The more I use social media, the more I’ve realised there’s social media and there’s social media.
Social media has gotten to the point where it is almost ubiquitous and inescapable. I use the main ones—Facebook, Twitter and Instagram (not TikTok)—as well as lesser known/more niche ones, like Ravelry (for all things knitting-related), Letterboxd (for all things movie-related), Clubhouse (mostly so I can listen to comic creators chatting about stuff) and Pinterest. There are ones that are less social and more interest-driven (like Ravelry, but also YouTube and arguably Duo Lingo, Apple Music and Spotify), and there are ones that are more like a private messaging service (like WhatsApp and Slack).
My favourite of the bunch is still Twitter, even after 13 years of being on the platform. I love that through it, I can keep a finger on the pulse of both the comics industry and the publishing industry. I can connect with comic people, writers, editors, publishing houses, Christians and friends. I can tweet and retweet whatever I like, and sometimes the things I tweet/retweet are appreciated by others—even by people I don’t even know.
I should qualify my love of Twitter, though: I love Twitter on Tweetbot, as the way the current algorithm works makes the actual Twitter website the dumpster fire everyone complains about. Viewing Twitter on Tweetbot removes much of the angst: there are no ads/sponsored posts, no Favourites from other people, no algorithm trying to control what you see, and no recommendations of people Twitter thinks you ought to follow. (Which is just ridiculous; why keep adding to your follows? No one could possibly keep up.) On Tweetbot, your feed in reverse chronological order and not much else. (That said, Tweetbot isn’t perfect: unfortunately polls don’t work and threads sometimes break, but if you really care about those, you can always view the tweets in a browser.) I read Twitter every day multiple times a day, and unlike most Twitter users, I read my entire feed (pretty much). But I do read it in actual chronological order—from bottom to top—so it’s not doomscrolling, it’s intentional. And when I reach the top and am up-to-date, I stop, because that’s a natural end point. I don’t need to read anymore.
I really wish there was something like Tweetbot for my least favourite social media network: Facebook. (I like calling it “BookFace” because I am petty.) Like I said, there’s social media and there’s social media, and Facebook is the worst of the lot. It’s large and unwieldy. It’s stuffed full of ads. It’s driven by infuriating algorithms that keep showing me stuff I don’t care about, despite my efforts to train it otherwise (and now that they’ve abolished the feed for Friends Lists, I can’t even do that). And it doesn’t always connect me with the things I care about.
Nevertheless, I can’t really excise Facebook from my life. Facebook connects me with a whole bunch of people who aren’t on my other social networks—family, friends, but also people in Facebook Groups. (The two I enjoy the most are the Sydney Comics Guild and the Australian Speculative Fiction Group.) People use Facebook to message me about various things. Also, I have to use Facebook for work as part of my job involves social media marketing.
Perhaps I’m being too hard on Facebook. Lately, I’ve been trying to work out what it is I dislike about it, and it’s not what people normally dislike about Facebook. I’m not addicted to it and it’s not a time suck for me; because I don’t enjoy it, my engagement with the platform is cursory, and whenever I find myself doom scrolling on Facebook, I tend to wake up and switch over to Twitter on Tweetbot. I know people post all sorts of trash on there, but I tend not to see it, or if I do, I ignore it; that’s their problem and I don’t need to engage with them. Facebook as a company is pretty awful: check out this eye-opening investigative podcast series by The Wall Street Journal for a behind-the-scenes look at how Facebook plays favourites, fails to restrict and moderate disturbing content, facilitates human trafficking, contributes to the body image issues of teen girls (through Instagram, which Facebook owns), and stokes outrage in the name of engagement. Its track record does make me feel uncomfortable about using it and contributing to its ad dollar revenue. But (unfortunately?) it’s not enough to make me break with it.
No, I think my main problem with the platform is the ways in which I’m forced to relate to people while on it.
Consider, firstly, their algorithm, which works using a formula the company calls Meaningful Social Interactions (MSI):
Ryan Knutson: What exactly is MSI?
Keach Hagey: It is a number that measures how much a post is interacted with by people who you are close to. The interactions can be things like comments, likes, re-shares, emojis. And then there’s another mathematical part of it that’s measuring how close the people who are doing that are to you. So it both measures the interactions and the closeness of the people who are doing the interacting and that has an impact on the number.
Ryan Knutson: Facebook used the concept of MSI to create a scoring system. The more likes, comments and shares and the more those happened among people who were close to each other, the higher the MSI score.
