Maybe it's too early, but that article was not clear to me. Is it alleged that Google incorrectly stated the value of the second place bidder in order to make it look higher, in order to pressure the first place bidder to bid higher? Even that doesn't make much sense, so I'm missing something.
> The shift sought “to raise the prices against the highest bidder,” Whinston told Judge Amit Mehta in federal court in Washington... To help eliminate that 20% between the runner-up and what the winner was willing to pay, Google gave the second-place bidder a built-in handicap to make their offer more competitive, Whinston said... Whinston’s comments Friday described Google’s technique, called “squashing,” that seeks to make the runner-up’s bid more competitive.
If they're inflating the second-place bid to be higher than the first-place bid, I'd wonder why that wouldn't backfire for them and result in the second highest bid winning a lot of the time. And if they didn't inflate it above the first-place bid, why would the winning bidder feel any pressure to increase their bid?
I'm sure it all makes sense, it's just not clear to me from this explanation.
There’s two common kinds of auction those with multiple rounds where the final bid wins and those where everyone puts in the maximum amount upfront and the winner pays whatever #2 bid. The second type is much faster but incentivizes the auction house to lie about the #2 bid.
Google did the second type and then got caught lying about bids. Doing so is really tempting but also generally fraud.
Edit again - I stand corrected, while it is true that the standard implementation is as I said, seems from TFA Google’s implementation is second-place+0.01.
They disclose its second place + 0.01, but the article says they lied and raised it by 15%.
“Google’s advertising auctions require the winner to pay only a penny more than the runner-up. In 2016, the company discovered that the runner-up had often bid only 80% of the winner’s offer. To help eliminate that 20% between the runner-up and what the winner was willing to pay, Google gave the second-place bidder a built-in handicap to make their offer more competitive, Whinston said, citing internal emails and sealed testimony by Google finance executive Jerry Dischler earlier in the case.”
“Google gave the second-place bidder a built-in handicap to make their offer more competitive” is the tamest way of phrasing it, given that the “handicap”’s only effect is to cost the first-place bidder more money.
It never helps the second-place bidder. I’d argue “handicap” is deceitful.
Ahh okay, the missing piece is that the winner pays the second-highest bid, plus a penny. I can definitely see how this would be fraud, and how it could be justified internally by Google as "it's still less than they bid".
I don’t think the point is that it’s fraudulent, remember that this is in the context of the antitrust trial. I think the point here is that they could only do that since they have a monopoly.
Though I’m not sure I even agree with that. It burns advertiser’s goodwill for sure, but they advertise on Google because it provides them with a ROI, not because there is no where else. This just forces them to tweak their bids more.
That’s without weighing in on whether Google has an illegal monopoly on search
Something can be both fraudulent and an antitrust issue. In fact there is probably a specific type of fraud which is just this (at least in the UK), which is "Fraud by abuse of position"
The UK legal definition of this would be met if a person/company:
1. occupies a position in which he was expected to safeguard, or not to act against, the financial interests of another person (e.g. operates an auction to tender their online marketing spend?)
2. abused that position (e.g. by putting in fake bids, which only they could do because of their position?)
3. dishonestly (e.g. intentionally failed to disclose or was otherwise intentionally dishonest)
4. intending by that abuse to make a gain/cause a loss (e.g. they intended to raise profits)
The abuse may consist of an omission rather than an act (e.g. number 3 can be a failure to disclose).
Winner expects to pay runner-up bid (this is a very important piece, because this changes the behaviour of bidders). Google quotes them an incorrect bid in order to collect more.
The allegation is that Google profited from lying, which is the definition of fraud. They stole, by making someone pay more than they otherwise would have, through deception. If the deal was “you pay what you bid” then this would be fine, but that was not the deal.(To be clear, I have no idea what the deal was, I’m just explaining the OP)
Exactly this. You can end up with some weird situations. I saw one guy get a criminal conviction for this: he repaired elevators. He left RepairCorp where he worked and set up on his own. BuildingCorp continued to pay him for their repairs not realizing it wasn't RepairCorp. In the trial they stated that they were always very happy with his work and the price was identical to RepairCorp. They were pissed he had lied to them though, and the guy ended up getting convicted for fraud.
