Hacker News Reader

The products that I care about actually do offer personalized ads without tracking: think podcasts (not from major networks, but podcasts like ATP or Relay FM), blogs (like Daring fireball), and YouTubers in an area of interest I have (MKBHD, Linus Tech Tips). I don't even mind sponsored content from these people because more often than not, the products are sensible, tailored to my interests, and actually appealing. Yet zero optimization or learning went into it outside of understanding what the source material is and marketing against that.

Obviously you have a bigger problem when it comes to larger sites like NY Times, but why not just let me say "hey, show me tech products because I like tech" to personalize, rather than trying to infer what I want and get it wrong? I'd be more than happy to tell every ad network ever what I actually care about, but right now all of the optimization is focused on guessing what I care about instead.

Podcasts should be the gold standard for online advertising. They're perfect for everyone involved: there's no creepy tracking involved, and the ads are appealing because podcast producers tend to know and respect their audience. I never skip podcast ads because they're often entertaining in and of themselves (e.g., Reply All).

I like this sort of "personalization" podcasts do in their ads: with human judgment where the content creators exercise editorial control, not some black box chum box any rando can dump ads into.

They're not perfect for everyone. It's usually a huge blackbox for advertisers who don't know what they are getting until they've dipped 50-100k into it and tracked the resulting conversions later in the funnel.

You nailed it. I've had personalized ads disabled for years, and a few years back I switched it back on to see if it's any better.

At the time I was shopping for a Pebble smartwatch and kept checking the site if it's on sale, because Pebble had a tendency to do short-time discounts without actually announcing them. After a few times, almost every Google ad was for "Pebble Smartwatch". It not only wasn't useful, it was annoying.

I'm 100% with this thought. Most "targeted" ads are garbage. I think this is also a chicken and egg problem. I don't want to give up more personal information so the the data set is limited on what i "like".

Instagram has an interesting model. Maybe because of the visual medium that it works well for me. I'll purposely do searches for a thing I'm interested in and click on related posts. After a couple days I'll start getting ads for companies selling things I'm looking for. Recently, new bedding & sheets. I'd imagine it's the same model Facebook uses but i rarely login there.

So in a world where targeted ads are actually useful you would probably opt-in? Would you ever try something like this?

You'd take a survey about your interests and/or things you're in the market for then you would get a monthly email with a variety of targeted ads. Over a year ago I wrote down in my "ideas" notebook "privacy respecting opt-in email advertisements" but then when I looked at it again later on I thought "I don't think anyone would sign up for this" haha.

. This idea goes in the right direction, in my views. I already sign ip to newsletters for things I want repeated reminders for, or updates about. If your service replaced newsletters, and could guarantee to work for things like ticket sales as well as commodity items, it might work.

One of the problems with ads is that they try to jump in front of me while I'm going about my business trying to do something else - it's the wrong time for a product recommendation in any case. I'm happy to go on sites like thewirecutter or cleverhiker when I want specific product recommendations for something. But that only works for products and product types that I already know about. If I could have an inbox of products that were likely to interest me, which I could check or ignore at my leisure (sorta like Spotify's Discover Weekly) - yes, I might just check that out, and even keep coming back if it kept giving good recommendations.

Criticism like this essentially boil down to the adverts not being targeted enough. Lumping you in the wrong group, not knowing enough about your intentions, etc.

An alternative interpretation is that the level of stalking, intrusion, & privacy you would have to give up to be targeted enough to make it useful isn't an acceptable trade off (& isn't even technically possible in the near future).

No, I'd say it boils down to ads being a bad way to discover products you are interested in. I love sites like thewirecutter or similar for getting actual recommendations, when I want to seek something out. Randomly getting products recommended in a drive-by fashion as I surf around the internet doesn't work, but even if it did work, it's neither the format nor timing in which I'd ever want to receive these recommendations anyway.

If targeted ads are no better, and maybe even worse, than untargeted ads, that doesn't imply you need more targeting. Maybe it implies targeting doesn't work (at least for the purpose of providing ads the user wants to see).

