A snapshot of the popularity of various stories in the last week · 12:31pm May 30th, 2014
Because I felt like doing this analysis, I randomly went through all of the stories submitted on the site for the last week and did a tabulation of the numbers for various pertinent categories of stories; they have been recorded here for the sake of posterity.
The FIMFiction story numbers were taken for the week prior to 5-20-2014 at 3:00 am Pacific time. There were 550 stories in the sample.
Rating
172 Everyone
248 Teen
44 Teen (sex)
24 Mature (no sex)
62 Mature (sex)
Clopfics
62 Mature (sex) - While mature-sex is not 100% clopfics, it is the best available approximation of it, but is doubtless an overestimate.
Breakdown - bolded categories represent stories where the majority of mature-rated stories had the sex tag
43 Mature - Romance (sex)
22 Mature - Human (sex)
21 Mature - Slice of Life (sex)
20 Mature - Comedy (sex)
12 Mature - Alternate Universe (sex)
12 Mature - Dark (sex)
11 Mature - Random (sex)
9 Mature - Gore (sex)
8 Mature - Anthro (sex)
5 Mature - Adventure (sex)
4 Mature - Crossover (sex)
3 Mature - Tragedy (sex)
3 Mature - Sad (sex)
Non-clop fics
444 stories without the sex tag
44 teen (sex) rated fics
Breakdown - bolded categories represent stories where the majority of mature-rated stories did not have the sex tag.
184 Slice of Life
79 Everyone - Slice of Life
81 Teen - Slice of Life
4 Mature - Slice of Life (no sex)
156 Adventure
36 Everyone - Adventure
110 Teen - Adventure
10 Mature - Adventure (no sex)
145 Comedy
49 Everyone - Comedy
72 Teen - Comedy (no sex)
21 Teen - Comedy (sex)
3 Mature - Comedy (no sex)
143 Dark
15 Everyone - Dark
111 Teen - Dark
17 Mature - Dark (no sex)
133 Romance
47 Everyone - Romance
54 Teen - Romance (no sex)
26 Teen - Romance (sex)
6 Mature - Romance (no sex)
133 Alternate Universe
26 Everyone - Alternate Universe
98 Teen - Alternate Universe
9 Mature - Alternate Universe
101 Sad
32 Everyone - Sad
65 Teen - Sad
4 Mature - Sad (no sex)
98 Random
39 Everyone - Random
55 Teen - Random
4 Mature - Random (no sex)
98 Gore
84 Teen - Gore
14 Mature - Gore
92 Human
17 Everyone - Human
62 Teen - Human (no sex)
7 Teen - Human (sex)
6 Mature - Human (no sex)
80 Crossover
11 Everyone - Crossover
60 Teen - Crossover
9 Mature - Crossover (no sex)
63 Tragedy
13 Everyone - Tragedy
43 Teen - Tragedy
7 Mature - Tragedy (no sex)
16 Anthro
3 Everyone - Anthro
7 Teen - Anthro (no sex)
6 Teen - Anthro (sex)
3 Mature - Anthro (no sex)
Discussion of results
So, what is there to say about these stories?
It is a very eclectic collection. All of the tags were used to a fairly great extent, and most of the tags were actually pretty frequently used.
Contrary to popular belief, clopfics are apparently not an overwhelmingly popular category of stories to write - there were only 62 mature stories with the sex tag in the sample of 550 stories, meaning that mature (sex) rated stories were actually less popular than the Tragedy tag, which was the second least popular story tag, above only Anthro. The Sex tag as an overall category clocked in between Human and Random in terms of popularity.
Likewise, in terms of rating, they did not fare so well - none of the ten highest rated stories from this week were sex-tagged mature stories. However, the distribution of sex-rated stories was skewed slightly high overall - there were 46 sex-tagged stories in the top 200 rated stories for the week, another 35 in the second 200, and 25 in the final 150, suggesting a slight bias towards them being higher rated (on average) than other stories. However, the bias was pretty small. The highest rated mature (sex) fic came in at #11, the next highest at #24, then #28, with #34 being the point at which sex-tagged fics started to get more common. Interestingly, #11, #24, and #34 are mature-sex rated but do not yet have any sex in them, though I strongly suspect at least #24 and #34 will. It is hard to say on #11 - Headless, Not Brainless - as while it possesses the Mature (Sex) rating, glancing through it it appears to be... well, first off a very strange premise and second off full of profanity but not sexualized at all.