Keach Hagey: And in the very beginning, the goal was just simply to get as much MSI as possible.
Ryan Knutson: If MSI is high, that means you’re not just a zombie passively scrolling and watching videos. You’re interacting, you’re engaged.
Keach Hagey: You are more likely to post something, to share a little tidbit about your life if you are more likely to get a comment or a like about it.
Ryan Knutson: There’s nothing more humiliating than sharing something on Facebook or Instagram and getting no response from anybody.
Keach Hagey: Exactly.
Ryan Knutson: The documents Keach reviewed actually break this MSI formula down. It provides a rare glimpse into the inner workings of the algorithm.
Keach Hagey: It’s actually a pretty simple formula. When they rolled it out, a like was worth one point. A reaction or reshare was worth five points. A significant comment was worth 30 points. And then they would add or subtract based on how close the people who were commenting or interacting were. So whether it was a group or a friend or a stranger.
Ryan Knutson: For example, an RSVP to an event was only significant if it was a yes, that would be worth 30 points. But if you RSVP’d maybe or no, it was only worth five points. Facebook would show these significant interactions to more people with the goal of spurring even more engagement.
“There’s nothing more humiliating than sharing something on Facebook or Instagram and getting no response from anybody”. This happens to me a lot: despite having 804 Facebook friends and 101 followers, I will post something and BookFace will respond with silence. No likes, no comments, no nothing. And then, because the post has failed to elicit a reaction—from the Friends closest to me and perhaps from those a little more distant from me—because it’s scored so low on their social metrics, BookFace buries it. And then no one sees it. (Unless they specifically visit my profile, but let’s not split hairs.)
The whole thing is daft because it relies on people’s engagement with me. I totally understand why those 804 BookFace Friends wouldn’t want to engage. Maybe the thing I posted wasn’t relevant or interesting or funny to them. (My sense of humour is a bit odd, and often most people don’t realise I’m making a joke online—perhaps because I’m usually so serious.) Maybe it was about something obscure—for example, sharing my excitement over the upcoming live action Cowboy Bebop series on Netflix , which the majority of my BookFace Friends wouldn’t share. Maybe my BookFace friends were tired or in a hurry or scrolling quickly. The point is, they shouldn’t need to engage with me in order to make piece of content more view-worthy. I shouldn’t need their engagement, their likes, their comments, their whatever in order to be seen on the BookFace platform. I believe that God created humans to be relational beings, which means we’re hard-wired to connect with each other, and the positive side of that is that we can help one another, love one another and care for one another. But the negative side of that is that we can demand things of one another, manipulate one another, and care too much about what other people think of us. I don’t want to be like that on BookFace.
Two years ago, I wrote (on BookFace, of course), “Facebook often feels like a popularity contest I never asked to participate in”, and since then, not much has changed. Don’t get me wrong: I don’t want to be popular on BookFace. I don’t need thousands of likes and comments; I’d never keep up. I just don’t like being ignored. That’s the problem: BookFace makes me feel ignored. It makes me feel like that person at a party who says something and everyone else turns their back on them and pretends nothing happened. It makes me feel invisible—like I don’t exist. Or if I exist, I am barely tolerated.
Now, it’s really important to note that my BookFace friends are not deliberately doing this. They aren’t the kind of people who would shun me at the party for saying something stupid. In real life, they wouldn’t ignore me when I talked to them. And even if they don’t share my (admittedly odd) enthusiasms, they would still understand them and even tell me about things related to them that I might have missed. (Bless you, my good friends, who went out of their way to make sure I knew that Neil Gaiman’s Sandman comic series is coming to Netflix!) It’s the BookFace algorithm that’s doing this.
And sure, I know I have psychological issues. (I mean, who doesn’t?) I know that I react more strongly to feeling ignored and abandoned because of stuff that’s happened to me in the past. Even so, let us acknowledge that the feeling of being ignored is not a nice one full stop, regardless of who you are.
In July of this year, fed up with the way the BookFace algorithm was handling my posts—particularly my public posts (which you’d think would reach a larger audience than my Friends-only ones), I finally caved and created a BookFace Page for myself. I don’t intend it to be a platform-building sort of thing; instead, I wanted somewhere where I could post stuff publicly and have people follow me as posting Public posts to my profile wasn’t working.
