Hacker News new | past | comments | ask | show | jobs | submit login

From what I keep reading on HN, Facebook ad network ads (which includes Instagram) are by far the best and most precise in terms of targeting, it doesn't even get close.

And from my anecdotal experiences, I have to agree with both you and what I see on HN in that regard. Not only they get the advertisers right up my alley, they even get the exact products I want to click on.

For a specific example: I don't get easily baited by random no-name "hip" clothing startups (that are mostly just alibaba dropship sort of places), so instagram keeps giving me ads for Adidas products as well. And the thing is, not only does it get correctly that I am likely to be interested in Adidas products, the specific products from Adidas that get suggested to me in those ads are the exact kind of products from Adidas that I would be interested in. Which is very impressive and surprising, given how wide the range of Adidas products is, and how most of their general stuff isn't super appealing to me. It is hard to describe to the point where I am struggling myself to define what exactly I am looking for if I am navigating Adidas website. But somehow Instagram ads get it right on target most of the time.

The only time when those ads let me down big time was when I saw an Adidas tracksuit advertised with the design I just liked a ton. Without looking at the details, I ordered it, only to realize a bit later that it was in "kids" section of their website, and I have no kids (and neither do I fall under the typical "people who might have kids" demographic by any metric; e.g., I don't search for any items even tangentially related to children, not a part of any FB groups that are heavily populated by parents or children, etc.). But damn, I would be lying if I said that Instagram didn't get the exact idea of what I wanted perfectly correct, sizing issues aside lol.




> and I have no kids

According to their algorithm, you do now.


With the story of Target figuring out daughter was pregnant before the daughters father, it gives rise to all kinds of "Movie plot" plausible sounding scenarios.

Movie plot: Two people meet up, have fun, then go their separate ways. 9 months later she has a kid, dad doesn't know. But FAANG knows...

The problem with "plausible movie plots" is they are often overrated in terms of % chance of happening.

With linked graphs and relationship trees, it doesn't seem impossible "The Algorithm" could figure out you have kids before you do, or at least base it's Advertising results on a % chance guess.


Which is crazy, but often does happen. I have a nephew and niece, so I've bought baby / small-child stuff, but mainly just around major holidays and their birthdays. And yet I still get plenty of random product suggestions on Amazon for kid-related products, even during the 95% of the year that I'm not looking for that stuff.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: