Aren't Search Ads based on a second-price auction? How can Google just raise the price? Or is it lying about the auction?
The only thing I can imagine is raising the minimum bid, but this would also result in less ads being shown (may still be overall beneficial but not a simple price increase).
Ad auction fraud is one of the allegations that landed them in court.
This former executive just testified under oath that they change the auction rules without telling advertisers, and do so in ways that increase Google's revenue specifically in order to cook their quarterly SEC filing numbers and mislead investors into thinking the ad business is healthy.
The executive also testified that they've fallen far behind Meta and Amazon on ad campaign quality and revenue.
edit: The bottom of this article explains one of the mechanisms they use. The top bidder gets second place, and the second bidder gets top place. This increases their ad revenue on average (tricks like this get them 5-10%, they think they could take 15% because of their monopoly power):
So... tl;dr: Their ad business is based on fraud built on top of other fraud.
He didn't even talk about the pop-under / youtube fraud that was reported a few months ago.
Each of the following guarantees YouTube makes to advertisers is demonstrably false:
- video ads will be displayed prominently
- video ads will be displayed on high-quality, reputable sites
- audio will be played back
- if the ad is over 5 seconds there will be a skip button that the user can click on, and the ad impression doesn't cost anything if the user skips it. (In many cases, the ads are rendered with an occluded or missing skip button.)
The difference between their claims and reality is responsible for a mid double-digit percentage of YouTube's revenue.
There are some lawsuits that allege that Google was effectively cheating before then, by sort of manufacturing a shill bidder, with privileged access to everybody's bids, that turned it into something closer to a first-price auction (https://www.wired.com/story/google-antitrust-ad-market-lawsu...). The quotes from within Google look pretty bad, e.g. one Googler allegedly wrote: "Doesn’t that undermine the whole idea of second price auctions? I.e., the assurance that you can bid the maximum you’re willing to pay with no negative consequence. But if the publisher manufactures a floor based on your bid to get you to pay more than the second price, this principle gets violated. It’ll transform the system into a 1st price auction where the bidder has a strong incentive to bid LESS than he’s willing to pay. (Only just enough to win.) I don’t think that’s desirable for either side in the long term."
> It’s important to note that our move to a single unified first price auction only impacts display and video inventory sold via Ad Manager. This change will have no impact on auctions for ads on Google Search, AdSense for Search, YouTube, and other Google properties, and advertisers using Google Ads or Display & Video 360 do not need to take any action.
No impact on Google Search - is that still the case?
Not a googler, but it sounds important that ads compete against no-ad bids.
Example: On search results, if all bids are 0.01$, then it’s worth showing no ad, because it will annoy the customer to have low-relevance ads. But what is the ceiling? 0.10$? 1.00$? Sounds reasonable to have a ghost bid which represents the weight of customer fatigue.
It’s hard to believe that google cares about customer fatigue as the search results are littered with ads.
I think the issue that the floor varies depending on google sets and there’s no visibility to how they do this.
If there’s only one bidder and the minimum price is $1 for one ad but $0.10 for another, why is that? It’s because google knows that one person can pay $1 while the other won’t.
It hasn’t been a proper auction for many years and is really just whatever google wants to charge. Because they have monopoly pricing power, customers don’t have any real alternatives to search ads.
It would be like if there was one magazine that had 90% of all readers and they charged whatever prices they wanted.
People are saying there is a "RGSP" (Randomized Generalized Second Price) mechanism where sometimes the second-place bidder wins instead of the first-place bidder.
Yeah, ad campaigns have budgets, and in order to spread traffic evenly during a day, a randomized ad suppression occurs. So sometimes it is the second place bidder who will win
The article says they changed auction rules that had the ultimate effect of increasing ad prices for some advertisers, at certain times, of between 5% and 10%.
Even if that's true, it's still an auction, with real competition and bidding. And it's not only about price, it's also about the quality of the ad and the quality of the advertiser.
Yes, absolutely, if the court can sort through the unfathomably complex layers of rules of real-time ad auctions and show that Google did something illegal, they should be held accountable.
But isn't it bizarre how everyone's so hot and bothered by a possible 5% to 10% swindle, meanwhile:
- Questcor bought the rights to Acthar Gel, which treats infantile spasms, a rare form of childhood epilepsy. Questcor raised the price of a vial from $40 to $23,000.
- Valeant bought diabetes drug Glumetza and raised the price from $800 for a 90-day supply to over $10,000.
Here's an industry that routinely runs 1000% arbitrary price increases, with no competition, while holding people's lives hostage. Seems like we should be far more outraged about that than an unproven, speculative, 5% to 10% issue.
