When reading things like this (and assuming it's true), I always wonder how it works in practice. I mean somebody has to actually implement the trickery in the codebase, right? Does that mean that one day there was a ticket that could be summarized as "Implement scam"? Or was there an attempt to hide the true purpose of the requested changes? Were the programmers not bothered by it? Were they in on it too? Are the implemented changes kept in Google's monorepo for everybody to see? Is it hidden somewhere else? Is it hot patched?
(1) BidManipulation is only used in some testing/market simulation environment.
(2) only used in cases where some pricing bug/weird legal issue happened that needs some tweaking to be corrected.
(3) only used in some weird markets where it is legal.
(4) implement many weird obscure features for "next generation ad experience marketplace stuff" and push them to production leaving it to higher ups to configure which combination of them are being used at which market.
(5) empower engineers to come up with additional "whacky ideas" and keep a culture where it is assumed that almost all of them are not passing legal review but they could be enabled at any time.
A DBA I know very well was charged with modeling the accounts for an intentional illegal financial trick (in the 80s for Service Merchandise iirc) that would shift the proceeds from the trick into regular accounts after the statute of limitations ran out. The way it was explained to them was that if they were busted on it they would immediately apologize for the "oversight" and pay up. They did it because they enjoyed being employed.
> They did it because they enjoyed being employed.
Not just that, when the whole product is broken down into features, managed by Feature Managers (aka product manager), that is wrapped up in business speak, and there are whole hierarchies of [Engineering, Product, Project] managers, there are very few people, if any, who understand the whole thing and various side effects and so on.
So in end, in many cases people don’t even realize the nefarious nature of what they’re working on.
On top of that one is surrounded in the echo chamber of fellow employees all looking at the rising stock price and resulting bank balance.
I don't even understand the point of the "feature" as described in the article. Dropping bids (possibly including the second highest bidder), sure, that happens all the time when the creative is deemed inappropriate for the website. Changing the way charging is done and doing all the necessary accounting to create a "slush fund" to use to boost other bids? That doesn't really even make sense. Keep in mind that Google is already choosing the optimal bid prices based on the page & individual so why would they charge their slush fund when just charge the advertiser's account.
It sounds a lot like someone is upset that Google values ads on their site lower than ads on other sites and is upset that Google consistently bids less on their site than for other sites for the same ad (again no shock there as Superbowl advertising costs more than daytime television).
> Google is already choosing the optimal bid prices based on the page & individual so why would they charge their slush fund when just charge the advertiser's account.
To artificially increase the competitiveness of advertisers who use Google's ad tools, as said in the article.
The people in charge are too far removed to care and the grunts implementing it can be compartmentalised. Optimise for the right KPIs. Throw diffusion of responsibility and you're golden.
> Or was there an attempt to hide the true purpose of the requested changes?
Google's custom in-house version control software has an extremely elaborate system of access controls governing who can see which pieces of the monorepo.
I imagine it would be implemented as mundane looking levers and configurations. Especially easy to do in ML powered systems with lots parameters and moving parts.
Except for the part where certain accounts would need to be hardcoded instead of dependent on the advertiser. Moving money is probably the most audited part of any code base and something like that would be a big red flag.
I really think this piece could be better written. Give us an executive summary or even bullet points of the activity. It went too fast into the details for me to see the overall trend.
It is not clear whether Google has fiduciary duty wrt their clients, but if they do and the accusations are proven, they basically looked up every way they could violate it, and did it.
It's not simply corporate responsibility; this is Enron levels of bad. In any decently regulated activity, executives would risk jail for that.
> In any decently regulated activity, executives would risk jail for that.
In the US, executives never go to jail. They negotiate "settlements" where the company doesn't even have to admit wrongdoing and the companies are fined slightly more than they made from the wrong doing.
I would love to to see the executives face jail time, both neither major party political party has any interest in changing things.
I'll grant that post-9/11 priorities have defunded white collar prosecution, but people go to jail in the US. If you really want a criminal-consequence-free jurisdiction, come to Canada. We really don't send people to jail for corporate fraud. The Vancouver exchange is a free-for-all, and Montréal boiler-rooms are a movie cliche.
When investors are the victims, white collar criminals go to jail. We'll soon (re)learn if mere customers have that kind of juice. IANAL, but I'd guess they best they can expect is be party to some class action settlement and the criminals get some token fines.