Looking at them in terms of views, though, tells a very different story; the most viewed story was a Slice of Life story (mine, actually), but 5 of the top ten most viewed stories for the week were, in fact, clop (though one of them does not actually feature any clop as of yet). Indeed, of the 200 most VIEWED stories for the week, 60 of them were fics tagged with sex, 43 of which were mature (sex). There were only 38 stories with over 1,000 views this week; of those, 19 of them were mature (sex).
On the other end of the spectrum, Gore stories fared quite poorly - while a fairly popular category, the highest rated gore story clocked in at #22, and of the 107 gore-fics, only 24 of them made even the top 200 rated fics for the week. Looking at it through the lens of views does it few favors; only a single gore fic clocked in north of 1,000 views. Gore does not seem to be terribly popular amongst the FIMFiction audience.
The same cannot be said for Slice of Life stories. This category was not only the most popular category of stories to write by a fair margin, but it also fared very well in terms of ratings - just under half (86 of 205) of the top 200 rated fics for the week were Slice of Life, and of the ten highest rated stories this week, 9 were Slice of Life, 4 of which had it as their only tag. The sole non-slice of life fic in the top 10 for the week was a Comedy fic. They likewise held up in terms of viewership, with 14 of 19 non-mature (sex) rated stories with north of 1,000 views being Slice of Life, and an additional 8 mature (sex) rated stories likewise sporting the tag.
In terms of other tags represented in the top ten rated stories, there were 3 comedy stories, two romance stories, one random story, one sad story, and one human story. Adventure, Dark, Alternate Universe, Gore, Crossover, Tragedy, Anthro, and (as noted previously) Sex all failed to make the top 10 in terms of ratings, and looking through it through the lens of views, only Anthro and Alternate Universe would pick up anything... but both for the sake of clop, though both do appear in the next ten by views.
Given the popularity of Adventure, Dark, and Alternate Universe, it is a bit surprising that none of them made the top 10 at all; Dark had several representatives in the top 20, and a Dark Adventure fic was #13, but only 37 of the top 200 rated fics for the week were Adventure stories, 39 were Alternate Universe, and 40 of them were Dark - all three categories underperformed relative to their popularity, though the significance of this is difficult to determine. In terms of viewership, they fared little better; only 3 Adventure fics and 3 (non-mature) Alternate Universe fics broke 1,000 views. On the other hand, Dark managed to get 5 such stories, as many as Comedy did.
Overall, the tags seem to fall into a few general categories - Slice of Life is king, with Adventure, Comedy, Dark, Romance, and Alternate Universe seeming to make up the next group. A fair bit below them lies Sad, Random, Gore, Human, and Crossover, with Tragedy bringing up the rear and Anthro being practically nonexistent - indeed, there were only 30 Anthro-tagged stories over the time-period, and of those, 14 had the sex tag (though only 8 were both mature rated and sex).
Being on the bottom, Tragedy obviously isn't a hugely popular story tag (indeed, sporting less than a third of the popularity of Slice of Life), but it actually is worse than that - while it has 66 tagged stories, a whopping 34 of them (over half) are also tagged Sad. They also do not fare well in terms of ratings - none of them are even in the top 50 rated fics for the week, and only one of them managed more than 100 upvotes.
Length, completion, and views
Unfortunately, these stories being NEW stories makes it hard to look at some of their other features. For instance, I considered looking at their length, but rather quickly realized that was a poor idea - of the 550 stories, 469 of them were less than 7500 words in length. Another 63 were 7501-20,000 words, 10 were 20,001-40,000 words, and 1 was north of 40,000 words (at a whopping 47,828 words in a week), and I'm not sure where the remaining 7 stories disappeared to (perhaps beyond the time horizon as I worked on this post). Only 205 of the stories were complete, though, so these numbers are likely to go up over time, making it difficult to draw any sort of useful conclusions from them.