>But isn't it bizarre how everyone's so hot and bothered by a possible 5% to 10% swindle, meanwhile:
No it is not bizarre at all. Honestly with this reply my response would be "How much is Google paying you to say this", but sadly it is likely you are shilling for them for free.
Now, maybe you are only able to focus on one thing at a time, but I do believe the average person can hold multiple opinions over multiple subjects at the same time, and while on a site about technical issues, complain about the technology companies implementing bad policies.
Sounds horrible but Glumetza is just metformin with something extra to stay in the stomach longer.
Acthar Gel is prednisone with a claim of less side effects. Interesting it got approved in the early 1960s and was grandfathered in and didn't need to provide trials so many questions exist around if it's any different from prednisone anyways. Either way patents have expired and a generic maker can make this.
You can be mad at both, either or none. But companies being shielded by patents doesn't really apply here so it's hard to generate outrage
Don’t most companies have budgets? Wouldn’t raising prices just cause companies to buy less? I feel like folks don’t just increase spend when prices go up, when it comes to ads.
Now if everyone is increasing budgets, prices of goods, ad spend, that story would make more sense. But that would correlate to whole market movements which honestly makes sense for something of this scale
For most companies ROI on ad spend is incredibly difficult if not impossible to calculate.
Setting aside all the challenges south multiple devices lagged purchasing etc there are a lot of sales processes that take a digital lead and convert it to real human contact so attribution is very difficult
I disagree with the "most" statement, most online commerce ads tracking is streamlined, user clicks on ads, goes to ebay, buys item, ads platform connects click and purchase through some cookies or fingerprinting and you got reliable tracking.
Google obviously tracks lots of people through multiple devises, because they are in eccosystem.
Drawing on my experience from working with an Ad-Tech:
> Don’t most companies have budgets?
Yes
> Wouldn’t raising prices just cause companies to buy less?
Not necessarily, the "ad-spend" typically remains a fixed budget until a given ad-campaign is complete.
The kicker is basically the cost per ad is increased, which means the budget for an ad-campaign is consumed more quickly. Thus, for example, rather than an ad-campaign lasting 4 weeks and generating X number of impressions, it lasts 80% of that time with 80% the impressions. Thus, even though the revenue is fixed per ad-campaign, Google would be able to cycle through more ad-campaigns
> I feel like folks don’t just increase spend when prices go up, when it comes to ads.
Depends on the company. In all probability the ad-spend would remain fixed and the companies just get less return on that spend (as an aside, the ROI for ads is highly debated to begin with)
>The kicker is basically the cost per ad is increased, which means the budget for an ad-campaign is consumed more quickly. Thus, for example, rather than an ad-campaign lasting 4 weeks and generating X number of impressions, it lasts 80% of that time with 80% the impressions. Thus, even though the revenue is fixed per ad-campaign, Google would be able to cycle through more ad-campaigns
This is, in fact, companies buying less (advertising), which is what I took the parent comment to be asking. It certainly isn't spending less.
Company X makes budget of $200k a quarter to buy ads. Agency Y setups up ads in Google, a budget and a keyword. They then tell Company X they got a certain amount of referrals/traffic/clicks/what ever.
Google increases ads 5% or 10%
Agency Y reports 5% or 10% decrease from last quarter. Company X goes well that's not great. Agency Y makes changes, Google changes back 5-10% change, next quarter it goes up. Company X goes great!
Google just made 5-10% extra an entire quarter.
Rinse and repeat.
Basically 5-10% is nothing, nobody would really notice. General traffic would probably change 5-10% every quarter, same with spending habits and other factors.
That's the thing though, with the Agency Y making changes. They would have to go back and ask for more money than before... which I just don't think that's happening (at least that's not what I do). I'm speaking from a place where I purchase ads for my product and know someone else who does as well. We just budget a set amount on ads with an agency and then measure what we get out of that. If it's worth it, then it's worth it for us. The tuning of 5% up and down doesn't really effect how much I spend per se. I only have so much to spend on that type of marketing. I guess the more expensive it is the more it grinds into profit margins because those clicks go down.
At one point though, it isn't profitable to market there anymore and I move onto something else, which wouldn't be what they want I would assume? That's where the ad "market" comes into play. The market decides the price supposedly? Is the argument that there is nothing else to move on to when they fudge the price?
If it is not a refactoring, a UI/UX redesign, or a wholly new ad product, then most likely it affects average prices of running ad campaings. Even if you think it does not, maybe you slipped in a coding bug. So before rolling out your change, you canary it and measure how much it changed stuff. A dumb raise in ad price will cause bad stuff for conversion costs, so it'll be a no-go.
