Google resumes shoveling stuff into its 'Privacy Sandbox'

Google is preparing another round of tests for the latest iteration of its purportedly private-preserving ad technology, after last year's Federated Learning of Cohorts (FLoC) experiment revealed the need for further refinement. In separate messages to Chromium developers declaring their "Intent to Experiment," Google software …

[Totally not a Cylon]

Totally redundant tech.

"The intent of the Topics API is to provide callers (including third-party ad-tech or advertising providers on the page that run script) with coarse-grained advertising topics that the page visitor might currently be interested in," Google says.

Uhmmm, you know what the page visitor is interested in; IT'S WHAT IS ON THE DAMN PAGE!!!!

If I'm reading TV reviews I'm probably thinking of buying a TV.....

[Ben Tasker]

Re: Totally redundant tech.

Ahhh but last week you bought a lawnmower, surely you'd rather have ads showing all the lawnmower deals you've missed out on for the next 6 months?

[Ken Hagan]

Re: Totally redundant tech.

If "you" are the ad provider, you know no such thing. This is just yet another page thst has been polluted with your shit. Unless you read the URL and parse the page contents, you can't tell whether this is a Mary Whitehouse tribute site or a Whitehouse magazine tribute site.

[Eguro]

Re: Totally redundant tech.

Presumably, though, any webpage that displays an ad would have an interest in giving at least coarse topic information about their webpage. After all it's also in their interest for the ad to perform well.

Given time and enough visitors you could potentially refine the topic, based on which ads seem to perform well, all without having to know who clicked any ad, but simply knowing that a given ad was clicked.

[Anonymous Coward]

@Eguro - Re: Totally redundant tech.

Problem is Google sold this to advertisers bragging they know who, when and where the ad was viewed. Now Google can't go back to them and sell less stuff for the same price.

[big_D]

Re: Totally redundant tech.

You set up an API, so that the page calling the ad can provide topic information about the site/page.

[Neil Barnes]

I wonder

What minuscule fraction of the billions of adverts launched actually result in a click through to the serving site; and what minuscule fraction of that results in a sale?

It seems to me that the problem is that advertising is too cheap, not that Google and friends make too much money at it. Perhaps an opportunity for governments to raise some revenue by taxing each and every advertising image that isn't on the originators own site?

(And they should also tax adverts which are sent but not seen, too.)

[Fred Daggy]

Re: I wonder

Internet advertising reminds me of what I've read about the old goldrush days.

Millions went, a lucky few got rich, some scraped by. But the ones that made the money sold the shovels, provisions and the other mining tools.

Internet ads are the same. Millions advertise, a few get rich, eg FAANG and the rest are left with scraps.

It's not going to get better. Certain websites content ratio is getting content down to less than 20% per page with scrolling boxes and ads inter-spaced with text. In which case, I mentally check out of the site. There is even more content on a free throwaway dead tree newspaper one finds on the morning commute!

Firefox or one of its cousins with a good adblocker or two.

[Anonymous Coward]

Re: I wonder

What minuscule fraction of the billions of adverts launched actually result in a click through to the serving site; and what minuscule fraction of that results in a sale?

It varies a lot, but the averages I hear are between 0.2% and 2% for both of these.

Some are much higher, e.g people searching for "car insurance" are so likely to buy one that each click is paid over $10. Conversely, many ads never see a click. It's very top-heavy, meaning comparatively few ads get most traffic.

[AMBxx]

Re: I wonder

I've never intentionally clicked an advert. 100% of the clicks from me are accidental so of no use to anyone other than the company displaying the advert.

I don't know what proportion of the 0.2% to 2% are accidental but I'd guess that with touch screens, it's getting higher.

[Graham Cobb]

Re: I wonder

What minuscule fraction of the billions of adverts launched actually result in a click through to the serving site; and what minuscule fraction of that results in a sale?