I prefer contextual ads, ie. related to what I'm currently looking at, not what my ad profile thinks my personality is based on something unrelated I did on another device last week.

> I prefer contextual ads, ie. related to what I'm currently looking at, not what my ad profile thinks my personality is based on something unrelated I did on another device last week.

That would be my preference for browsing the web, or looking at results of a search. But what about the ads that rely on doom-scrolling? The ones that are inserted every two posts on a Facebook feed would be hard to make contextual, for example.

Reading that, I’m SO glad I left facebook.

I don't mind some Instagram ads which are directly related to accounts I follow. For example, follow some musicians on Instagram and get an advert for some other artist's new album. I don't use Facebook but I guess it could work similarly if you follow certain interest-related groups. However, I wouldn't like if it was using my interactions with friends to target me.

Both of these are better that tracking your activity all over the web and using that to target you.

It is a fair criticism, is it not? My go-to for this issue is my Kindle Paperwhite for which I did not spend the extra $20 to turn off the ads. Amazon has over ten years of my reading habits, which mainly consist of sci-fi, philosophical/religious non-fiction, and miscellaneous dregs. You know what I have never, ever read in my life? Some along the lines of the "Witches of Paragrene Point" series of books. Or the "Defending $POPULAR_90S_GIRL_NAME" series (which I take to be "manly hunk is bodyguard for a female has to rescue her at some point"). But those show up on my Kindle, at which point I assume Amazon just isn't even trying.

After about ten years or so of Kindle ownership, I recall clicking the ad and buying the book exactly once. Yup, a sci-fi book, and it was pretty damned good as I recall. The title, you ask? Hell if I know, Amazon never showed me another book from that author ever again. And I wish they would, because I can't remember which one in my "stacks" of sci-fi is the one I want.

Perhaps you didn't read the last part of that comment:

> So long story short, personalized ads are just as useless to me as un-personalized ads, only creepier, and I'd prefer not to be stalked.

The ad platforms have very clear incentives to deliver the ads that perform best for marketers so that the marketers spend more with the ad platform.

As much as you think the ads are doing a poor job, they're likely doing a better job in aggregate than alternatives.

Retargeting ads deliver higher value than prospecting ads so that's likely why you get so many more of those.

Honest question: Do retargeting ads just appear to perform better because I already had purchase intent and I was already going to go back there to buy it later anyway?

It's hard to tell, since that data isn't publicly available. But I read this article, which makes the case that this is often the case, and that marketing departments are willfully ignoring this, because it's in their interest to appear useful to the rest of their organization, and it's easy to hide using bad statistics: https://thecorrespondent.com/100/the-new-dot-com-bubble-is-h...

This is a completely fair question and is something anyone working in performance optimization (marketing, product, etc.) should be asking themselves all the time when looking at data and finding ways to suss it out.

In the case of retargeting ads, you can run tests to figure it out. When I was doing more on the marketing side, I used to find a non-profit and then show ads for the non-profit to my retargeting audiences. That way we could test to see what kind of value our retargeting ads had vs. a completely irrelevant retargeted ad.

For the most part, the relevant retargeted ad outperformed irrelevant by enough magnitude to make the retargeting campaign profitable.

I think the illusion of "cheap" confuses most marketers. Marketers are left with an unbalanced sense that they are getting higher bang for their buck, ignoring centuries of marketing best practices about expanding reach and exploring non-traditional targets for secondary/tertiary effects. But I think it works well for the ad platforms themselves in the way that it encourages whales just like microtransactions in games. A marketer has an easier time considering opportunity cost for one big ad spend, versus not even paying attention to opportunity cost as they buy a lot of individual micro-targets. (And relatedly not paying attention to how much they are over-spending on people that already have their message versus missing opportunities on greater reach.)

The other aspect of it is these ads are supposedly targeted yet they look like they were produced for a general audience like you'd get on TV. They target the lowest common denominator. They are so safe and inoffensive they end up being simply boring to watch.