right?
00:07:46
Mike:
Yeah, it can totally vary by case. It’s estimated that Amazon raked in billions in additional revenue because of this. But that’s exactly it—with this proxy bidder, they calibrated this carefully because they wanted to get as close to the first bid as they could without anyone noticing, basically.
00:08:12
Mike:
We talked a couple of episodes back about that concept of minimum acceptable returns, and they were probably good at anticipating that, right? They inflated the bid to that minimum acceptable return.
00:08:36
Mike:
They calibrated this, they worked on it. In some ways, it’s both their greatest defense and their greatest sin that no one noticed, actually.
Chris: When did this come to the surface?
Mike: As we speak. At the time of recording, this was in the past week or so, but when the episode airs, it’ll be a couple of weeks ago.
00:09:02
Chris:
Shocking to me, if you look at the stock price of Amazon, it’s not really moving downwards—slightly, but nothing crazy.
Mike: This is part of their business model. They just put this in the business case, to be honest. They know there’s a probability they’re going to get busted, and they think about how much they are going to spend in legal fees and a fine compared to how much they will make. Then there’s a waterline for sure.
00:09:24
Chris:
Amazon especially is a cold-calculating business. Probably this tradeoff made sense for them. I’m not even worried about the potential fine of billions of dollars—that’s already factored into the stock price.
00:09:43
Chris:
I’m rather thinking about this lack of trust in the bidding ecosystem. The end consumer probably won’t give a shit about it, right?
Mike: Well, we can talk about that.
Chris: What I want to ask you is: because this might have an impact on the stock price long-term, do you see any backlash whatsoever in terms of end consumers visiting Amazon or buying on Amazon? Or maybe advertisers are like, “Okay, I will just shift my budget somewhere else, to Google where the trusted ecosystem is”? Or where it’s not there, I don’t know!
00:10:55
Mike:
No, I don’t think consumers care about this at all unless it was affecting their wallets somehow. They don’t understand these systems, they don’t care. But we can talk in a minute about whether it touched their wallets or not.
Mike: As for the advertisers, again, they’ll be angry, but I don’t know what they can do about this realistically. It’s water under the bridge at this point. In the end, they’re going to keep spending there. We saw this with Google; no one boycotted Google or walked away from Google.
00:11:19
Chris:
I can remember during the DOJ hearings, the craziest thing back in the days was that it wasn’t really that big of a topic for a lot of big retailers. They just accepted it or didn’t really read into it. Even when we talked about it with them, it was like, “Water under the bridge, spilled milk, I don’t care, let me move forward.”
00:11:41
Mike:
Exactly. That’s the point of these market power conversations.
Chris: Yes, it was probably even bigger with regards to Google, right?
Mike: When was that breaking? I think that was September a couple of years back, and they were already like, “Well, I’ve got to hit my Q4 numbers.”
Chris: So what do you think? Yeah, I don’t like that at all, man. I really don’t like that at all.
00:12:01
Chris:
Because what really rubs me the wrong way is look at all these high margins these big companies have. Are you kidding me? Do you really have to have this fake Jeff bidder to cross a couple billions more a year?
00:12:18
Mike:
Yeah, well, for me, I’d love to find out who made the decision. There has to be a decision chain for that. It’s not a small decision, right?
Chris: No, and come on, man, I really don’t like that at all. I tell you one thing: I have never been a big Amazon buyer, but now I will think twice about buying on Amazon again.
00:12:42
Mike:
It won’t hurt them one bit!
Chris: I know!
Mike: I’ve always had my ethical or value-based qualms about shopping at Amazon, and at the end of the day, I still shop at Amazon pretty frequently. So that’s why I’m cynical when I hear topics like sustainability being so important to shoppers—to a point, people end up following the price tag so often.
00:13:27
Mike:
Speaking about Amazon segues…
Chris: What a segue!
Mike: What is the argument on Amazon’s side? Probably they have one, right?
Mike: They wrote a lengthy defense, but to me, the crux of it was not really denying that they did this, but rather saying that there’s no evidence that this harmed advertisers or consumers. We’re building toward that.
00:13:49
Mike:
Their core argument is that the improvements in targeting that were going on in Amazon Ads were so powerful that it was actually dropping CPCs and boosting return on ad spend. Technology was improving so fast that it was creating more or less a deflationary environment.
00:14:15
Mike:
I think that is exactly their motivation as well as their defense. They have to put the story on the street that Amazon Ads is growing and money was pouring in. But at the same time, at what point are these advertisers too profitable? Like I said at the start, Jeff Bezos’ classic motto: “Your margin is my opportunity.”
00:14:43
Mike:
And we’re talking about ad margins! They literally said, “We looked at it…” It can be interpreted differently now, but seriously, they were like, “These ads are too efficient, too profitable, working too well, and we are getting left behind. It’s a deflationary environment, so let’s inflate the bid.”
00:15:09
Chris:
Hell of a statement! Hell of a defense! Look, we can discuss the thesis they put on the table, which is that the end consumer didn’t get harmed. Fair enough.
00:15:31
Chris:
But there can also be a case made that if my ads are just not running as efficiently anymore because there was this fake bidder Jeff which led to a 10% higher bid, there might be some price dynamic yield management going on. Maybe you sold or bought the shoe for a couple of euros more.
00:15:53
Chris:
One thing is for sure: advertisers got harmed, right?
Mike: That’s right.
Chris: What about them? Exactly. And advertisers at the end of the day are also employees who work at these companies—they’re consumers as well.
Mike: This is the thing: the FTC is also alleging consumer harm—that basically this made advertising more expensive, which needs to be baked into prices, which hurt consumers. So ultimately, Amazon passed on this profit-taking at the expense of consumers.
00:16:21
Mike:
That will be very hard to prove at the end of the day.
Chris: Again, because I know that a lot of clients are listening here as advertisers, it is a fact that I probably paid too much for thousands of clicks, potentially, right? So the harm on the advertiser side is pretty clear. My question is: what is Amazon doing about it? Is there some money flowing back to them? I assume not, but what’s going on here?
00:16:43
Mike:
I don’t know if there’s a discussion yet about the fine amount.
Chris: Are there any statements from Amazon on how to restore trust again?
Mike: No, because they’re denying this or saying no harm was done, everything’s good, it’s just paying the fair price.
00:17:08
Chris:
Yeah, “you can trust in the system.”
Mike: Exactly. It’s too early for them to talk about restoring trust because they’re too busy saying that nothing happened, and that actually the platform worked amazingly well for advertisers, which I think it did. But yeah, it’s going to be interesting to see what happens down the line on that.
00:17:32
Chris:
Let’s see. Quite interesting.
00:17:54
Mike:
On a more positive note… Meanwhile, why would you talk about some FTC thing when it’s way more exciting to partner with ChatGPT?
Chris: Yeah, we love ChatGPT! By the way, this is a good move. It’s a great move for both platforms. Very smart. What is it about, Mike?
Mike: Basically, Amazon is linking their DSP—their demand-side platform—with a very premium supply publisher, and that is ChatGPT, this new consumer surface.
00:18:48
Mike:
The idea here is… I actually have Amazon’s DSP page open to read how they present it. In case you’re not familiar, I think we did an episode about that probably like a year ago. They’ve been doing really good work on their DSPs. Two of their big USPs here… They talk about having premium supply, which is reaching customers across a huge supply of premium ad-supported content.
Mike: A lot of it is open web crap, to be honest, but there is premium stuff in there, and probably no better example than OpenAI or ChatGPT. Everyone’s excited to be on there.
00:19:10
Mike:
The other thing that they talk about—and this is really what actually powers everything—is unique audience signals and targeting. This is their first-party data, which spans browsing and search activity, your transaction history, even Prime Video streaming history. They know an awful lot about your interests in the market and everything. They know you very well.
00:19:34
Chris:
Yes, and this is exactly the idea: what if we could activate that first-party data on ChatGPT? It’s like a home run. Of course people are going to love this idea.
00:19:57
Chris:
Let’s talk about what that could mean tactically for ChatGPT, for Amazon, and maybe for the end consumer. Why this might be a win-win-win situation: I, as a user of ChatGPT, probably get better content served by ChatGPT and Amazon because there will be some customer matching going on. Secondly, ChatGPT has a clear path to monetize this potential click inventory. And Amazon gets clients back on their website.