I'm aware of what fraud is, I just didn't understand based on the parent comments what fraud was being committed (what lies were told, etc). I didn't pick up on the fact that Google was advertising paying the runner-up bid plus a penny but then marking up the runner-up bid substantially.
But that's not what the witness expert claims. He said "squashing".
Google used a second price auction, and also it ranked ads in the auction by bid multiplied by click through rate. Squashing is something like ranking ads by (bid * power(ctr, gamma)) where 0 < gamma < 1. In auctions where the 3rd bid (or lower) wins under the (bid * ctr) rank switching to squashing may increase revenue, because the actually higher bid will win the auction.
I meant that if they claimed that the winner would pay the runner-up bid plus a penny, and then the winner actually paid the runner-up bid plus a percentage, plus a penny, then that would be fraudulent (assuming it was not disclosed).
I have the same question too, say on ebay if I bid for something for a $1.00 but the second highest bidder was $.80, would i pay the $1.00 or $.80? (Its been a long long time I used ebay or did any bidding)
> New way: Winner bids $1.00, 2nd place bids $0.80, Google adds 15%* markup to 2nd bid, Winner pays $0.93.
This is true. But the story gets crazier. Let's say that CorpA and CorpB are both bidding on the term "Valuable Widgets". Here is who it goes:
CorpA bids $0.80
CorpB bids $1.00
--
CorpB wins the auction, pays $0.80 + 15% = $0.92
--
CorpA is sad they lost and wants so they try to outbid. They must pay $1.00 + 15% or $1.15
CorpA wins the new action, paying $1.15
--
CorpB needs those auctions to sell more widgets, so they now must pay $1.15 + 15% or $1.32
It obviously goes back and forth, but as you can see, Google is widening the gap each time to escalate the bid up faster. It increases competition and raises prices. In order to win you can't just pay what the previous person paid, you must pay a premium of 15%.
Now the part that is unclear is what happens if you ignore the winning bid price. So if CorpA has a max bid of $1.00 and CorpB tries to bid, they will see that they need to bid $1.15 to win the bid, because Google is saying that is what is needed to win the auction, but if CorpB only bids $1.10 they should still win the auction at $1.10 since that is their max bid and CorpA is only willing to pay $1.00. So CorpB in this case would pay their max bid and win. But either way, google is artificially pushing the price up higher than it technically needs to be.
That doesn't sound right. If the loser can see what the winner bid, they'd just need to bid for $1.01 next time to win. They'll get charged $1.15, but their bid will be $1.01, which means the loser next round will see that that was the winning bid and they'll know to bid for $1.02, and so on.
Since almost every comment is missing the nature of the change and its effects, let me put it in a clearer way:
Google stated that they were holding naive second price (with an asterisk) auctions and moved on to holding more optimal auctions. Optimal auctions are the ones where you set reserve prices and assign handicaps to each individual buyer to make the auction more competitive and yield higher revenue, like Google has done. (How to calculate optimal reserve prices and handicaps is left to reader, should be fairly trivial because economics is not a real science.) So, in effect Google has increased advertising prices without communicating the fact, in a way that the participants could also attribute the cost increase to changing marketing landscape. The fact that Google expected 15% more revenue puts them over the standard 5% test for significant market power, and hiding the price hike is easy evidence for the abuse of such power.
How much of this ties to enhanced CPC bidding (enabled by default, pushed by account reps as a must have)? It allows your bids to exceed your maximum CPC in scenarios that Google defines-- which, so it seems, are likely scenarios that Google views as having profits left on the table.
The heavily mathematical part is the proof that goes with the justification for the bidding system. Explaining the merits of the auction or explaining what the fraud is is pretty easy.