That being said, being complete seemed to help stories' ratings - while only 205 of the 550 stories were marked as complete, 7 of the 10 top rated fics for the week were marked as complete.
In terms of overall views, the distribution appears to more or less follow the Pareto Principle; only 38 stories managed to get 1,000+ views, another 31 getting between 500-1000 views, and 47 getting 250-500 views. Another 170 stories got 100+ views, and the remaining 264 stories - nearly half of the stories published on the site in the last week - got less than 100 views. These numbers are likely to increase over time, and are not perfectly representative, but it is quite striking - the bottom stories got about 90,000 views between them, while the 1,000+ views stories got 86,294 views between them - in other words, the most viewed stories got about half the views of all the new stories between them.
Unfortunately, it is rather difficult to get a good handle on what percentage of site views that is overall - according to the tracking numbers on the site statistics, there were about 3.6 million story views, but because those 180,000 views represent the peak views of the individual stories it is hard to say what percentage of the overall site viewing that really represents.
Conclusions
So what conclusions can we draw from all this?
The clearest conclusion is that clopfics and slice of life stories appear to be by far and away the most popular types of story on the site - both dominate the over 1,000 views category, and slice of life stories are very well regarded in general rating-wise. Slice of Life stories are the most popular type of story to write, while clopfics appear to be very popular relative to how frequently they are written - there is a high demand for these sorts of stories.
The second most clear conclusion is that clopfics very highly viewed, but not very highly rated. Indeed, it seems clear that while view counts and ratings, while tied together, are not the same thing; the most highly rated story for the week has a whopping 866 upvotes on 4,939 views, but the next most upvoted story - a clopfic - has 4,822 views but only 459 upvotes. Indeed, the clopfics in general seem to have a much worse upvote conversion ratio, with most of them clocking in at an upvote per every 8-10 views, whereas the non-clopfics seem to convert upvotes at a rate closer to 1 for every 5 views. Additionally, the clopfics accumulate downvotes at a massively higher rate - the three most viewed clopfics managed to accumulate a whopping 64, 75, and 52 downvotes respectively. All of this seems to contribute to their lower ratings.
Thirdly, viewership appears to fall off very rapidly - the stories with more than a thousand views make up nearly half of the views that all new fics on the site get. Most stories on the site get extremely few views overall, and only a tiny fraction of new stories - probably on the order of 10% - get north of 1,000 views. Overall, the viewership distribution of new stories appears to match the Pareto Principle fairly well.
Finally, it seems that the mature category is very strongly correlated with the sex tag, with three-quarters of mature fics also bearing the sex tag.
It would be interesting to repeat this experiment, but it took me about three hours to put this post together, so I'm not sure how inclined I am to do it again next week. Still, it is interesting to mull over.
One very random aside: One story this week managed to acquire a whopping 70 downvotes off of only 186 views.
Can this help me become more popular?
Probably not.
Looking through the data, the most strikingly clear piece of data was this:
People who write popular stories have written them before.
Of the 10 highest rated stories from the last week, 8 of them were written by people who had hit the feature box previously - of the other two, one was a brand new writer and the other had about as many stories under their belt as I do. Of the 10 most viewed stories from the last week, again, 8 were written by people who had hit the feature box previously - and the remaining two were both brand-new writers.
And if we are going to go with "well, of course the good stories are that way, but clop is drivel", of the 10 most viewed mature-sex stories from the last week, 9 of them were written by people who had hit the feature box previously - the only one who hadn't had half as many upvotes and had by far the fewest views of the top 10.
Thus, drawing conclusions from the data about what one should write in order to attain popularity is questionable, as it may well be that what sits in the feature box in any given week is determined not by the audience, but by what the writers who can draw an audience (or who are best able to promote stories - because RainbowBob's signal boosting of my own story may well be what pushed it over the top and got it into the feature box in the first place) wanted to write/promote that week.
Well, this is very highly detailed, but I cannot say that this is anything really new. Clop has always been more popular, though the amount of clop actually written is rather low. However, it is basic supply and demand, where the demand is high but supply low, meaning that clop fics will generally be featured faster.
