What's wrong with targeting ads based on subject matter?

TV channels and magazines know how this works. If the page is about fashion, sell ad space to clothing companies. If the page is about travel, sell it to hospitality companies and tourism boards. If the page is about politics or finance, sell it to people making catheters and stair lifts.

It's much less creepy than trying to spy on everybody who reads the page, and then looking for an ad that fits your warped understanding of their preferences.

My cofounder and I used to manage a £100k/year ad budget and we did TV, news papers, radio and Facebook. The level of engagement and ROI on Facebook was insane compared to the others specifically because we could tweak and cohort up the audience so specifically - we even commented how terrifyingly powerful it was and that it was a good job we were only selling a home efficiency central heating product because in “the wrong hands” it could clearly be devastating - you just don’t get that level of granularity with more mass media outlets, even when you break it down by readership/demographic

How much of that was useful engagement though? One of the hot topics a week ago was ad fraud being so prevalent on Facebook and Google that it caused Uber to shift their whole strategy

https://news.ycombinator.com/item?id=25623858

Our sales numbers, the only metric that mattered to the kind of business we ran was 10x even the best of the other methods

Ads based on the subject matter isn't the problem. "Targeted algorithmic advertising" in this context is building a profile of the viewer, and running ads based on that profile.

At its simplest, it's the ad network deciding that you were recently seen visting toyota.com, so it'll show you 4 ads over the next hour for Toyota vehicles, to help cement that association. At its worst, it builds up an individualized profile over the course of months from thousands of websites (with google ads everywhere, you bet!) and allow advertisers to practically target viewers based on a psych profile.

Ah yes, re-targeting is where the real conversion comes. The brand/message with the biggest budget is suddenly everywhere you look

It doesnt work as well as targeted ads. Why would hundreds of billions in ad spend willingly choose a lesser option?

Is there evidence for this?

Google and Facebook being two of the most valuable companies along with hundreds of billions of dollars and petabytes per hour.

What kind of evidence are you looking for exactly?

That it works better. I'm wondering because of this submission the other day:

https://news.ycombinator.com/item?id=25623858

Uber was a victim of fraud, not a failure of advertising. They should've known better and used more checks to catch it quickly but they already run plenty of targeted campaigns precisely because they work well.

If it’s the only option that exists, why wouldn’t they?

> I believe advertising does not increase consumption per-se, but that it redistributes it among a different set of product and services, so who would be the winners and losers of such an effort?

Why do you believe this?