ann-arcana
@ann-arcana

I mean ... it's lies.

We all know it's lies.

They want the algorithm because the algorithm lets them funnel users toward their real customers: advertisers.

Never mind that every other social media site is presently failing, surely if we do the same thing, it'll work this time.

Meanwhile Google snickers to itself in the corner like Lucy setting up the football.


nex3
@nex3

Annie is definitely right that algorithmic timelines are intrinsically, dramatically more advertiser-friendly in a number of ways, but that's not the whole story. There are two other effects of an algorithmic timeline that are very relevant here:

  1. It makes a user's feed functionally endless, while remaining at least somewhat relevant to their interests.
  2. It gives users who are new to the platform a way to find new people to follow who are at least somewhat relevant to their interests.

Both of these effects are highly user-visible, the first for invested users who just want to engage with the platform more and the second for new users who want to get a sense of what the platform is like. As a consequence of years of algorithmization of the hegemonic social media platforms, they're also firm user expectations among the broad suite of people who, unlike you and me, don't have a principled vendetta against algorithmic timelines as a concept.

Users coming to Cohost absolutely do complain about these things. Despite my strong antipathy towards algorithmic timelines, I have personally complained about how difficult it is to just get more good Cohost posts when I'm sitting around with nothing to do and I've exhausted my timeline. Cohost has made the principled choice not to add an algorithm despite the fact that it would probably be a net positive for short-term growth, and I salute that with all my heart. But it's important to understand that the decision to go algorithmic looks very attractive even from a user-focused perspective when you're mostly listening to feedback from new users who are used to other algorithmic websites.


You must log in to comment.

in reply to @nex3's post:

I mean, the thing about those is though, is that those are both "features" for advertisers too.

I am chronically internet and I get the urge to click more and more but ... so do the platforms, it's what they want. That's what "engagement" is all about: keep you clicking and clicking and clicking so they can serve more people more ads.

In the end there is no growth limit in their world: they must consume all of our attention.

I mean, yes, ultimately everything done by a website whose primary source of profit is advertising is "for advertising" but that doesn't make their explanations of why this might give the website more new users a lie.

I just object to their framing this as if "ease of use" is anything they cared about. If a non-algo timeline served more ads or made advertisers happier, they'd do that instead, regardless of how cumbersome it supposedly is for users.

Are you saying that companies are lying any time they explain their actions in terms of intermediate goals without explicitly saying "and then meeting those intermediate goals will make us profit"? Are you objecting to my framing because it could potentially be read (by someone who pointedly ignores all the indicators to the contrary) to mean that Tumblr is genuinely primarily interested in its users' well-being?

My apologies, I thought by the use of "their" I was clear I was referring to Tumblr's framing, not yours. Because no, I do not extend any corporation that benefit of the doubt as a rule. Public benefit is not what they're for.

i found timelines to be pretty bad, but then i tried tiktok, and well, the algorithm actually works pretty well for a firehose of content, and it doesn't supplant the "following" timeline either

one minor point: outside of being advertiser friendly (hi instagram) or user friendly (tiktok) they're also service friendly

one of the big pushes to algorithm timelines is that for the most part, follower based subscriptions are very expensive to build and very rarely used

Another point I see a lot of people skipping over is that Tumblr specifically mentioned creators in this point as well - asserting that basically, because users pretty much only ever see the people they already follow, new creators tend to be extremely slow to grow and often give up entirely.

I wouldn't be surprised if at least some of that was just an excuse, but it also sounds like what they're outlining is a problem in failing to acquire new creators over the long term b/c too many of them are just failing to attract any audience and burning out early - basically that ALSO becomes a big user acquisition problem for Tumblr.

Not that an algorithmic timeline is going to necessarily be great for the small creators but it does add more reasons for them pursuing it beyond "haha tumblr thinks their users are stupid"

Very true. These are all real problems that algorithmic timelines solve adequately (albeit with a lot of negative side effects) and it's not surprising that Tumblr is reaching for the "industry standard" solution to them. That said, I'm very pleased that Cohost is taking the higher path and looking to blaze the trail to better ways to solve these problems.