Yet that seems to be the opposite of Slice of Life. It seems to be the most popular and the most highly written, meaning that the supply matches the demand evenly. Though, that doesn't mean that every story is going to be popular.
Your observation of the Pareto Principle is valid indeed; it seems that popular stories take up half of the views, while the rest take up . . . the rest. However, this brings up a larger and probably more intensive study that would be interesting to look into: how many of the most popular stories were by authors that are considered to be popular? I would think that more than 300 followers would suffice to call someone relatively popular. But you might also look into whether or not their individual heats have a part in making their stories get featured.
I remember reading the formula by that one guy I forget the name of that made a massive blog post about it, and I think it is followers/story views or something. I will need to dig up the heat index formula later just for personal intrigue.
In the end, though, this blog post was indeed very interesting. I love how anthro has the least amount of stories overall even though an entire tag was made for it thanks to Equestria Girls (and most likely the cries of those who would like to write anthro fics) coming out.
But I would be interested in this being a weekly segment, if only to create a pool of data of which we can once and for all extrapolate real data to prove or disprove fandom stigma against clop fics or anything else.
2158712
I added a section to the document about this, but unfortunately it is, again, difficult to draw any useful conclusions from this because I didn't go through and look at which stories DIDN'T get very popular - maybe some folks who have hit the feature box wrote some of the real stinkers this week as well. It is hard to say for certain. But the fact that the top folks overwhelmingly were people who had previously written popular things suggests that who is writing the stories may well have a very large impact on what stories end up popular in a given week. That being said, that didn't necessarily mean they all had piles of followers; there were four people in the top 10 with under 400 followers, and two of them had only double digit numbers of followers, likely entirely from their new hit story.
I do know that RainbowBob (may his pores never dry) wrote a story this week that got under 1,000 views despite the fact that he has 2,604 followers, and I know bats writes stories from time to time that don't get featured (or don't get featured for very long), so there is certainly evidence that writing stories which get featured and having gigantic piles of followers does not ensure that every story you write gets featured, though having lots of followers does, I suspect, lower the bar in terms of quality that you are allowed to put out and get however many hits.
It also is harder to tell whether or not someone got signal boosted; I got signal boosted by Bob this week, for instance, so it is hard for me to tell whether or not my being featured was a result of my story itself or because Bob signal boosted it; it was in Popular Stories at the time of the boost, and then overnight went up into the feature box. Would it have made it without Bob's blog post? That's hard to say.
Was it the best story on the site in the last week? It was both the highest rated and the most viewed. But I'm not sure that it would have been those things if Bob hadn't promoted it. On the other hand, Bob probably wouldn't have promoted it if he didn't think it was good, so it is kind of a chicken and egg thing.
2158749
If I could find the heat index formula, it actually proves that how much heat someone generates has an effect on whether or not they enter the feature box at all. It actually showed that Pen Stroke has a rather low heat index even though he has tons upon tons of followers. The obvious consensus is that he has heaping loads of inactive followers, but it seems that this index formula compensates for that and creates an accurate number that describes the ability for one person to enter the box.
Anyway, you do bring up a valid point. I actually found your story before Bob made his blog post, but it was definitely his post that made me read it without letting it sit in my read later folder.
This just gets more and more complicated, then. I do believe that there is no real way to be sure of anything with people getting into the feature box other than attributing it to a good story or . . . clop. As they say, "sex sells."
2158749
I at least have the formula, though who it is from I don't know. He is a popular author, so it shouldn't be too hard to find, but I really don't care to find it.
Heat Index = Follower Count/(Total Story View Count/100)
That puts your heat index at .997, and that puts mine at 4.776.
2158769
I actually keep track of when I gain followers, so I actually have some idea of where they come from. I had 143 followers prior to the release of my latest story; I now have 197. I have gained, thus, 54 followers from that story.
The second place recruiter for me is Shotgun Wedding, which got me 29 followers. The Collected Poems of Maud Pie got me 19 followers.