Did you read the linked article? The context of this discussion is VP of Ads (Jerry Dischler) explaining on the stand what this Ad process of making changes look like, in his exec-level lingo, and news reporting on it adding another layer of obscurity. I explain what it means for the HN audience
Yeah, Ads Googlers perform a massive numbers of these price-affecting canarying and rollouts yearly, and it's not about raising minimal bid prices that folks ITT think of. Even if Google didn't support automated bidding strategies, there would be other considerations, for instance hourly budgeting. Not diving into details.
Yes, Mat, I read the linked article. And I think I know a lot more about this than you do. You use the word "canarying" and "rollouts" as if they meant anything.
The answer is monitor & adapt. If Google has indeed increased the price for ads, you can choose to be upset with them or to see it as a signal for your company to diversify the methods through which you drive traffic, leads, sales, and brand awareness.
I'm into SEO, few months ago with the rise of AI it felt like the SEO industry is about to end. The ranking signals of today, might not be as useful tomorrow. So does your PPC budget.
There really aren’t many other options. If Google raises prices 10x, then your budget will be 10x less effective, but it will still probably be much more effective than if you were to advertise on, say, Bing or Yahoo instead.
Till a competitor puts ads out for your brand and shows up at the top of results in formatting that's barely distinguishable from an actual search result.
* Cause or allow any person to bid at a sale for the sole purpose of increasing the bid on any item or items being sold by the auctioneer, except as authorized by Section 2328 of the Commercial Code or by this title. A violation of this subdivision includes, but is not limited to, either of the following:
* (1) Stating any increased bid greater than that offered by the last highest bidder when, in fact, no person has made such a bid.
* (2) Allowing the owner, consignor, or agent thereof, of any item or items to bid on the item or items, without disclosing to the audience that the owner, consignor, or agent thereof has reserved the right to so bid.
"Without disclosing" sounds like a gigantic loophole. They'll just reserve the right to bid on page 53 of their phonebook contract and since they own the ad market everyone just has to agree to let them.
Disclosure is an effective deterrent in a competitive marketplace - the implication being that once this behavior is disclosed nobody will want to work with you, or they'll negotiate it away somehow.
In a consolidated market disclosure is a water sandwich.
Anecdotaly, I find Google Ads hard to use despite their market dominance.
I have been trying to recruit participants for my academic observational health study. My ads get incorrectly flagged as clinical trials and there seems to be no way to reach anyone who could override it. I also wanted to use a specific advertising strategy, which is only available via Google's poorly documented desktop software. Facebook Ads, while far from ideal, are much easier to set up.
The way Amazon Advertising utilizes dark patterns to boost search revenue is by setting default bids artificially high. The UX preselects "suggested bids" at the top of the bidding range and daily budgets for some ad types at $100.
Even if you choose "custom bids" for keywords the default will be $1 or $3 depending on the ad type.
Of course, thousands of people accepting the defaults results real-time bid levels rising across the board for everyone.
They just accept the defaults even when it means more money comes out of their pocket?
How is anyone supposed to design products when customers don’t consider their money to be scarce? And why do we not expect Amazon to not take advantage of it?
It’d be like if I was selling a used bike on Craigslist and I said “default price $10k” and then people were like okay guess I’ll just pay way too much money
You'd be amazed at how many junior people at an SMB or publishing imprint who are thrown into the deep end and told to spin up an Amazon ad campaign with zero knowledge of the platform. It's not their money, and their bosses don't get it either, so they accept the defaults and the suggested bids, and blow through a pile of cash. Some realize the mistakes and quickly optimize, and learn how to set up future campaigns the right way.
The other group getting burned are desperate passive income newbies sitting on a pile of junk from Alibaba and grasping at anything which might move the needle on sales. To pay for campaigns, they'll even take out loans from Amazon Lending or predatory finance partners like Parafin Cash Advance, which take a percentage of your gross sales every other week.
Self published authors are also getting tricked by Amazon suggested bids, high default budgets, and ridiculous default levels for custom bids.
That doesn't seem sustainable. Amazon's advertisers seem like the type to want to sell something directly rather than bolster their brands, and if they aren't getting a return on their spend they'll stop.
63 comments
[ 3.1 ms ] story [ 260 ms ] threadThe only thing I can imagine is raising the minimum bid, but this would also result in less ads being shown (may still be overall beneficial but not a simple price increase).
This former executive just testified under oath that they change the auction rules without telling advertisers, and do so in ways that increase Google's revenue specifically in order to cook their quarterly SEC filing numbers and mislead investors into thinking the ad business is healthy.
