Дата публикации: 28-07-2026 03:00:19

Hey! Those are good questions. I don't know of a ton of relevant data, but I'll throw it out to my readers -- does anyone else know of relevant studies/data?I do have this comment data I collected recently:ALTWe can see that nearly three quarters of fanworks on AO3 have fewer than 10 comments. I've looked at popularity metrics a number of times in the past, and when I examine the data from this post in 2014, I see that 12 years ago, over 40% of AO3 works had 0 comments (vs. 25% now). And at that time, only 5% of works had 25+ comments, whereas now 6% of works have 50+ comments.So it looks like maybe things have actually gotten better, right? And possibly they have, in spite of the general sense from authors/creators that they're getting less feedback than they used to. But OTOH, the older works have now been around longer and had time to accumulate a bunch more comments. This histogram can't tell us about the age of the works with more comments, nor how quickly the comments have accumulated.You are also right that fannish attention is probably spread thinner than before. There are a huge number of fandoms on AO3 (76K as of June 2026), and most of those fandoms are very small, meaning that a lot of attention is spread around -- and most people in those fandoms are probably not getting nearly as much feedback as people in some of the really big/really active fandoms. (Though I will note as someone who has made the first/only work in a fandom that the people who DO comment on such a work are often extremely enthusiastic to have found it! :) )ALT(Sidenote: It's worth me & other fandom data enthusiasts noting how few fandoms have 10K+ works, given how many fandom stats are about the big fandoms!) I don't have historical numbers on fandom sizes for comparison. But I expect that there has been a significant shift toward more smaller fandoms on AO3 (and elsewhere), given how the general cultural attention to media has fragmented over time. I have seen indirect evidence of such fragmentation... I recently compared the popular ships from 2013 to those now (forthcoming work!) and shockingly, most of the top 5 ships are the same now as in 2013. Which is undoubtedly due to a variety of factors, but it is consistent with the idea that fannish attention has spread out from the early mega-fandoms to a much wider set of new fandoms and ships now, and so very few new ships have gotten as much attention as the early mega-ships. I do think commenting culture has almost certainly changed some over time, too. But it's a bit tricky figure out how to quantify shifts. Even if we looked at a bunch of random works over AO3 's history and compared their numbers of comments to see how average comment number is changing over time, there are complications. As I mentioned before, newer works haven't had as much time to accumulate comments as older ones. And I've also seen in the past that average numbers of kudos/comments/etc. can vary wildly by fandom/ship/tags. But that said, I could imagine sampling a whole bunch of works from different time periods and looking at how quickly they got their first comment, second comment, etc. If commenting has slowed down, then the time to the Nth comment would be longer now than before. (That would only partially address your questions, but could be interesting.) I kinda expect that maybe that's not true for the new hotness like Heated Rivalry but is true for smaller fandoms, but that's pure speculation!I guess I could also write code to scrape my own AO3 inbox info (and my own Works posted dates) to gather the timing info for all the comments on my works. Then I could look at how commenting rate has changed over time for me and my own stories. But that wouldn't tell me much unless I also got a ton of other people to use my code to scrape their own comment/work timing info and then shared that with me. On the one hand, that could help us look at longitudinal changes on the same creators over time, which would be interesting. On the other hand, it would introduce sampling bias (which creators are sharing their data? probably not a representative set!) and also be a heck of a lot of work. :)I'm curious whether anyone else has ideas for methods of investigating! Maybe there are survey questions that would also be useful to ask, or other ways to investigate. I think the questions around how much writers overlap with commenters (or anything else involving individual commenter identities), are ones I would hesitate to try to address through scraping due to potential privacy issues -- so something like survey questions might be a better approach there. (Note also that I'm not signing up to do further research here, mind you, but it's interesting to at least think through what would provide answers to these questions. :) )One final note: For any histogram of things like number of comments or sizes of fandoms, it's almost always the case that most things are in the smallest bucket due to power laws. But interestingly, that's not what we see with comments when we zoom in:ALTIt's very cool to see that there are more works with two comments than one, more with four than three, and so on. I think this speaks to how much authors appreciate comments: they often reply, and so comments frequently occur in pairs! destinationtoast:kadstats:kadstats:kadharonon:I wonder if checking one-shots posted in specific time spans might result in some interesting patterns. It’s true that stuff that’s been up for longer has longer to accumulate comments, but my anecdotal experience is that if a one-shot is going to get a comment, it’s usually within the first 24 hours or so, while chaptered fics get tossed to the top of the pile every time they get a new chapter and therefore have more chances of getting comments.Gonna go poke at that idea a bit…Okay. I went and pulled complete stuff with expected number of chapters: 1 to get just one-shots, both overall on the archive and in 2-year chunks starting in 2007; the only distinction I was making was whether they had a comment or not.Overall, out of 12,459,761 oneshots, 9,111,944 had at least one comment, so 73.13% had at least one comment and 26.87% had zero comments. Here's what the two-year-segment breakdowns look like.But it could be that the stuff earlier is getting larger numbers of comments than the stuff later on average, so now I have to decide what counts as a large number of comments and where to make my breaking points. Probably on even numbers, because authors replying to most comments will tend to make the counts even more often than odd...Another set of quick and dirty numbers, pulling what percentage of works in 2-year segments have 1-10 comments, 11-100 comments, and 100+ comments:And there's nothing really... notable? Like, you're almost as likely to get between 11–100 comments as you are to get zero comments in most of these date ranges.Maybe it's worth breaking down that 1–10 range into smaller segments—there's a big perception difference between 1 comment and 10 comments (though I was pulling 1 comment on its own for a bit and it was holding steady at about 7% and I decided that was a boring line to have), but it doesn't look as if things have changed THAT dramatically, you know? They've probably changed on a fandom-by-fandom basis, sure, but these lines are much steadier than I was expecting.Wow, those lines do indeed look flattish! I was expecting more of a U-shaped curve (where recent things haven't had as much time to gather so many comments yet, and really old stuff from when AO3 was baby didn't get as much traction, but the middle is a sweet spot of commenty accumulation). Okay, If we discard the early data from before AO3 was in open beta (pre-2009) and zoom in just on your "are there any comments at all?" data, we do see a U-shaped curve:ALTThat's a noticeable dip of the sort I'd expect and is similar to data from @inimitablereel (see my previous reblog), though different methods and different timing yield somewhat different numbers (check out that post for medians with a very nice U-shape, too). But I am still surprised the lines are as flat as they are, especially in the 1-10 range! That is... interesting in its non-interestingness?? XD