Professor Whinston didn't work at Google, did he? He just makes his opinion based on some emails found in the discovery?
For all we know, it is equally plausible that Google at some point launched a new prediction model, and in a traffic experiment it showed up the following stats: increase on campaign average cpc price, decrease on average auction discount (or whatever the metric for the gap to 2nd bid is called), and no harm to conversion cost. Team manager reported the metrics highlights to their higher ups. Emails found in discovery... ends with guys on HN throwing fraud accusations, pretty easy.
That person is an expert witness. He's there to explain to non-experts the meaning of what's been found in discovery. What he's saying to the court has likely been vetted by the prosecution.
If your claim is that the prosecution is making up wild shit, then Google's attorneys should have a field day rebutting it.
This trial is antitrust, not fraud, so Google lawyers will be rebutting it only to the point it helps them. And they will try to keep it sealed anyways. Since it doesn't seem that either party is willing to describe the auction algorithmically, we may never get to learn it, apart from reading Albert Cory substack
In particular, the linked article mentions 'squashing', and a quick research finds https://www.theregister.com/2010/09/16/yahoo_does_squashing/ where it is mentioned that ads are ranked not by bids, but rather by "bid multiplied by click probability".
In the ideal world, the first question to the auction expert should be "what price is charged, if due to click probability, the second bid ad wins the auction?" and the follow ups would be "but surely, the price charged isn't higher than the bid, or else the advertisers would notice that?" and "supposing the third bid ad wins due to the click probability, how would disabling that multiplying by click probability affect google revenue?"
I work with NGO (property law) and couple of months ago we were awarded a grant for ads, around €2000/mo. Initially, I was pretty surprised with how generous this was, but later realized that it's both - a way for Google to brag how much they help "good causes", but most importantly - to raise the ad costs for others.
All in all - we made a few spammer operations less profitable, a couple wealthy assholes were angry about us helping people defend themselves and we brought quality traffic to the website and FB group. Thanks Skynet!
The auction code was worse than a Google-only black box, when I was in Ads. The code itself was in a "high intellectual property" subdirectory that most engineers could not access.
There's an elaborate economic rationale for second-price auctions, including that they avoid "winner's remorse." Leave it to the late-stage "improvers" to say, "hey, we can capture some of that gap between first and second place!"
>There's an elaborate economic rationale for second-price auctions, including that they avoid "winner's remorse." Leave it to the late-stage "improvers" to say, "hey, we can capture some of that gap between first and second place!"
The reason Google uses second price auctions is because first price auctions incentive constant bid changes. The stability increases the money Google makes.
edit: For more context, and ignoring the whole generalized second price auction parts, the revenue from first and second price auctions is the same. In theory with optimal bidding and the usual assumptions of auction theory papers. The difference is that in a second price auction your optimal bid is independent of the other bidders. In a first price auction your optimal bid is dependent on the expected distribution of other bidders. In theory with all the usual auction theory assumptions. In reality with online auctions this meant a lot of churn in bids and a preference for bidding more conservatively. Which isn't good for advertisers or the revenue of the company running the auction. Second price auctions mostly solve this especially if you adjust the minimum auction price (ie: reserve price) per auction.
edit2: A great way to get more money is to have advertisers think they're bidding in a second price auction while actually running something closer to a first price auction. Which is essentially what the article talks about.
I’ve been in the field for 20 years and I know the difference between what companies do and what theoretical papers say they do.
As the other person said, bids repeat across multiple auctions on the same keyword and even in a generalized second price auction optimally your bids are adjusted based on the expected dynamics of each auction. Based on past information and so on. In a first price auction you’d keep lowering your average bid until it’s on average right above the next bid. The next bidder would then increase their bids. You’d then do so as well. Etc. Etc.
You’re moving the goal post from not knowing auctions to not knowing Google. Will the next question be if I made the decision at Google for this approach when they were competing with Overture?
I always thought that HIP vs. rest of google3 would make a great case study for developer experience/productivity research. Fewer eyes, emergent private style guide variant, etc.