The executive also testified that they've fallen far behind Meta and Amazon on ad campaign quality and revenue.
edit: The bottom of this article explains one of the mechanisms they use. The top bidder gets second place, and the second bidder gets top place. This increases their ad revenue on average (tricks like this get them 5-10%, they think they could take 15% because of their monopoly power):
https://www.msn.com/en-us/money/companies/google-tweaks-ad-a...
So... tl;dr: Their ad business is based on fraud built on top of other fraud.
He didn't even talk about the pop-under / youtube fraud that was reported a few months ago.
Each of the following guarantees YouTube makes to advertisers is demonstrably false:
- video ads will be displayed prominently
- video ads will be displayed on high-quality, reputable sites
- audio will be played back
- if the ad is over 5 seconds there will be a skip button that the user can click on, and the ad impression doesn't cost anything if the user skips it. (In many cases, the ads are rendered with an occluded or missing skip button.)
The difference between their claims and reality is responsible for a mid double-digit percentage of YouTube's revenue.
It’s clear from Facebook for example that they have dramatically lowered the bar for what can advertise recently.
Bidding against a larger pool of advertisers would increase the clearing price of the auction.
There are some lawsuits that allege that Google was effectively cheating before then, by sort of manufacturing a shill bidder, with privileged access to everybody's bids, that turned it into something closer to a first-price auction (https://www.wired.com/story/google-antitrust-ad-market-lawsu...). The quotes from within Google look pretty bad, e.g. one Googler allegedly wrote: "Doesn’t that undermine the whole idea of second price auctions? I.e., the assurance that you can bid the maximum you’re willing to pay with no negative consequence. But if the publisher manufactures a floor based on your bid to get you to pay more than the second price, this principle gets violated. It’ll transform the system into a 1st price auction where the bidder has a strong incentive to bid LESS than he’s willing to pay. (Only just enough to win.) I don’t think that’s desirable for either side in the long term."
Google... disputes this (https://blog.google/outreach-initiatives/public-policy/ag-pa...). I guess we'll see.
No impact on Google Search - is that still the case?
Example: On search results, if all bids are 0.01$, then it’s worth showing no ad, because it will annoy the customer to have low-relevance ads. But what is the ceiling? 0.10$? 1.00$? Sounds reasonable to have a ghost bid which represents the weight of customer fatigue.
I think the issue that the floor varies depending on google sets and there’s no visibility to how they do this.
If there’s only one bidder and the minimum price is $1 for one ad but $0.10 for another, why is that? It’s because google knows that one person can pay $1 while the other won’t.
It hasn’t been a proper auction for many years and is really just whatever google wants to charge. Because they have monopoly pricing power, customers don’t have any real alternatives to search ads.
It would be like if there was one magazine that had 90% of all readers and they charged whatever prices they wanted.
https://threadreaderapp.com/thread/1703969874698011098.html
Even if that's true, it's still an auction, with real competition and bidding. And it's not only about price, it's also about the quality of the ad and the quality of the advertiser.
Yes, absolutely, if the court can sort through the unfathomably complex layers of rules of real-time ad auctions and show that Google did something illegal, they should be held accountable.
But isn't it bizarre how everyone's so hot and bothered by a possible 5% to 10% swindle, meanwhile:
- Questcor bought the rights to Acthar Gel, which treats infantile spasms, a rare form of childhood epilepsy. Questcor raised the price of a vial from $40 to $23,000.
- Valeant bought diabetes drug Glumetza and raised the price from $800 for a 90-day supply to over $10,000.
Here's an industry that routinely runs 1000% arbitrary price increases, with no competition, while holding people's lives hostage. Seems like we should be far more outraged about that than an unproven, speculative, 5% to 10% issue.
No it is not bizarre at all. Honestly with this reply my response would be "How much is Google paying you to say this", but sadly it is likely you are shilling for them for free.
Now, maybe you are only able to focus on one thing at a time, but I do believe the average person can hold multiple opinions over multiple subjects at the same time, and while on a site about technical issues, complain about the technology companies implementing bad policies.
Acthar Gel is prednisone with a claim of less side effects. Interesting it got approved in the early 1960s and was grandfathered in and didn't need to provide trials so many questions exist around if it's any different from prednisone anyways. Either way patents have expired and a generic maker can make this.
You can be mad at both, either or none. But companies being shielded by patents doesn't really apply here so it's hard to generate outrage
Now if everyone is increasing budgets, prices of goods, ad spend, that story would make more sense. But that would correlate to whole market movements which honestly makes sense for something of this scale
Setting aside all the challenges south multiple devices lagged purchasing etc there are a lot of sales processes that take a digital lead and convert it to real human contact so attribution is very difficult
Google obviously tracks lots of people through multiple devises, because they are in eccosystem.