Unfortunately, that is the problem which is causing all this tracking. The reality, as I understand it, is that virtually no one clicks on an ad and buys the product in that transaction. So, Google promise advertisers that they can tell them when the ad contributes to a future sale! I can understand why advertisers who believe this would see this as a useful metric - likely to encourage them to spend more on that type of advertising.

That is what this Privacy Sandbox approach is promising to do: it tracks which ads were displayed to you, and it tracks what you bought, and it tells advertisers which ads resulted in a (later) sale. I have no idea whether it really does that very well - but that is what Google is promising advertisers.

At the same time, Google is trying to promise to us that it will not let advertisers know who we are and will not allow (for example) things like web sites adjusting prices because they know what we are looking for or what we have bought.

I have no idea whether what it is proposing can do either of those things. However, I am not willing to even let Google know any of this stuff (what about competitors to Google? do I have to tell them as well?). What I feel like doing today is none of anyone else's business.

[Anonymous Coward]

Re: I wonder

The clever bit of this from Google's point of view is this: the biggest corporations tend to buy the most ads. They also sell the most goods (that's what makes them the biggest, duh). So if I'm Google and I'm selling my yellow pages ad space, it's very easy for me to show that "placing an ad retrieved by a browser" and "browser being used to purchase an advertised product or brand" are highly correlated. Well, of course they are! You have managed to spew thousands of ads into every browser every day, and most of them are for products people buy in great quantities. Even if an ad has negative effectiveness there will still be a strong correlation. Ignorant ad space buyers eat it up, their employers lose money, Google get richer, and the rest of us choke on ads or play the arms race game.

When it was literally the yellow pages, the ads were very, very effective. But they were advertising goods and especially services offered by local providers, not global brand marketing departments. When you opened the yellow pages, you did so because you wanted, or needed, to purchase something and were ready to do so *right there* and *right now*. That kind of ad space was an easy sale to make because anyone could understand the premise and almost certainly had personal experience of it. Now it's all spreadsheet voodoo to trick the gullible. The ads that are still highly effective (precisely because a search for them practically guarantees that the user intends to purchase something *right here* and *right now*) are now so expensive that their buyers give up nearly all their profits. The rest are just marketing budgets set ablaze. Of course someone who was served your ad bought your product 3 months later; you have a 70% market share.

Post hoc ergo propter hoc. Still going strong after three millennia.

[Anonymous Coward]

@Neil Barnes - Re: I wonder

That's irrelevant. Google is not paid for the ads you're clicking on. They're paid on the promise they will show you the ad.

[ArrZarr]

Re: @Neil Barnes - I wonder

Incorrect. Google show ads on a CPM (Cost Per thousand impressions) and a CPC (Cost Per Click) basis.

CPM targeting has been falling out of style for a long time now, and the value of an impression is hotly debated. For some reason, these debates usually have proponents from places that stand to gain the most from CPM models (Those who put far too many ads on site). No idea why it skews that way, of course.

When it comes to spending the client's money, CPM is very efficient. When it comes to making the marketing basically worthwhile, you should just use CPC.

[Anonymous Coward]

And yet can I find a clock that ticks ?

Can I fuck ?

Unless and until Google and their ilk deliver the ability to understand what I mean when I search for a "ticking clock", and not return pages of clocks being marketed proudly as "non ticking" then I think we are pretty safe from the rise of the machines.

(When I was given a tour of IBM Hursley and broke a Watson demo with some questions, I was told that understanding language isn't really an AI problem ...)

[Dinanziame]

Re: And yet can I find a clock that ticks ?

Why would you want a ticking clock?? Are you looking for a gift for your mother-in-law?

[DishonestQuill]

Re: And yet can I find a clock that ticks ?

No, he's in possession of a very young crocodile.

[Norman Nescio]

Ticking?

The AC is not looking for a noisy clock, but one made of ticking. I guess an odd form of plushy (do an internet image search for "IKEA HEMMAHOS clock").