Hacker News Reader

Comment by beloch | Hacker News Reader

"But changes to major web browsers to crack down on third-party data collection and new internet privacy rules are making that practice less viable."

Notice that the NYT's new ad system has not been prompted by privacy-minded people who install privacy and ad-blocking plugins. NYT is concerned that coming changes to base browsers will make 3rd party tracking so ineffective for targeting ads that a homebrew solution relying on users checking off boxes that describe themselves will work better.

Given the rapacious ingenuity ad hucksters have shown in the past, I suspect 3rd party data collection will find a way to keep doing it's thing and, likely, become even more invasive, at least for a little while. However, the fact that some companies are finally jumping ship is significant. I've long maintained that the ad industry's war on user privacy was ultimately self-defeating. This might be the beginning of the end.

View on HN

It will likely lead to centralization.

If third-party cookies and device fingerprinting are blocked, then how can you track users who saw an ad twice, before deciding to buy?

You can’t with much certainty, unless they’re using Chrome, and you’re Google.

For the NYT who can rely on “brand” (as opposed to “performance”) advertising this isn’t a big deal. They can charge a lot just to show the image on their site. They have strong enough prestige that many advertisers will not require a third-party arbiter of tracking or impressions.

But I worry that the direction of change will favor Google as an advertising behemoth, and extend their influence over the web and publishing.

Maybe I’m older and from a different time, but conventional advertising was simply based on the topics at hand. Selling camping gear? Advertise in a camping magazine or the travel outdoors section of the newspaper.

These are all completely worthwhile ways to advertise that generated results for decades. This recent change to picking a group of people as opposed to a selection of content providers is very new and many would argue this was pushed heavily by the fact that these companies like Google Facebook and others were able to wedge themselves between the content producers and the content.

I’m not sure we ever truly definitively proved it was worthless to target content subject matter as opposed to personal data.

Sure, they were done, but not nearly as effective - hence all the effort to gather data for targeted ads.

If you want completely integrated ads you'll need about twice as many to get the same revenue. And I'm not even sure what's so terrible about targeted ads in the first place (leaving aside how the data is handled which varies enormously).

While we fill unsold slots with adsense, we make far more revenue with directly sold ads. Either the big ones like Sony everywhere, or smaller merchants and producers targeting specific products (for example: show the ad for this lens on all articles about it, as well as articles about cameras it connects with)

It partly depends on how small your target group is. If you want to reach "People who travel" then targeting the travel page of the NYT should work nicely. But to play devil's advocate, what if you want to reach the 300 oncologists who prescribe your competitor's drug for Specific Rare Cancer and not your drug? Advertising in the cancer journals might be too expensive for your niche brand and you might want to run different messages to different groups within the 300.

That is outmoded. Free services begot personal data (i.e. behavioral surplus) that could be put into algorithms that allows marketing and advertisement firms to bet on our futures. This is truly valuable. Even if the marketing is only 1% more effective, it represents millions in revenue.

https://en.wikipedia.org/wiki/Surveillance_capitalism

Zuboff's book is great.

My concern with describing your ad interests manually is that it will most likely lead to fingerprinting. You’d essentially be building your own unique identifier by choosing ad categories.

If fingerprinting wasn’t an issue (either because the industry was honest or we had effective laws to punish offenders) then I’d say that’s a great idea and could even be implemented in browsers as an “Accept-Advertising” header (just like the headers we already have for language and encoding).

https://science.sciencemag.org/content/347/6221/536

That study gets into the idea of fingerprinting "anonymous" data. I'm working on a study of how people reveal their unique preferences on ways that are visible to advertisers.

> homebrew solution relying on users checking off boxes that describe themselves will work better.

Sounds like a wildly successful website from 2004.

I'm not sure it will be just "users checking off boxes that describe themselves".

I don't see in the article any mention of how they will collect data. While some will definitely come from users, I suspect they will build profiles based on what a user reads.

I wonder how long until they become the third-party data collection service for other companies.

Reply on news.ycombinator.com