My other stories have gained me about 8 or 9 followers apiece, though it is hard to disentangle how many followers I gained from The Wraith of Ponyville and Proper Anatomical Terminology respectively due to having close release dates, getting 1-2 followers on release, and then the next week getting 9 followers which were hard to attribute.
I also got 11 followers from a promotion by RainbowBob a while ago.
The remainder trickled in.
I also got my first six followers before I posted any stories at all.
2158789
2158790
That double post tho . . .
But interesting nonetheless. I actually have personally gained most of my followers to do being active all over the site, whether it be through constantly being the first to comment on a Writer's Group thread or editing. My stories honestly haven't done much for me overall, but they have had an effect nonetheless. XP
The distribution is roughly power-law, which means that the longer you observe, the more views are concentrated in the top stories. I did a complete count of all stories published on fimfiction once, and I think the top 2% of authors got half of the views. That would almost certainly mean less than 1% of the stories got half the views.
Holy crud on a shoe, this is useful.
I've wanted to do something like this for a while, but I don't have the wherewithal. Thank you for doing it.
2158829
I ended up spending far too long looking at this statement. You are right (or at least very close to it). I spent way too much energy obsessing over it.
I think there's something extremely anomalous going on with the extreme end of the curve, though, because looking at stories like My Little Dashie and their present viewership rate, and doing some back-calculation, it has "too many" views, especially given that our site readership is massively higher than it was when it came out.
Indeed, of the stories with 25k+ views, only six of them (debatably) are 2013 era stories, and zero started in 2014. Even of the 2013 stories, several started in 2012.
Today, according to the site statistics, we see 500,000 views/day, which is more views than My Little Dashie has... every single day.
I dunno. Something feels wrong to me that all of these stories are so old.
Then again, the stories with multiple chapters also have that ridiculous multiplier where you've got 100k views times twenty chapters, minus people who quit midway through and end up with 500k-1M views overall. Anthropology has like 1.1M chapter views in total, which means that it makes up about 0.3% of the overall chapter views on FIMFic by itself, so maybe it all makes sense after all. University Days still gets like 600 views/day, which is 0.1% of the overall chapter views on FIMFiction. You don't need that many stories like that to end up adding up to 50% of the overall site.
Though I'm still wondering why there aren't more recent stories with obscene viewcounts.
2158918
Isn't that pretty likely to be related to the expanding number of stories? We have more views now, but we also have more stories, so those views will be more distributed (even if not evenly distributed.)
For example, Bad Horse mentioned that the top 2% had most of the views at the time he looked... that might still hold true, but now the top two percent includes more stories. Add to that that views are cumulative, and you're more likely to have 25k+ views on a story that was published three years ago, with much less competition at the time. Every single day a writer has more competition for a view and is still competing against stories that had less competition when they came out many of which now have a reputation (especially in those 25k+ stories.)
The question then, would be how views per day relate to stories published: If you could show that the 25k+ stories had the same ratio of overall story views on the site to stories published that we have today, then it would be an anomaly. But I suspect that the earlier in fandom you go, the more views there were for each story published (again, not evenly distributed.)
That breakdown verified a lot of my observations about this site and also answered a few questions I've been wondering about. Thanks.
2158829
Wow, something even more lopsided than American wealth distrtibution!
2158749
You mentioned in the post that most featurebox authors have gotten featureboxed before — how do you track that? Is there a way to tell whether a given story once upon a time reached the 'box?
In any case, interesting statistics. It might be interesting, even if you don't do it week over week, to take snapshots over time — I wouldn't be surprised to learn that story genres trend in and out. I know it used to feel like the "X Verbs a Y" comedy was a lot more prevalent than your numbers say, and comedies in general; Slice as 9/10 of the top stories seems like an outlier, but I don't have data to back that up.
Yes it can. We've all learned two things from this. In order to become more popular, you must:
1. Write sex.
2. Already be popular.
Easy peasy lemon squeezy.
2159013
I think the real problem is in part that I screwed up while looking at these; now that I'm looking at their charts in more detail, more of them obviously DID update (or get sequels) in 2013/2014.
That being said, a lot of them did not.