> There's an elaborate economic rationale for second-price auctions, including that they avoid "winner's remorse." Leave it to the late-stage "improvers" to say, "hey, we can capture some of that gap between first and second place!"
Are there any more examples or reading around this concept?
Mathematically, all fair auctions provide the same optimal price to the seller. It's called the Revenue Equivalence Theorem.
Google's auction is not fair in the sense that Google is both the seller and the auctioneer. Squashing enters the situation when the second price bidder has an ad that is more likely to be clicked than the first price bidder and therefore has additional revenue potential (known only to Google) beyond the revenue offered by the bid. Google then assigns this additional expected revenue to the value of the second price bid and declares this to be the market value of the bid. I see where their argument comes from, but it's a pretty lousy one.
That’s what happens when you don’t have open source and open protocols. I see people on HN constantly complaining but very little support for distributed systems, especially byzantine fault tolerant with proper incentives, why? Incentives are key to such distributed systems working, but it sounds too much like Web3? FTX? Or what? I guess many people don’t want to see viable alternatives appear. So then all you can do is keep complaining about black boxes and secret actions.
I believe he is mocking the average HN commenter, who simultaneously complains about these types of opaque-black-box systems but attacks transparent distributed systems.
That's how I interpreted it. Everyone seems to want the advantages of the FOSS model for general benefit, but want firms to have their advantages at the same time, and not acknowledge the inherent adversarial aspect of that.
The formula is interesting: ad selection uses bid ⨯ relevance^q, where q is the squashing value, which is lowered when the second-highest bidder is likely to bid low, creating a tradeoff:
> [David Pennock] adds that squashing can significantly improve revenue, at the expense of advertiser and user satisfaction. As a result, it is necessary to set acceptable thresholds for loss of relevance, and then optimize revenue based on those thresholds.
Just a reminder that the EU started a preliminary antitrust action over Google's ad business earlier this year.
> Google’s advertising business should be broken up, European Union officials said Wednesday, alleging that the tech giant’s involvement in multiple parts of the digital advertising supply chain creates “inherent conflicts of interest” that risk harming competition.
“@Google controls both sides of the #adtech market: sell & buy,” tweeted Margrethe Vestager, the commission’s top competition official. “We are concerned that it may have abused its dominance to favour its own #AdX platform. If confirmed, this is illegal.”
To be clear though, the article is talking about search ads which run on a separate auction. The actions you are referring to concern the display side where Google runs a different exchange with multiple sellers for the inventory.
Sure, but if the EU is looking for evidence of Google abusing it's dominant position in ad tech for it's own benefit, the US antitrust case certainly gives them a good place to start looking.
My company spent a lot (relatively to our size) of money in AdWords withour a correlated conversion rate. Not saying we could not be completely wrong but the intuition is that more conversions are expected.
We tried different strategies, even hired competing agencies targeting different locations. Obviously a common feedback was that we need to increase the budget. I would love to to that once I see good conversions. We were using AdWords since 2003.
In this kind of auction, the winner pays the runner-up bid. Google simply increased that bid before it was communicated to the winner.
The only effect, as this doesn't change the bid order, is that Google pockets the "built-in handicap" on top of the runner-up bid, without telling the winner what the actual bid is.
This is so obvious to everyone who has used their ad system.
They get away with that because they know marketing departments don't care - they have ad budgets to spend. Google gives them enough dashboards and excuses to satisfy their bosses. It is still unknown which half of advertising works.
> They get away with that because they know marketing departments don't care
Even if they cared, they have no leverage.
In the financial sector, any market-maker in Google's position would be heavily regulated, and the kind of shit Google pulls would land people in jail. Even without the market manipulation, there's no way a company like Google could operate without a clear separation between departments. The reason for that is that centuries of playing that game have shown that it's the only way to keep market makers honest.
I believe this is true in terms of the amount of money in the market, but do search ads actually work? I can't imagine them being even half as effective as banner ads or video ads or native ads. So why is it such a big business?
