How Good are Facebook Algorithms? A Very Unscientific Test
I have been reflecting on my own relationship with social media since watching the documentary “The Social Dilemma” on Netflix. One of the key ideas in the documentary is about the amount of data these sites collect about us and how they can use it to manipulate us to alter our behaviour into engaging more and more with the site and hopefully clicking some ads along the way.
There is no doubt this is true, but in my case, I do feel somewhat insulated from the advertising. I use an ad blocker and it prevents most ads — except on Facebook, which is apparently one step ahead of my ad block software. In any case, I essentially never click on ads and I don’t remember a time when I purchased something that I saw an ad for online, specifically because I saw the ad. (Of course, I realize I could have been affected by the advertising without knowing it.)
But I decided to look at the five most recent ads on my Facebook feed, wondering how effective these were. How good is the algorithm at knowing what I want or am likely to click on? Most times, I just scroll past the ads, not giving them a second thought, and I realize even in this case, by closely looking at and thinking about them, maybe I will be more affected by them than I might otherwise have been.
So, here are the five ads, and using my very unscientific metric, I rate the ad from 1–10 on how relevant it was to me:
1) The first ad was for Sobey’s (a grocery store chain in Canada), announcing a charity donation thing going in stores. I go to Sobey’s now and then, but mostly I shop at a cheaper grocery store and I’m a bit poor at the moment so I likely wouldn’t be donating in any case! Relevance score: 3
2) The next was for Temptations cat food! I like looking at pictures of cats (who doesn’t?), but I don’t have a cat and don’t foresee getting one anytime soon! Relevance score: 1
3) The next was for Jack Daniel’s. Now, I am a whisky drinker, but I almost exclusively drink Canadian whisky and don’t recall ever having bought a bottle of Jack Daniel’s. Relevance score: 4
4) The next was announcing the shortlist for the “Cundill History Prize,” which I learned is a history prize for best book from McGill University. I actually did click on this one out of curiosity (but I probably wouldn’t have clicked it if I wasn’t doing this “experiment”). Relevance score: 7
5) The final one was Canada Dry premium tonic water. Now, I do drink gin and tonic, but I would probably buy Canada Dry anyway (although maybe not the premium kind!). Relevance score: 5
Based on these highly unscientific metrics, it seems as though the ads were not terribly relevant to me!
But I do wonder whether the algorithm has sort of given up on me. Maybe they realize I don’t often engage with the ads, so they’re just like, “ah, just show him whatever!” Could it be that some people are more susceptible to advertising? Or am I also susceptible to their manipulation — it’s just that the manipulation runs deeper than clicking the ads?
I don’t want to pretend that I am immune to the effects of social media. One of the things from the documentary I really related to was when someone mentioned how scrolling platforms (like Facebook or Twitter) use the same kind of principle as slot machines, namely, that you might not get something good with your first pull, but maybe the next one will be something great. Okay, that one wasn’t good either, but what about the next one? I definitely find this when I use Facebook or Twitter. There is a lot of stuff I’m not interested in, but you keep scrolling down in search of something really cool or relevant. Just scroll down one more time!
I do feel like this has had an effect on my attention span. Ten years ago, I think I was able to focus a lot better on work without feeling the need to always check email or social media. But now, I feel like my concentration sort of slips away and I wonder what’s happening on Twitter, etc. And I can feel it even as I’m writing this!
Of course, there are larger harms of social media and the film does a good job of highlighting the negative mental health effects, particularly for kids and teenagers. They also discuss the way social media can be harnessed for sinister ends, like QAnon conspiracy theories. But one problem I had with the film is that it tries to be too even-handed here, making it seem as though both sides were equally at fault for conspiracy mongering, as if this were a problem for everyone and not especially for the right. It even features clips of speeches from Republicans Marco Rubio and Jeff Flake about the problems of polarization! Are Republicans really in the best position to be lecturing anyone about that? I think it’s true that left-wing people are also susceptible to believing distorted narratives, but it seems like people on the right are much, much more so.
All of these are important issues, but from my perspective, perhaps the biggest problem with social media is not so much the data it has about me, but the fact that it collects data about humans in general and can then use that data to learn things about our psychology that would even be unknowable by the person themselves – and all without anyone’s permission or anyone’s oversight.
I guess at some point we all checked a box and clicked okay… but did we really know what we were getting into?