At the moment, only 79 people Like the Page and 84 people follow it, which is obviously not a large number, compared to the number of BookFace Friends I have. Even so, I am liking using it far more than my BookFace profile, because the BookFace Business portal lets you see exactly how many people have seen what:
The number of reactions and comments I get on those posts rarely exceeds single digits. But I don’t care because I can see at least I’m not being ignored.
Comparison, the thief of joy
Secondly, BookFace adds insult to injury by seeming to boost my BookFace Friends’ content. I don’t just mean in terms of engagement: from a cursory glance at my News Feed, it seems like my BookFace Friends’ posts attract a higher amount of likes and comments than mine, but I also know that News Feed manipulates what I see because it’s run by BookFace’s algorithm, and that there may be posts of theirs that remain buried like mine. If BookFace operated like Twitter on Tweetbot, I could get more of an idea of whether that’s actually the case.
That said, there have been a couple of instances where BookFace has buried my post and has not seemed to bury my BookFace Friends’ posts. In April of this year (and thankfully way before lockdown!), I went to see Hamilton at the Lyric Theatre with two friends. I took a lovely photo of us in front of one of the posters:
Then I posted it to both Instagram and BookFace. On Instagram, it garnered 15 likes. On Facebook, it garnered 9 reactions and zero comments and therefore (I presume) was buried. My friends also reposted it (with my permission!) to their Facebook profiles, where one got 25 reactions and the other, 62. (She paired it with her review of Hamilton, which probably helped.)
62 reactions! It’s essentially the same piece of content, but look at that difference. I’m lucky if my reactions ever tally beyond 20. Clearly, BookFace didn’t bury her post.
The thing is, I am not in competition with my friend. It’s nice she got 62 reactions! People obviously liked her review. (It was very good!) But it’s hard not to feel like we’re competing against each other when BookFace’s algorithm buries my post and boosts hers.
Thirdly, BookFace doesn’t actually serve me the content I’m actually interested in. In mid-August of this year, I was trying to think through the issue of Christians and medicine (and the related areas of God’s sovereignty and human civic responsibility vs individual freedom) because of a work email I received. I wanted to reply with something useful—something that outlined how Christians should view medicine and medical procedures, written for a popular audience. I asked my BookFace friends for recommendations and also popped the stuff I found useful that I had found through Googling in the comments to that post (which, by the way, only garnered three reactions).
The following day, I discovered that my brother-in-law—who BookFace knows is my brother-in-law (I am listed under his family members in the “About” section of his profile)—had posted the perfect article for my situation on his wall on the day I asked the question. But not only had BookFace failed to show him my post on the subject, BookFace had also failed to show his post to me.
Furthermore, the article in question was “The Suspicion of Science” by Lewis Jones on The Gospel Coalition (Australian edition) website. I’m actually BookFace friends with Lewis. The article went up on TGCA on 16 August (the day I posted my question) and he didn’t post about it until 17 August. Even so, BookFace didn’t show me his post either.
Which leads me to ask, is the BookFace algorithm broken? Isn’t it supposed to serve me the stuff I’m interested in? If it doesn’t, what is the point of it?
Hellooooo? I’m talking to you!
Fourthly and finally, BookFace does not notify the people I care about about my direct interactions with them. Here’s an example from yesterday: while on Twitter, I stumbled across this tweet by journalist Talia Shadwell
The correspondent the BBC sent to cover the petrol shortage this morning is called Phil McCann pic.twitter.com/t64piutcg6
— Talia Shadwell (@TaliaShadwell) September 25, 2021
I posted it to Ben’s wall because I knew he’d find it amusing. (You know: Ben, my husband? Arguably the most important person in my life after Jesus Christ?) BookFace knows we’re married. BookFace failed to notify him.
(How do I know this? Because I noticed that Ben didn’t like the post and though that was weird, and so I straight up asked him today and he had no idea what I was talking about.)
The problem with algorithms
Looking over what I’ve just written, I’m starting to wonder if I’ve been included in one of Facebook’s mood experiments (like the one they did for one week in January 2012). I know that makes me sound like a conspiracy theorist. Still, given the age we live in, it’s not that far-fetched.
Anyway, you can see why all of this has made me quite interested in algorithms. They affect so much of our lives now, I think it’s useful to understand how they work and how we’re being manipulated by them.
At this point, let me reproduce part of a post I put up on BookFace back in June (4 reactions/9 comments), which BookFace buried, because it’s relevant:
I was listening to this NPR Planet Money episode on recommendation engines/algorithms and it was fascinating. They interview Doug Terry, the guy who invented the domain name system for URLs and also, coincidentally, the system that would eventually become the “Like” button on Facebook.
