Here's an amusingly written artical on the uselessness of tracking:

Now to find my favorite quotes...

"It starts with a brilliant idea: we'll collect information about every click someone makes on every page in our app! [...]

And then they do track all that. Tracking it all is easy. Add some log events, dump them into a database, off we go.

But then what? Well, after that, we have to analyze it."

"The state of personalized recommendations is surprisingly terrible. At this point, the top recommendation is always a clickbait rage-creating article about movie stars or whatever Trump did or didn't do in the last 6 hours. Or if not an article, then a video or documentary. That's not what I want to read or to watch, but I sometimes get sucked in anyway, and then it's recommendation apocalypse time, because the algorithm now thinks I like reading about Trump, and now everything is Trump."

"This is, by the way, the dirty secret of the machine learning movement: almost everything produced by ML could have been produced, more cheaply, using a very dumb heuristic you coded up by hand, because mostly the ML is trained by feeding it examples of what humans did while following a very dumb heuristic. There's no magic here."

"Even today you can still use every search engine web site without logging in. They all still serve ads targeted to your search keyword."

Read it for more!

@alcinnz thanks for the quotes! It's made me add the article to my to-read list 📃

@alcinnz Or when you read about your favourite foobtall team's next **opponent** and then the recommendations you get are all people you can't stand the sight of.

@alcinnz This is a serious mistake by simpleminded engineers who can't tell the social graph apart from the web graph. They're converging, technologically, but they're not one and the same thing.

@alcinnz You really need to stop going to whatever evil site that is.
You can rule your own little world here in the fediverse. No drama or manipulation.

@gemlog Tell that to the author of the blogpost I'm quoting.

And seriously, read it! He's encouraging developers to handcode simpler heuristics so they know what data they actually need, not get in trouble when it inevitably leaks, and yet end up with better recommendations.

Not to mention startup time. It takes a lot of data before ML actually starts to kick-in, but a handcoded heuristic can kick in after just a few samples. The blogpost cites Pandora web radio as a great example of pulling this off.

@alcinnz I can confirm that.

When I worked in digital agencies in the 2000s we had several clients who spent big dollars on collecting data for every interaction, but there was no budget to analyze - or even store - any of it.

I personally deleted giant databases with email addresses from ad campaigns where nobody had thought of a follow-up to their cunning marketing strategy.

@alcinnz I’ve been telling folks this for years! I used to work as an advertising operations engineer for A Big Software Publisher; the Entire display Ad industry is a scam, top-to-bottom. It’s not just that tracking is garbage, but in a twist perfectly suited to our capitalist hellscape, the orgs that pay for the “services” know it’s garbage too.

Sign in to participate in the conversation

For people who care about, support, or build Free, Libre, and Open Source Software (FLOSS).