Search ads are valuable because when you search for "best laptop 2023", googles decision of what to show will have a big determining factor whether you buy an hp, a macbook or a thinkpad.
Companies are prepared to hand over literally hundreds of dollars to persuade Google to direct you to them rather than a competitor.
Whereas most people are fairly used to ignoring banner ads, and the vast majority of ads don't lead directly to a purchase, so they tend to be worth only a few cents at most.
It's always useful to show the otherwise fraudulent nature of the business, and the moral bankruptcy of the people running it.
Inevitably some google exec will testify that they "strive to bring the best to their customers and bring a public-good service". By then, everyone will have been warned that this person is unquestionably a swindling weasel.
Ads being the main funding source of media, it's an essential piece of a functioning democracy.
Without decent revenue streams for quality news outlets, the alternative we're slowly heading to is "sponsored" media, with patrons such as Jeff Bezos funding papers like the WP, and distorting their ability to report news accurately, or clickbait media not contributing to news discovery.
> This is so obvious to everyone who has used their ad system.
This. I tried once and am still a bit confused if I understood it correctly. Google seemed to refuse "too low" bids completely for no other disclosed reason than that they were too low. And the UI was hellbent to make me give google authority to bid on my behalf on googles own ad auction. Like, WTF? Who on their right mind would agree to that? I'd very much like these folks to be my customers. I guarantee, I will sell you whatever (legal) you want, as long as you agree that I can set the price to whatever I want. It just boggles my mind.
Ad attribution is difficult, but it's difficult everywhere because customer behaviour is complex. It's much easier to attribute digital ads sources than it is to attribute out of home advertising. This isn't anything Google specific.
Hey look, it's another online-first startup reinventing media mix modelling!
(Sorry for the snark but this is a very well studied technique in the offline world that gets completely ignored by online advertisers, rather like AB testing gets mostly ignored by offline advertisers).
Because compared to the alternatives, they significantly easier.
Attribution is arguably the biggest challenge in marketing. If you know that someone bought because of an ad you know you can spend more on that ad, if you know no one is buying from it you know not to.
With almost all digital marketing there's some amount of tracking, not necessarily of users, but of ads. When you make a sale you know which ad it came from. Now that's not perfect attribution, users might take a few views to click through, may go direct rather than clicking through, or may click through multiple different ads, but it's pretty good.
With non-digital marketing there's nearly zero information. Billboards, TV adverts, magazines, radio, even podcasts which are digital, etc. The way that attribution is done is based on times of day (did you get a spike after your TV advert aired), or it's based on location (did you get a spike in the area of the billboard). Sometimes discount codes are used for attribution (use code TUBE10 to save £10 on your next order), but those only work for some types of product and users forget more often than not (yes, even with discounts).
Marketing on Google, Facebook, affiliate marketplaces, etc, is all automatable, and more accurate, more precise, and therefore it's easier to understand marketing ROI, and easier to justify more spend.
This is the case for "auction" based ad platform, I don't see how there is anyway to ever police this especially when they run the auction and have full view of all bids, essentially letting them perfectly price discriminate. Where are customers gonna go? Bing?
This and all the dark patterns to nudge the user to turn on "ad optimizations" without explaining what it does, anyone who has dabbled in google ads can surely attest
The modern insight is that you should be estimating your ads results instead of looking into detailed proxy metrics. And all that they lying on their metrics means is that you'll have a hard time optimizing your ads there, and thus move into other, more friendly platforms.
But, of course, I have no idea how much of this is real (it certainly is real to some extent), and how much is propaganda created to fool people into spending more on ads (it certainly is also propaganda to some extent). On some contexts, that estimation is incredibly hard to do correctly, and I don't know if people even try it.
> To help eliminate that 20% between the runner-up and what the winner was willing to pay, Google gave the second-place bidder a built-in handicap to make their offer more competitive, Whinston said, citing internal emails and sealed testimony by Google finance executive Jerry Dischler earlier in the case.