I don't know if the allegations in this antitrust are true or not. However, about a decade ago I wrote a paper (https://www.cis.upenn.edu/~sga001/papers/vex-sigcomm13.pdf) discussing all the ways in which ad exchanges could misbehave without such actions being detectable and some technical mechanisms to prevent violations. I recall some colleagues telling me that surely ad exchanges would never misbehave because deviations from honest behavior would affect companies' reputations, and ethical people within these companies would whistle-blow immediately.
I'm really curious to see where the government acquired their evidence (and what solid evidence do they have), and whether it ends up holding up in court.
Because of millions of people on HN and elsewhere who think that as long as you can get away with it, anything you do to make yourself or your shareholders more money is just fine.
That maybe true but in my opinion is a lesser evil compared to the corruption of government officials who are complicit in allowing this to happen. What is the point of government agencies if all they do is bow to those with the most money?
Technically, yes, but the FTC is not geared to enforce financial market rules.
Searching FTC rulings on ad exchanges, I found one example of a settlement in a COPPA case.
Implied by your question is the suggestion that the solution is just to abolish government agencies entirely, thus leaving people with no recourse against massive megacorporations, as opposed to trying to fix the problem in some meaningful way.
It's certainly true that there are problems with the current implementation of the mechanics of our government, but the idea that throwing them out entirely would leave us with something better just doesn't pass the sniff test.
> What is the point of government agencies if all they do is bow to those with the most money?
To legitimize the transfer of wealth upward, and protect the system. "Sure didn't we all vote for Trump/Biden". "If you're not voting you can't complain". Etc.
From housing to banking to the environment to food and drugs; from healthcare to the military, there is no shortage of examples of revolving doors, veneers, and gaslighting to choose from.
This is why 'regulatory bodies' are fundamentally a scam.
Tort law remains the worst way to prevent these kind of abuses, except for all the others that have been tried. But the same scam artists (lobbyists, politicians) who convince the public that what we need is just a little more regulation also engage in "tort reform" and it's no coincidence that means removing private causes of action.
"Tort reform" people are also "less regulation" people in every case I can think of. What they usually claim they're for is "better enforcement" while also defunding and reducing the scope and powers of regulatory agencies.
I guess not really. The ad exchange is sort of an auction of auctions; for bare efficiency, every participating network submits only their top two candidates. This is communicated to ad networks; I guess, there is no obligation that the networks pass the bids of their advertisers unchanged. Google has some secret sausage there.
To Google Ads advertisers, Google says "you participate in a second-price auction, competing with other Google Ads advertisers and other networks; if you win you will pay the bid of the runner-up, which is the second highest bid from the Google Ads unless another network submits a bid in between" and that holds true.
To publishers, Google says "the price is determined in a second-price auction, every ad network submits up to two bids"
IANAL but it may be considered as an anti-trust, because it is the owner of the ad exchange who could decide if that obligation exists (regardless of whether it could be plausibly enforced).
Even worse, the sponsored native ad has the same text styling (bright orange title) as the rest of the actual content, so even I struggled to tell where the article continued. This ad actually goes against google policies.
Yep, confused me too! Doesn’t help that I skip over most social media call outs and ads. Way to confuse your reader. As a reader, I’m here to read your work, not expend mental energy on figuring out where your content is.
If only there were evil people somewhere insidiously committing evil deeds and it were necessary only to separate them from the rest of us and destroy them. But the line dividing good and evil cuts through the heart of every human being. And who is willing to destroy a piece of his own heart?
It's a nice aphorism, but it's too defeatist to be of any use. Yes, you have acknowledged human nature. Now what?
The simple truth is that most people do not pursue evil as a goal, but are opportunistic: they take the easiest solution that won't harm them personally. If we're serious about building a better world, reducing opportunities to be evil is what we should be focusing on. Not raising our hands and acting like there's nothing to be done about it.
You provide incentives that make the good part win out, in proportion to the temptation and reproductions of being evil.
Is that so hard? We do it with pets, children, even ourselves, (ice cream after work). But writ large we just say "Nope, they're evil, let's <extreme action>"
When your business depends almost entirely on ads, you have to do anything you can to keep from losing revenue from ads and that results in screwing anyone and everyone and hoping your size makes it impossible to lose if it gets out.
