Related Content Sucks
Hint: It’s probably because it’s never related.
I’ve been a big fan of content for as long as I can remember. My first job was as an intern at a local magazine, and I ended up leaving because the publisher valued the revenue we generated over the quality of the content. I ended up working for them again later in life, and left yet again for similar reasons. My sixteen year-old-self had it right the first time, go figure.
To be fair to the content publishers out there on the web right now, times are tough. They still haven’t been able to figure out how to properly monetize what they make, and that’s really, truly unfortunate. That’s not something I have any advice around — it’s a tough problem that will need to get solved someday, but I have just as little clue as the next guy as to how that’s going to happen.
We’re a number of generations in to seeing content-related advertising on the web, and it’s not getting any better. We’ve gone from the good-old-days of webrings to linksharing, to ad placement, to content syndication, and now, finally, to recommendation engines. This is becoming an extremely common sight:
So this is the perfect example. This screenshot was taken from right below an article posted on The Atlantic. The article was titled “ Are Student Loans Really Killing the Housing Market?” So, our sponsored content starts off with an article about universities, which seems promising, but then the other two are clearly not even remotely related. The third one doesn’t even make sense, as the photo and the title don’t match at all. I love that in this example the first one seems marginally related, because that’s just dumb luck. Here’s what I got when I refreshed the page:
Yeah, not even mildly related.
To be fair to Taboola, they’re not trying to claim these articles have anything to do with the content on the page, but that’s not really the point. These schemes make me cringe just like “advertorial” advertisements used to back in the day. The goal is to make your advertisement look as much as possible like the native content (without getting too close, because that’s a no-no), in a bid to try to get the reader to believe it’s information just as trustworthy as the rest of the articles you just read.
A strategy such as this irks me especially when it’s from a source that I have come to trust, such as The Atlantic. They’re a well respected organization with a long history of creating some of the better, more thoughtful writing out there, so when I’m reading something on their website I expect it to be of high quality. If I wanted to read poorly written articles that were vapid clickbait, I would have gone to one of the purveyors of one of Taboola’s shitty articles in the first place.
And that’s the real failure, right there. I have no problem at all with the concept. Taboola’s slogan is “Content you may like”, which sounds great to me. I read this great article, now show me something else I might like, Taboola. I just read a 10,000 word piece about genetically modified foods, so yes, of course you would think I would love to read about Katie Holmes’ latest unbelievable shopping spree. That makes a lot of sense. This is clearly content I might like.
If instead, you had suggested four articles from similarly respectable sources that also dealt with agriculture, GMO laws, biochemistry, food science, cuisine, or anything else even remotely related — I would understand and agree that you were living up to your goals, but instead you’re just feeding me whatever random bullshit you've been paid to hawk.
I think there’s a wildly untapped value in truly powerful recommendation engines. If you respect your audience and deeply understand their interests, that’s a proposition that holds a bright, bright future. I read almost every single article posted to Ars Technica. I check it a few times a day, and keep up with it religiously. I do this because Ars Technica seems to have a managing editor and staff that has tastes roughly approximating my own. Their interests are my interests, so most of what they write is interesting to me. I’m a subscriber there, even though they don’t have any paywall, for this exact reason.
There’s no reason a content promotion/recommendation service can’t do this. It’s not rocket science, it’s advertising (and magazine editorial) basics. You need to do your demographic research, segment your markets properly, and have content editors that understand a single fuck about what they’re doing.
What boggles my mind the most is how easy this would be for a company as large as Taboola to do. They’re already on some of the most heavily trafficked websites on the internet, so they already know what I spend time reading there. How hard would it be for them to collect enough information about me to form an opinion that I’m not interested in anything having to do with Katie Homles or pretty pictures of sunsets?
Come on guys, I’m a reader who really wants to find more good stuff to read. You’re a content syndication network who has access to lots of stuff to read. Go take the good stuff, and show me where it is. I’m happily reading, you’re happily making money off my traffic. It’s not hard — it’s your job.
That being said, until these guys get their act together, I’m blocking them outright from my internet experience. I’ve set up an Adblock custom filter that knocks out content from Taboola, Zergnet, Outbrain, Pubexchange, and has a couple of site-specific filters so far. I plan on maintaining it and updating it as I find more and as things change, so it should grow over the coming months.
If you want to keep supporting the sites you love, there’s no reason you can’t turn off all the other Adblock filters and only use this — that’s what I’m doing currently.
You can grab my filter here: http://e0a94a2909382aaa3ec1-24ef4fae91757aa3c51f6506fde75c73.r51.cf1.rackcdn.com/relatedcontent.txt
(Adblock for Chrome allows you to add filter subscriptions via URL, use this with that, or just paste in the ruleset manually)
And if you are so inclined, you’re more than welcome to contribute to it via Github, here: https://github.com/mocha/related-content-blocker