00:20:24
Mike:
Exactly, exactly. This is a way for them to offer real value to their advertisers and attract more advertiser dollars.
Chris: Yeah, so there’s a flywheel.
Mike: Yeah.
00:20:53
Chris:
One second, Mike—my question would be (and I think we briefly touched this a couple of days ago): they talk about a deep technical partnership here. I was wondering, because with this first-party data, customer matching, Amazon probably will be doing this on Google anyway as part of their features… why is there this big flag of a “deep technical partnership”? Do they do more than anyone else can do, or do you have any information about it?
00:21:17
Mike:
It’s a good question. I don’t know that I can fully answer it yet because it’s very early. It’s at the level of press releases, which always have a certain level of detail, and I’m not the ad tech pipes guy. But it suggests to me that, yeah, there is an integration deeper than just, for example, Amazon creating a customer match list for high-propensity buyers and using it on Google without needing a deeper integration to buy ads on behalf of advertisers.
00:21:45
Mike:
Here, I don’t know if there will be something in terms of ad tech that’s a level deeper than that. But the bottom line is: if you’re buying through Amazon’s DSP, you’re going to be able to buy placements on ChatGPT using Amazon’s targeting capabilities, which I would say are a cut above ChatGPT’s targeting capabilities and the limited options that you as an end user would be able to do in the platform by yourself.
00:22:12
Mike:
So I think it’s probably going to be the best targeting in town for ChatGPT.
Chris: By the way, this is a big statement because we talked… I think a couple of weeks back we did a feature-for-feature comparison, and targeting was one big weak spot right now. I love that move.
00:22:40
Mike:
Yeah. As a user of ChatGPT, I’m someone who believes that personalized ads suck less than non-personalized ads. So I think that you will have a less annoying and more satisfying experience when you see more relevant ads, because ads are going to be there whether you want good ones or bad ones. From ChatGPT’s standpoint, this is giving access to Amazon as a demand platform—a huge amount of demand there.
00:23:22
Mike:
So this is very attractive for them to get that in the door as they’re trying to grow.
00:23:43
Mike:
Also, by default right now, a lot of these post-chat intents that get warmed up or generated end up just getting captured by Google Search, by the Big G. We’re talking about 25-30% of traffic, based on some analysis we’ve been doing.
00:24:06
Chris:
Remember back in the day, exactly!
Mike: So, given the alternative between “I’m just going to invisibly warm up intent and lose this stuff to Google, and their results are going to look great” or “I can have this super cool ad tech story and refer that traffic to someone who isn’t my arch rival and monetize it,” there’s for sure revenue upside for ChatGPT. Great move. Everyone wins there.
00:24:28
Chris:
Super interesting move.
Mike: From Amazon’s standpoint, hopefully it’s enough to smooth over… You know, you get bad PR one week, good PR the next. It’s all part of the game.
00:24:53
Chris:
Can you imagine an email is found now as part of the investigation and this fake bidder is literally called Jeff? That’s awesome!
Mike: Well, the FTC shared… You could see that one of the product managers or product marketers at Amazon was talking with ChatGPT like, “What is the best word I can call this? Like, I’ve built a proxy bidder that’s ripping people off—should I call it a surcharge or should I call it…”
00:25:23
Mike:
It’s so funny to call this a smoking gun. There are smoking guns left, right, and center.
Chris: My prediction is it won’t harm their Q4 results.
Mike: No, it won’t.
Chris: You’re boycotting on behalf of all the retailers who suffered from this shady behavior!
Mike: I’m boycotting Amazon, okay, but trust me, it won’t make a big difference!
00:25:45
Chris:
Mike, a great episode. Thanks for joining me.
Mike: That’s all right. Thanks everyone for tuning in. This was another episode of Growing Ecommerce, brought to you as always by Smarter Ecommerce, also known as smec. You can learn more at smarter-ecommerce.com. And as always, if you want to leave us a review, a rating, a comment, or a shoutout on socials, we really appreciate it. Every little bit helps. Thanks, we’ll see you next time. Bye-bye!