Technology / 

02 Apr 2019

For Whom The Chronological Timeline Tolls

I recently posted on Mastodon about a theory I’ve been mulling over for some time, about the apparent lifecycle of major social networks, as they move from neat places that scratch a particular itch, to all-consuming vortexes that suck up personal information and promote rage-bait and extremism in the name of driving ‘user engagement’ metrics.

I have a theory: the turning point in any social media service is when it moves away from having a chronological timeline. That’s the shark-jump. When you go from displaying content that the user largely seeks out to actively “curating” or promoting content, that’s a big step. It’s almost always driven by a desire to control what users see. FB/Twitter/Insta have all followed the pattern—they bait-and-switch with a chrono timeline, then mix in the “secret sauce”. Every time.

The switch from chronological to “algorithmic” feeds started with Facebook back in 2009, although it was initially possible for users to avoid it and get the traditional chronological timeline. Between 2009 and 2015, Facebook made it progressively harder and harder to avoid the “algorithmic” version, burying the chronological feed in overflow menus and preventing users from setting it as the default. By 2011, at least as I recall things, it was pretty firmly entrenched.

In retrospect, 2016 was the year of the “algorithmic” feed across the rest of the industry: Instagram, which began as an independent service but was purchased by Facebook in April 2012, moved to a non-chronological feed in that year, as did Twitter. Tumblr held out a bit longer, only getting into the timeline-rigging business in 2017 with their “best stuff first” Dashboard.

Perhaps as a result of persistent commentary that Twitter’s design actively promotes and facilitates what might be politely called ‘low-effort discussion’—what us old folks used to call “flamewars” (or sometimes “harassment” and “stalking” or just “being a real asshole”)—Twitter has recently backed away from the non-chronological timeline, and have started to let users switch back.

The usual defense of the non-chronological, algorithmically-managed feeds is that they’re a necessary evil, because chronological feeds can’t keep up with the volume of posts from people’s ever-expanding social networks. Of course, this explanation doesn’t consider that—absent the change—the “problem” of too much stuff might be self-correcting. If you have a lot of crap in your social media newsfeed, such that it’s burying content you’re actually interested in, maybe it’s time to start muting/unfollowing people, or just dropping them altogether.

Furthermore, it’s not clear that Zuckerberg’s Law—as the prediction that “next year, people will share twice as much information as they share this year” was once called—and which was frequently cited as a justification for non-chronological feeds, holds much water. There is presumably a maximum amount of content and information that the average person wants to share, and as social networks get larger, that amount may actually go down rather than up. But the companies really, really want it to be true. And to try and keep the amount of information they can harvest trending up, they started optimizing for “user engagement”.

And here is where we get to the crux of my argument. It’s not that an algorithmically-driven feed is inherently a bad idea; in some platonic ideal of a universe, it might actually be a useful thing. That depends, of course, on the details of the algorithm, and how well it leads to a maximization of the utility of the service in the lives of its users. But the algorithm that drives Facebook’s feed isn’t designed to maximize the utility of the service in their users’ lives; it’s designed to maximize “user engagement”. And that’s when things get toxic.

The switch from chronological to algorithmic feeds is a symptom; it’s indicative of a shift towards trying to optimize the service so that users give it more and more of their time and attention. And in doing so, it’s the difference between a pocket calculator—something you take out when you need it, use, and then put away; a solution to a problem—and a casino slot machine—something that actively tries to get you to use it, tries to keep you using it, and leaves you poorer, when you finally tear yourself away, than when you started.

A service with a chronological timeline / feed is likely to be a ‘pocket calculator’ or a simple toy; it’s a thing that serves a purpose, and acts in a predictable way. Chronological feeds have nothing up their proverbial sleeve. What you see in your feed is an direct representation of the underlying data. Algorithmic feeds, at least those developed by for-profit companies trying to harvest information and sell advertising, are almost universally indicative of a slot-machine mindset. They provide no obvious indicator of how they work—what does “best stuff first” mean, anyway? And through this opaqueness, they provide a perfect avenue for manipulating what users see. This can be used to drive ad sales, both by raising the visibility of traditional advertising or ‘promoted’ content, and decreasing the visibility of content that isn’t being paid-for. This is pretty sketchy by itself, although one could be forgiven for suggesting that only a very naive person in 2019 wouldn’t expect it.

Worse, in the long run, is promotion of controversial “flamebait” content–including conspiracy theories, hatemongering, stereotype reinforcement, and the like–that necessarily follows user engagement maximization, because flamebait causes people to get angry and engage with the service and with other users (even if that ‘interaction’ is a negative one), spending more and more of their time on it. This is not the sign of a normal itch-scratching, problem-solving product or tool. It’s a dark pattern: something that exploits a particular vulnerability in the human psyche to the user’s detriment.

I’m not an ideologue when it comes to social media; use what you like, as long as you feel like it’s a positive influence on your life. But know when it’s time to jump ship, preferably before the service gets its endlessly-optimized, gamified, attention-seeking hooks into you. And a good heuristic to determine when a particular service has jumped the shark might be paying attention to when it goes from straightforward chronological feeds, to one driven by opaque algorithms and somebody else’s idea of “most interesting” or “best” content.