> Don’t most companies have budgets?
Yes
> Wouldn’t raising prices just cause companies to buy less?
Not necessarily, the "ad-spend" typically remains a fixed budget until a given ad-campaign is complete.
The kicker is basically the cost per ad is increased, which means the budget for an ad-campaign is consumed more quickly. Thus, for example, rather than an ad-campaign lasting 4 weeks and generating X number of impressions, it lasts 80% of that time with 80% the impressions. Thus, even though the revenue is fixed per ad-campaign, Google would be able to cycle through more ad-campaigns
> I feel like folks don’t just increase spend when prices go up, when it comes to ads.
Depends on the company. In all probability the ad-spend would remain fixed and the companies just get less return on that spend (as an aside, the ROI for ads is highly debated to begin with)
This is, in fact, companies buying less (advertising), which is what I took the parent comment to be asking. It certainly isn't spending less.
Google increases ads 5% or 10%
Agency Y reports 5% or 10% decrease from last quarter. Company X goes well that's not great. Agency Y makes changes, Google changes back 5-10% change, next quarter it goes up. Company X goes great!
Google just made 5-10% extra an entire quarter.
Rinse and repeat.
Basically 5-10% is nothing, nobody would really notice. General traffic would probably change 5-10% every quarter, same with spending habits and other factors.
At one point though, it isn't profitable to market there anymore and I move onto something else, which wouldn't be what they want I would assume? That's where the ad "market" comes into play. The market decides the price supposedly? Is the argument that there is nothing else to move on to when they fudge the price?
They being sold less product for a high price and their tools to identify that are manipulated opaquely.
But there is literally no one who understands it all, so I'm sure there are ways.
have fun
For example for the keyword “foo” google sets the minimum bid at $1. I set my max bid at $5. There’s only 5 bidders so we all pay $1.
Google ups the minimum to $2 and it’s under our max bid so now we all pay double.
That’s how.
If it is not a refactoring, a UI/UX redesign, or a wholly new ad product, then most likely it affects average prices of running ad campaings. Even if you think it does not, maybe you slipped in a coding bug. So before rolling out your change, you canary it and measure how much it changed stuff. A dumb raise in ad price will cause bad stuff for conversion costs, so it'll be a no-go.
Yeah, Ads Googlers perform a massive numbers of these price-affecting canarying and rollouts yearly, and it's not about raising minimal bid prices that folks ITT think of. Even if Google didn't support automated bidding strategies, there would be other considerations, for instance hourly budgeting. Not diving into details.
There are a lot of VPs at Google.
And why isn't the answer also take Google to your state attorney for fraudulent and anti-competitive practices?
If organic doesn't work, it's because you are not ready, ads won't fix your issues.
Prohibited in California:
* Cause or allow any person to bid at a sale for the sole purpose of increasing the bid on any item or items being sold by the auctioneer, except as authorized by Section 2328 of the Commercial Code or by this title. A violation of this subdivision includes, but is not limited to, either of the following:
* (1) Stating any increased bid greater than that offered by the last highest bidder when, in fact, no person has made such a bid.
* (2) Allowing the owner, consignor, or agent thereof, of any item or items to bid on the item or items, without disclosing to the audience that the owner, consignor, or agent thereof has reserved the right to so bid.
In a consolidated market disclosure is a water sandwich.
Fantastic
I have been trying to recruit participants for my academic observational health study. My ads get incorrectly flagged as clinical trials and there seems to be no way to reach anyone who could override it. I also wanted to use a specific advertising strategy, which is only available via Google's poorly documented desktop software. Facebook Ads, while far from ideal, are much easier to set up.
They don't have any incentive to improve the product, only to increase revenues (e.g. by raising prices).
Even if you choose "custom bids" for keywords the default will be $1 or $3 depending on the ad type.
Of course, thousands of people accepting the defaults results real-time bid levels rising across the board for everyone.
How is anyone supposed to design products when customers don’t consider their money to be scarce? And why do we not expect Amazon to not take advantage of it?
It’d be like if I was selling a used bike on Craigslist and I said “default price $10k” and then people were like okay guess I’ll just pay way too much money
The other group getting burned are desperate passive income newbies sitting on a pile of junk from Alibaba and grasping at anything which might move the needle on sales. To pay for campaigns, they'll even take out loans from Amazon Lending or predatory finance partners like Parafin Cash Advance, which take a percentage of your gross sales every other week.
Self published authors are also getting tricked by Amazon suggested bids, high default budgets, and ridiculous default levels for custom bids.
Amazon is laughing all the way to the bank.