Looking at story viewership distribution, a story gets the most views at its start, and a declining number of views over time - indeed, it seems to mostly follower a power law curve on a day to day basis (i.e. the older a story, the fewer views it gets). If we look at a lot of the stories, over half of their upvotes (and thus, presumably at least, views) came early in their history; My Little Dashie, for instance, accumulated half of its upvotes by October 2012. Anthropology is the same way, so is Princess Molestia and Cheerilee's Garden. Thunder Struck is a new story (new being 2013), has vastly more views/day than the stories above it (even the chapter stories above it), and yet it acquired about 2/3rds of its upvotes when it was posted, showing only fairly modest growth since then. And indeed, looking at the curve of most of these stories, they seem to be acquiring declining number of views over time, assuming that their upvote curves are representative of their views. The only real exception to that trend I found on a story which wasn't updating or having sequels written about it (that I'm aware of) was Past Sins, which appears to be increasing in popularity fairly linearly. Even amongst the stories which are still updating, though, it still seems to be the case that a lot of their popularity is very old.
Added to this, the overall chapter views per day over time, according to the site statistics, has gone up faster than the number of stories submitted has gone up; in fact, stories per day actually appears to have diminished somewhat in 2014 in terms of the rate of approval (though the overall number of words appearing on the site appears to be constant).
This would seem to imply, to me at least, that we actually have more viewers with fewer new stories being thrown at them, which should, in principle, help drive those stories to greater heights; while it is true that it takes a while to acquire 25k views, on the other hand a story which hits the top of the feature box and stays there can accumulate five thousand story views in its time there, and thus many thousands of views per chapter if it is a chapter story.
I think there's a few things at work here. One is the feature box bug - looking at the stories which DID update, they all got quite considerable boosts every time. This would naturally favor stories which started in 2012 and continued to update, which may explain the prevalence of such stories in these lofty heights - over time, now that the bug has been fixed, we would expect this to reverse itself.
The second, as you noted, is a change in viewership behavior; I suspect here that the proximate cause may be Equestria Daily and the rise and fall thereof. In the past, it may be that Equestria Daily was significantly more important than the feature box; nowadays, it is probably somewhat worse than getting featured. This would hyperpromote a relatively small number of stories, which may explain part of the phenomenon - however, the one problem with that that theory is the persistent popularity of some of these stories. People aren't looking through old EQD posts to find MLD; they are finding it via Google and YouTue primarily, apparently. While people searching by most viewed stories may explain this to some extent, there are massive variations from story to story, suggesting that searching by viewership, while it may explain MLD, doesn't explain the others very well - several of the stories on the most viewed page get only about 25 views per day.
The third, as you also noted, is just time; it may be that the site's turnover rate is relatively low, so getting more views is really hard - even a phenomenonally popular story, like My Sediments Exactly, only managed to pick up 8k views, of which it looks like about 80% of which were from its time in the feature box. It may well be that there are only so many eyes you can put on a story in a given time frame, so a story which gets 7k views in the feature box is getting a very significant fraction of the total possible number of views of all stories on the site. So the only way to get enormous numbers of viewers is to be around for an enormous amount of time, and still be visible - which means that some of the older stories, due to their presence on TVTropes and elsewhere, get more clicks.
It is also probably true that chapter stories have exaggerated numbers - Brony Hero of Equestria is only 19k words long, has 62k views, but has like 1.2M views because it has 60+ (very short) chapters. Something like Background Pony, which is more than an order of magnitude longer, has 40 fewer chapters, and gets about half as many new views probably represents more time spent on the part of users, and only has somewhat fewer story views. It may also be that the numbers on many of these stories are so high because when they update again, people more frequently reread the earlier chapters, spiking the total view count numbers, though I'm not sure how likely this is, or how we can tell it apart from the usual readership degredation on chapter fics. The Collected Poems of Maud Pie gets about half the chapter views of My Little Dashie on a daily basis, for instance, because it has 20 chapters and is only 1000 words long; its rate of accumulation of upvotes and story views, however, is obviously much lower.
2159354
It was more of an educated guess based on upvote numbers and my general observations of upvote numbers for featured vs nonfeatured stories - I simply assumed that any story with 300+ upvotes had been featured, which is, I think, mostly a good bet.