Originally comment by Robert Maxwell at https://insurgency.quora.com/Tat…
I think Quora doesn’t know quite what it wants. It’s odd.
See, Quora claims to be a Q&A site dedicated primarily to spreading knowledge. And that’s fine – but Quora’s own design implies that it sees itself far, far more as a social site. Quora tends to isolate writers more often than topics, and writers you’ve followed seem to dominate feeds more than what they’re actually writing about. If you follow Mr. X because you’re interested in, say, France, but Mr. X goes on a spat of answers talking about Trump, you’ll get those answers, too.
On Quora, you’re linked far more often to a person and not a topic of interest. If you build your social network, more people will be exposed to your answers – “power users” are given disproportional influence not due to their expertise, knowledge, or even quality, but due to the visibility as a function of your network. (The argument that “high quality answers attract upvotes and followers,” and hence that one is an indirect expression of the other, is fallacious – but I’ll get to that in a bit.) The fact that people are tied primarily to you as a writer inherently promotes the creation of personae to which people can easily attach themselves: an author who’s able to create a personal connection with a reader is more likely to gain a follower, who will then be redirected to other things that author wrote. See, for instance, the large number of high-follower “power users” that inhabit the survey section.
Quora’s core “quality” mechanic, upvotes, reinforces this, as does the idea of “followers.” No one who’s been on the internet for long could take a look at “upvotes” and confirm that it’s a reliable measure of quality. Especially on sites that promote social connection, those upvotes become an expression not just of quality (if they express quality at all), but their personal connection to the author, and if the author they’re upvoting has interacted with them in the past. Who hasn’t upvoted people only to notice an upvote in return? As to Quora’s follower system, liking an author sufficiently enough to follow them does not inherently mean they are producing intelligent, high quality responses: they may simply agree with you or make you feel good about yourself.
As an aside, I’d say that the only system that works well in even slightly avoiding this is Reddit, which is explicitly decentralized, fragmented, and thrown under a huge veil of anonymity. Even then, though, there are certainly cliques in subreddits.
And all that’s fine. That’s not a critical failing in Quora – it just means that Quora inherently has, abets, and is built around social interaction through the medium of Q&A. That’s not a problem.
But what is a problem is that, despite the system promoting such interaction, normal expressions of that interaction are forbidden. People interact socially in a huge number of ways, and those ways are to a huge (even dominant) degree dependent on the context of the relationship. Social communication is inherently contextual and discarding context in favor of looking only at the message in isolation is inherently inimical to social interaction and to the appropriate policing of social interaction.
If we adopted this standard in the real world (and Quora Moderation, certain individuals argue, is in and part of the real world), we’d be arresting people for openly carrying knives in the street – but we’d also be arresting chefs carrying them in the kitchen. Context is key there, and it requires us as adults to think critically and evaluate non-verbal cues.
I suspect, but cannot prove, that the ultimate justification for this is two-fold.
First, that Quora Administration is badly understaffed, and therefore cannot both pay reasonable attention to the context of a comment and issue moderation decisions at a brisk enough pace. Disregarding context means faster work (in their opinion). I wouldn’t be surprised if this were the case – Quora’s has generally not scaled well at all.
Second, it’s becoming clear that Quora is becoming increasingly reliant on algorithms to supplement their staff shortages – in essence, as an attempt to deal with pressing concerns with scaling. In Marco North’s answer to What do Quora users think of Quora Product Management?, Marco North (as explained in a comment), confirmed via a direct conversation with a member of Quora staff that things such as TW selections are mostly done via algorithm, hence certain rather odd selections for TW. We also know that, to one degree or another, certain aspects of moderation have also been offered up to the algorithm. Social context is difficult for an algorithm and, rather than work on that or subject the reported comment to an investigation of context, it’s easier to simply disregard it and treat all positives as true positives.
Then again, this could be a genuine expression of how they feel policing should be done: without reference to context or relationship on a social networking site.
I think TWs are not selected entirely by algorithm—some glaring omissions from the list give the lie to any strictly objective process. But this is quite insightful, especially the “Q is a social thing but they don’t want to admit it” angle.
Robert Maxwell:
I can’t speak to the degree that algorithms are being used in the selection of TWs – and I certainly don’t find it hard to believe that there’s some bias involved – but relying heavily on algorithms would explain the number of very anomalous TW selections.
Marco North’s explanation according to what he was told (he used to be a volunteer moderator back when Quora had those):
“There was a rash of new TW’s that had less than 100 answers a few years ago. It was explained that they were writing answers in very specific threads, which merited the TW nod. Meanwhile, the answers were truly mundane at best. The algorithm looks for what is “best” in certain categories and sub topics, for example. How did i get to this answer? It started when I asked who edited the three-volume anthology of TW answers a few years ago as there was no editor credit. The admission? Much of the selection was an algorithm, and (shocker) many, many TWs asked “why the hell did you choose THAT answer for the book?” well, it was edited by a fancy server – what do you expect?”
Because we have zero clue on the actual process, it’s basically all hearsay, but it would certainly explain a lot of things. All in all, I think you’re right – there’s some input somewhere, we just don’t know how much latitude that overseer is granted. Perhaps the algorithm comes up with a huge list and they whittle it down; perhaps the algorithm comes up with a list and they’re only allowed to alter it under very specific circumstances.
As with most things when it comes to how Quora works, the answer is that we just don’t know beyond, at best, a rough sketch.