Основное содержимое страницы с новостью.

Anonymous asked:

Hello! Thank you for all your hard work~ I was wondering if you were aware of any stats or papers about commenting trends? I always see writers talking about getting less feedback these days, and I wonder if it's actually that people comment less, or if it is that there are more fandoms / writers / works spreading the fans who do comment thinner. Or if commenters used to be more likely to also be writers as well as readers and are now more likely to be one or the other. Thank you!

destinationtoast answered:

Hey! Those are good questions. I don’t know of a ton of relevant data, but I’ll throw it out to my readers – does anyone else know of relevant studies/data?

I do have this comment data I collected recently:

ALT

We can see that nearly three quarters of fanworks on AO3 have fewer than 10 comments. I’ve looked at popularity metrics a number of times in the past, and when I examine the data from this post in 2014, I see that 12 years ago, over 40% of AO3 works had 0 comments (vs. 25% now). And at that time, only 5% of works had 25+ comments, whereas now 6% of works have 50+ comments.

So it looks like maybe things have actually gotten better, right? And possibly they have, in spite of the general sense from authors/creators that they’re getting less feedback than they used to. But OTOH, the older works have now been around longer and had time to accumulate a bunch more comments. This histogram can’t tell us about the age of the works with more comments, nor how quickly the comments have accumulated.

You are also right that fannish attention is probably spread thinner than before. There are a huge number of fandoms on AO3 (76K as of June 2026), and most of those fandoms are very small, meaning that a lot of attention is spread around – and most people in those fandoms are probably not getting nearly as much feedback as people in some of the really big/really active fandoms. (Though I will note as someone who has made the first/only work in a fandom that the people who DO comment on such a work are often extremely enthusiastic to have found it! :) )

ALT

(Sidenote: It’s worth me & other fandom data enthusiasts noting how few fandoms have 10K+ works, given how many fandom stats are about the big fandoms!)

I don’t have historical numbers on fandom sizes for comparison. But I expect that there has been a significant shift toward more smaller fandoms on AO3 (and elsewhere), given how the general cultural attention to media has fragmented over time. I have seen indirect evidence of such fragmentation… I recently compared the popular ships from 2013 to those now (forthcoming work!) and shockingly, most of the top 5 ships are the same now as in 2013. Which is undoubtedly due to a variety of factors, but it is consistent with the idea that fannish attention has spread out from the early mega-fandoms to a much wider set of new fandoms and ships now, and so very few new ships have gotten as much attention as the early mega-ships.

I do think commenting culture has almost certainly changed some over time, too. But it’s a bit tricky figure out how to quantify shifts. Even if we looked at a bunch of random works over AO3 ’s history and compared their numbers of comments to see how average comment number is changing over time, there are complications. As I mentioned before, newer works haven’t had as much time to accumulate comments as older ones. And I’ve also seen in the past that average numbers of kudos/comments/etc. can vary wildly by fandom/ship/tags. But that said, I could imagine sampling a whole bunch of works from different time periods and looking at how quickly they got their first comment, second comment, etc. If commenting has slowed down, then the time to the Nth comment would be longer now than before. (That would only partially address your questions, but could be interesting.) I kinda expect that maybe that’s not true for the new hotness like Heated Rivalry but is true for smaller fandoms, but that’s pure speculation!

I guess I could also write code to scrape my own AO3 inbox info (and my own Works posted dates) to gather the timing info for all the comments on my works. Then I could look at how commenting rate has changed over time for me and my own stories. But that wouldn’t tell me much unless I also got a ton of other people to use my code to scrape their own comment/work timing info and then shared that with me. On the one hand, that could help us look at longitudinal changes on the same creators over time, which would be interesting. On the other hand, it would introduce sampling bias (which creators are sharing their data? probably not a representative set!) and also be a heck of a lot of work. :)

I’m curious whether anyone else has ideas for methods of investigating! Maybe there are survey questions that would also be useful to ask, or other ways to investigate. I think the questions around how much writers overlap with commenters (or anything else involving individual commenter identities), are ones I would hesitate to try to address through scraping due to potential privacy issues – so something like survey questions might be a better approach there.

(Note also that I’m not signing up to do further research here, mind you, but it’s interesting to at least think through what would provide answers to these questions. :) )

One final note: For any histogram of things like number of comments or sizes of fandoms, it’s almost always the case that most things are in the smallest bucket due to power laws. But interestingly, that’s not what we see with comments when we zoom in:

ALT

It’s very cool to see that there are more works with two comments than one, more with four than three, and so on. I think this speaks to how much authors appreciate comments: they often reply, and so comments frequently occur in pairs!

I wonder if checking one-shots posted in specific time spans might result in some interesting patterns. It’s true that stuff that’s been up for longer has longer to accumulate comments, but my anecdotal experience is that if a one-shot is going to get a comment, it’s usually within the first 24 hours or so, while chaptered fics get tossed to the top of the pile every time they get a new chapter and therefore have more chances of getting comments.

Gonna go poke at that idea a bit…

Okay. I went and pulled complete stuff with expected number of chapters: 1 to get just one-shots, both overall on the archive and in 2-year chunks starting in 2007; the only distinction I was making was whether they had a comment or not.

Overall, out of 12,459,761 oneshots, 9,111,944 had at least one comment, so 73.13% had at least one comment and 26.87% had zero comments. Here's what the two-year-segment breakdowns look like.

But it could be that the stuff earlier is getting larger numbers of comments than the stuff later on average, so now I have to decide what counts as a large number of comments and where to make my breaking points. Probably on even numbers, because authors replying to most comments will tend to make the counts even more often than odd...

Another set of quick and dirty numbers, pulling what percentage of works in 2-year segments have 1-10 comments, 11-100 comments, and 100+ comments:

And there's nothing really... notable? Like, you're almost as likely to get between 11–100 comments as you are to get zero comments in most of these date ranges.

Maybe it's worth breaking down that 1–10 range into smaller segments—there's a big perception difference between 1 comment and 10 comments (though I was pulling 1 comment on its own for a bit and it was holding steady at about 7% and I decided that was a boring line to have), but it doesn't look as if things have changed THAT dramatically, you know? They've probably changed on a fandom-by-fandom basis, sure, but these lines are much steadier than I was expecting.

Wow, those lines do indeed look flattish! I was expecting more of a U-shaped curve (where recent things haven't had as much time to gather so many comments yet, and really old stuff from when AO3 was baby didn't get as much traction, but the middle is a sweet spot of commenty accumulation).

Okay, If we discard the early data from before AO3 was in open beta (pre-2009) and zoom in just on your "are there any comments at all?" data, we do see a U-shaped curve:

ALT

That's a noticeable dip of the sort I'd expect and is similar to data from @inimitablereel (see my previous reblog), though different methods and different timing yield somewhat different numbers (check out that post for medians with a very nice U-shape, too). But I am still surprised the lines are as flat as they are, especially in the 1-10 range! That is... interesting in its non-interestingness?? XD

See more posts like this on Tumblr

#other people's analysis #toasty replies #ao3 comments #reader feedback #kadstats

Схожие новости