At the end of the day, Google can set the price however they like for ads on Google Search.
Previously they ran an auction. Now they have a very complex price-setting algorithm that has elements of an auction within it, but also secret factors which push bids up or down.
But they're well within their rights to just roll a dice.
I'm kinda surprised they don't have some human/AI combination that estimates how much profit you are earning from Google Ads, and charge most of that.
The whole idea of having a price for each individual impression seems outdated - what businesses care about is "how much did we pay Google this month, and how many widgets did we sell, and was it worth it or should we advertise elsewhere?". Therefore, why bother with the middle man - just make an algorithm to predict exactly how much the business thinks the ads are worth.
They do have this, you can choose to bid for a target conversion cost (I want to pay at most $x per conversion). In general I find all of this pretty silly as a Google employee.. I don't think you can distill the complexity of any change down to one article. Say for example that Google did just raise how much advertisers paid per click across the board, why would that lead to billions of dollars in profit? I would bet that the majority of advertisers on Google already spend their whole budget, so that would mean that now they would just be getting less clicks and conversions per dollar of budget spend. Logically they would then move some of this budget to other advertising channels such as Facebook. Charging more per impression always has second order effects, but who cares about that on hackernews
But thats why modelling businesses has such a large opportunity. Some businesses will be prepared to pay more if you send them more conversions... Others have a fixed budget. Some use many platforms, and certain things might persuade them to move more budget to you. Some might be persuaded by "tricks" like handing out airmiles, vouchers or free stuff to the marketing team in return for ad spend.
All of those things require a higher level strategy than just doing everything per-impression.
Seriously, nobody has a right to any particular algorithm or outcome. Google can rank ads first if they contain the word "orangutan" if that suits them. Buyers should be looking at their ROI, not expecting the ad platform to optimize every aspect of it.
Ok, then do they need to be truthful about anything?
How about giving themselves credit for extra clicks based on the value of the word? (With the defense they are doing it for you!)
What about if they switch to a click tracking system that uses statistics (so they can use the defense that the method is just inaccurate and they aren't lying), which just happens to over counts clicks?
How about charging you based on very erroneously projected traffic (one further than the step above, basically projecting the future very genrously in their favor) and not actual traffic?
What about if they enable the same third parties to both participate in an ad exchange as a seller and a buyer, with real time bid information, such that they can place phantom bids to up advertisers spend, just like HFT dark pools?
I could go on, every thing is all good, right? No guarantees?
Which ways is it ok if they misrepresent, overcount, badly estimate, fail to deliver, bill based on future estimates, and otherwise enable third parties to steal from you are ok?
If Google is allowed to arbitrarily set the price of their ads, and they are, then any legal pricing scheme (i.e. not running afoul of discrimination or price gouging laws) ought to be fine as long as they're lying to you about it or making false claims.
What laws did they break? I’m confused by the sudden interest in monopolies when the duopolies have developed over 20 years. It’s like after Microsoft government assumed Google / Facebook and Google / Apple and EBay / Amazon had each other so all was good?
Is anyone really shocked that when you tell the seller the maximum rate you are willing to pay that will be motivated to get that maximum rate when the seller is running the auction?
> The shift sought “to raise the prices against the highest bidder,” Whinston told Judge Amit Mehta in federal court in Washington... To help eliminate that 20% between the runner-up and what the winner was willing to pay, Google gave the second-place bidder a built-in handicap to make their offer more competitive, Whinston said... Whinston’s comments Friday described Google’s technique, called “squashing,” that seeks to make the runner-up’s bid more competitive.
If they're inflating the second-place bid to be higher than the first-place bid, I'd wonder why that wouldn't backfire for them and result in the second highest bid winning a lot of the time. And if they didn't inflate it above the first-place bid, why would the winning bidder feel any pressure to increase their bid?
I'm sure it all makes sense, it's just not clear to me from this explanation.