Not long ago, a debate of two Mexican "influencers" [1] was doing the rounds in Mexican networks where one guy was debating another about how Marketing in itself is an evil/bad career: That people who go into the Marketing field basically learn how to deceive and lie to people for the sake of profit.
I agree with that that assertion. I think the current marketing trends are very exploitative and highly misused. Marketing tools like Google, Facebook and Twitter (marketing tools in that their main goal is to take Ads to certain demographics) give advertisers incentives for bad behavior .
business is marketing, not some separable element. if you work, you're working for a marketer. as with any facet of socioeconomic life, the problem isn't marketing so much as it is the perceived prevalence and acceptability of exploitation and dishonesty, and how that influences our individual and collective behaviors.
marketing, at its core, is bringing to market a useful product or service and exchanging it with others who find it more valuable than the cost. this core is net-positive, but on top of that is plenty of leeway to create net-negatives.
we have norms and shared ethics for a reason (to rein in negatives & benefit from positives).
If you go maximizing short term revenue at any cost, you will get evil. It doesn't matter what market you are on.
It's not because of the ads. Plenty of business relied on ads for very long times in the past without going evil. They are evil because they want the profits.
As has been noted elsewhere, if a group has to say "Don't be X" it is because they are X. By definition, a company whose slogan is "Don't be evil" is evil.
Imagine if their motto was "Don't cut the testicles off our male employees." Like, er, sure, obviously. Wait, what? Was that a possibility?
I've wondered this same thing. Were they always the same, but with more effort (as a ratio) towards PR and image, or did the Ruth Porat era signify a cultural shift? Maybe the influx of "tech as a high paying job" type people versus "tech as an inherent interest/tech to solve problems" crowd?
You’re right. The description on this report is faulty, but I think the gist is pretty apparent.
Google tells the website owner that the winning bid was smaller in value than what it really was, thus paying the owner less money.
Google tells the advertisement auction that the winning bid was higher (than what the agreement required), but they only pay the website owner the smaller value. They are paid by the higher bidder and they (Google) keep the difference.
Google can use that excess money however they choose. In this case, they seem to be using it to pump up the value of the second-highest bids (it’s an auction), so that they can receive an even greater difference in the next auction.
WIRED 3 days ago reported on it in a less confusing manner. Also it included a comment from Google
> In a statement, Google said, “AG Paxton’s latest allegation—that we generated a ‘third price auction’ or manipulated our ad exchange—is entirely inaccurate. As of September 2019, we have been running a first price auction, but at the time to which AG Paxton is referring, AdX absolutely was a second price auction.”
I think you are misunderstanding how second-price auctions work. In a second-price auction, the FIRST price bidder pays the bid of the SECOND price bidder. This sentence is saying that the FIRST bidder is still paying the SECOND price rate, but the publisher is only receiving the THIRD bidder's rate. For example:
* Bidder A - Bids $100
* Bidder B - Bids $90
* Bidder C - Bids $80
Under a true second-price auction, the publisher receives $90, and Bidder A pays $90. The article is alleging that Google changed it so that in this scenario, Bidder A pays $90, but the publisher receives $80.
It's two different parties. You've a website. Google says your ad space is worth '3rd place bid' and pays you out 'your share' of 3rd place. The advertiser that 'won' the bid for ads on your site pays out the 2nd place bid' to Google.
The monetary difference between second and third place bids? Google pockets that plus the fee they take for 'brokering' the ad sale.
The author analyzed one section of the pdf which amounts to a competitor creating a product, and so google creates their own competitive product. That sounds like a normal business practice.
It’s kind of silly to write a short article with an outlandish title, and then your only justification for the title is “just read the table of contents”. That pdf probably has tons of supporting information, but this blog post provides none of it.
(The embedded Twitter thread did not load on my initial read through, and it does provide actual content. But again.. embedded Twitter content when you have a brand new pdf filled with info?)
It appears that when google says "All participants in the unified auction, including Authorized Buyers and third-party yield partners, compete equally for each impression on a net basis" they aren't exactly being honest. Am I misinterpreting?
If the point you're making isn't "why mention anything unless you're willing to personally write a book on it" I don't understand it. You mention that adequate links are provided, so I'm confused.
