This is a very cynical view of things. I'm not a product manager - but I can't imagine one sitting in a room thinking about all the evil ways to make the feed harder for people to find relevant posts. I can, however, perfectly imagine a PM who is "extremely numbers driven" to look a bad version of the algorithm that increases the time users spend shifting through shit and assume that the algorithm is a huge success because it increased "retention" and reduced bounce-rate.
Seems to me like PM's would think in the most positive terms about the same issue, e.g. what is the maximum amount of ad space that we can use before users lose interest?
No one would directly attempt to be evil, it'd just turn out that way.
“The truly terrible thing is that everybody has their reasons.”
― Jean Renoir
As a PM, though not one that works at Facebook, I would probably approach it something like, how do we ensure that our users stay interested in the homepage, coming back as often as possible and staying on it as long as possible?
There are lots of other sites competing for attention, and to sacrifice a good UX would be shortsighted, especially with all the “FB is dead” comments people have been making the last few years.
And, to be sure, many will justify it by saying "by maximizing this revenue, we're able to provide even more services that people love for free." Is it a Faustian bargain if practically everyone is begging to be Faust?
Well to be frank I've been in these meetings and yes people do intentionally think of dark patterns. If people are willing to sit in war rooms to murder people and Wall Street board rooms to steal pensions, it shouldn't seem so unbelievable that someone could intentionally obfuscate a social media timeline. In fact they use more technical language, such as "engagement", to discuss these topics so that it seems less repugnant.
"Lard it up with debt, fire down to a skeleton-crew and stop all maintenance" sounds bad, like you're killing the company or something. "Unlocking value", hey, doesn't that sound better?
Similarly, "Let's get rich creating the second-coming of AOL by building a tacky nextgen panopticon" doesn't sound that great. "Creating a community of technical greats to bring people together and foster blah blah blah", while a completely bullshit nothing-statement, seems to shift focus off of the surveillance capitalism. For a bit.
Eventually, though, if you make your living sniffing other peoples' panties, eventually they notice and they take measures to limit your access to their laundry. FB and Google are both blocked entirely by IP at my gateway (along with a bunch of other surveillance shops), and the internet at home is so much nicer than what I see other people putting up with.
>I'm not a product manager - but I can't imagine one sitting in a room thinking about all the evil ways to make the feed harder for people to find relevant posts.
Really? Especially a FB PM you mean? Because PMs otherwise have been known to use exactly such techniques and worse:
Easy, launder it through abstract metrics like “engagement” and “ad relevance” and the talks are about what make those numbers go up, not about holding pictures of your friends hostage behind more ads.
Agreed, pathogenic designs can come out of poor success metrics.
Some low level PM was tasked with optimizing time-on-site. They had some negative user feedback, but it didn't show up in their low sample UR and segmented rollout. Since they had good results in their primary success metrics and the secondary metrics did not tell a consistent story, they rolled it out to 100%. Any negative feedback from there was chalked up to change aversion among a small segment of users.
That's still too optimistic. You're assuming that they weren't fully aware of the problem, and would have fixed it if they could. The truth is that they only care about problems that could impact their bottom line.
> I'm not a product manager - but I can't imagine one sitting in a room thinking about all the evil ways to make the feed harder for people to find relevant posts.
Increasing "time spent" means spreading out relevant posts. I'm sure that Facebook is tuned to optimize for some combination of "time spent without drop-off in engagement". This is a "Good" version of the algorithm because it nets them the most revenue, while still fulfilling the need for the larger user base.
It's milkshake marketing[0]. People don't want the most efficient method of content consumption. That's why most "feeds" are no longer chronological. People on Facebook want to scroll around, look at posts, comment on articles, like a few things, etc. for X minutes/day without seeing things they've seen before, and without seeing things that are boring.
The job-to-be-done isn't to consume a certain relevant piece of content, it's to waste time and not get too bored.
They are optimizing for posts that generate revenue. Since users don’t pay, the news feed is optimized for advertisers and the user posts that lead to more advertiser engagement.
True, no one goes around thinking "how can I make my product as obtuse as possible," presumably while twirling their mustache and saying "bwahahaha" occasionally.
The thing to realize is that there's not actually a functional difference between that and the "numbers-driven" guy you describe. Evil isn't just the sadistic maniac; it's the affable businessman who doesn't care, doesn't even consider, whether his product helps people or hurts them, so long as it maximizes profit.