But it started with email: basically Doug was frustrated that his inbox was cluttered with all sorts of things that weren’t necessarily important or relevant to his day-to-day job. So he created a system where you could rate emails that would then prioritise certain emails (e.g. ones from your boss) over others (e.g. spam or forwarded jokes [remember those?]). And then he got his colleagues to try it out. And then he realised that he didn’t have to rate all the emails himself, because his colleagues were rating the exact same emails; he could use their likes and dislikes to filter his inbox in a process that he called “collaborative filtering”. It cut the time they were spending on emails in half, which was a big win for all of them!
Fast forward about 16 years: Netflix was using technology similar to what Doug had invented in its own recommendation system. But it wasn’t improving. So they held a contest called the “Netflix Prize”—one million dollars to the team that could improve their recommendation engine by 10 per cent. And a man called Bob Bell and his team won it by building on Doug’s collaborative filtering technique. They did this by looking not just at explicit data points (e.g. what rating you had given something), but implicit ones (e.g. whether you had rated something at all). From there, they could work out, say, whether you had an interest in science fiction films or police sitcoms. And then they added other layers to the algorithm designed to try and reach other people’s interests.
Which was all very well until, in the wider field, research starting noticing how recommendation engines were affecting our decision-making and preferences. A researcher named Jing Jing Zhang started looking into this and found something interesting: she and her team took the top 100 songs from this Billboard chart and then manipulated the recommendations, which was based on a five-star system. They then told students that the ratings were tailored to their preferences. The students had to listen to the whole song and then were asked if they wanted to buy it. If so, they were then asked how much they’d pay for it. And they found that increasing the star rating on a song increased a student’s willingness to pay for it by between 7 and 17 per cent. Here is the most significant part of the entire episode:
Mary Childs (Planet Money host): The students offered significantly more money for higher-rated songs, even when those ratings were totally manipulated. Jing Jing tested this and retested this. And the results were clear. When a machine tells us that we’re going to like something, we trust the machine more than ourselves.
Kevin Roose (Planet Money host): And, like, look, recommendations aren’t all bad. Sometimes they’re great. They save us time. They help us avoid decision fatigue. Sometimes I just don’t want to, like, manually curate my own playlists of vibey electronic music. But here’s what I worry about. These recommendation systems are getting so good that if we aren’t vigilant, we’re just going to end up drifting toward whatever the machine tells us we like.
Mary Childs: This isn’t just a problem of human psychology. It’s also a computer science problem. Jing Jing says it becomes a feedback loop. Those little drifts add up.
Jing Jing Zhang: Over time, this will make the system less effective, less accurate and provide less diverse recommendations. Eventually, I know this longitudinal impact on the system will make the system provide similar items to everybody, like, regardless of personal test.
“If we aren’t vigilant, we’re just going to end up drifting toward whatever the machine tells us we like”: does anyone find this prediction as chilling as I do? Algorithms are affecting our tastes and preferences as much as we are teaching them about them.
Furthermore, algorithms are starting to edit the world around us for us. This Reply All podcast episode about what makes the TikTok algorithm so good had this insightful tidbit:
A lot of our social media today is only positive sentiment oriented. There’s no dislike button on Facebook. There’s no dislike button, necessarily, on Twitter. And when you only capture positive sentiment, the danger is you have a blind spot to things that mildly annoy or disturb people. In real life, humans are very attuned to this. You know, if you’re with your friends or your family or your significant other, and you do something that bothers them, they might not actively come out and say, “Oh, you’re annoying me,” or something like that. But you pick up on their body language and you realize, you know, and you adjust based on that. That’s a really important feedback loop in just the social world generally.
What happens when we live in a world where we never come across anything that affects us negatively—anything that annoys or irritates us—anything we disagree with? What sort of people will we become? Furthermore, what will we do when we’re faced with that sort of content—or even people who believe in that sort of content? How will we treat them? How will they treat us?
There is social media and there is social media. Some of it is changing us. It’s hard to see because we’re currently swimming in this water. But we need to be aware of what it’s doing to us.
(Postscript: I’m currently dealing with my frustration with BookFace by not posting anything to my Profile wall. However, I haven’t completely given up reading BookFace or posting to my Page. That might not be the most mature response, but I’m interested to see what the platform does with that—and also how it affects my Page stats.)