Actions

Work Header

Rating:
Archive Warning:
Fandom:
Additional Tags:
Language:
English
Series:
Part 7 of Fandom Stats , Part 1 of Popularity Metrics
Stats:
Published:
2013-05-18
Words:
786
Chapters:
1/1
Comments:
7
Kudos:
86
Bookmarks:
2
Hits:
2,200

[Fandom stats] Do longer fanworks get more hits on AO3?

Summary:

How do AO3 hits relate to number of chapters or length of fic? (As of 2013)

Notes:

Originally posted on Tumblr.

I've since learned that AO3 does make efforts to de-duplicate hits coming from the same person to the same work. Which wouldn't be 100% effective, but does complicate a bunch of my assumptions about how posting a new chapter might work. Overall, a bunch of this post is dubious (both in terms of methods and inferences), but I'm archiving it anyway for completeness/reference.

(See the end of the work for more notes.)

Work Text:

So, I started geeking out with fanfic stats again, and made some more charts.  Today I was interested in answering the following question:

Do multi-chapter fanfics get more views than single-chapter fanfics on AO3?

(Spoiler: Not exactly.)

Introduction: 

I started thinking about this because every time I posted a new chapter of May Your Heart Purr Like A Bumblebee (my only multi-chapter fic), I noticed a substantial boost in number of hits that the story got.  It occurred to me that perhaps every time an author posts a new chapter, a story is re-exposed to roughly the same number of new viewers (whoever happens to be browsing recent AO3 posts, e.g.).

Methodology:

I decided to compare the average number of hits per work on AO3 between single-chapter and multi-chapter groups.  AO3 doesn't let you search for specific numbers of chapters, but if my hypothesis were correct, I would see a higher number of hits for multi-chapter works than single-chapter works.  If my more specific hypothesis were correct -- that an author gets about an equal surge in readership every time they post a new chapter -- then I should expect to see at least twice as many hits per multi-chapter work as per single-chapter work (because each multi-chapter work has at least two chapters).

I decided to compare median number of hits rather than mean, for a couple reasons.  Median is less susceptible to outliers.  But, more importantly, Median is much easier to determine on AO3!  Just sort the results by number of hits, then go to the the midpoint of the list and see how many hits that work has.

Big caveat:  I was able to get median number of hits for single-chapter works, or for all works.  However, I wasn't able to select only for multi-chapter works.  So every time I refer to the "median" for multi-chapter works, it's actually a misnomer -- I instead assumed that the median number of works overall was a weighted average (weighted by number of works in the category, shown in Fig. 1) of the single-chapter median and multi-chapter median.  This is highly dodgy statistics -- it does not actually give the correct multi-chapter median.  However, it's probably close enough to accurate for these purposes, where I'm mostly interested in relative size and orders of magnitude.

(NB: I'm not sure why Fig. 1 turned out several percentage points different than when I ran the same search a couple weeks ago -- it seems unlikely the ratio would have changed substantially in that time.  So that may cast some doubt on the how much we can trust these figures are overall -- perhaps AO3 doesn't return consistent sets of results?)

Edit to add: I got a useful further caveat from one of the AO3 wranglers:

[M]any of those "one chapter fics" aren't fics at all. They're art, or vids, or podfics, or cosplay, or meta, or whatever. It'll be easier to draw conclusions about fic specifically once there's a filtering option for that.

Results and discussion:

Examining Fig. 2, we see that multi-chapter works do, on average, get more hits than single chapter works (though not by a factor of 2 or more, as I'd hypothesized).  However, I decided to dig deeper.

If we control for the length of the work, as in Fig. 3 (I created word count buckets that are admittedly arbitrary -- they correspond to what I personally think of as similar-sized fics), then we see that the story is not so straightforward.  The main trend is that longer fanfics get more hits.  There are only very small differences between the median number of hits for single-chapter vs. multi-chapter fics, generally, and the direction of the difference is not consistent.  And, for very long fics, single-chapter fics actually get substantially more hits than multi-chapter fics (there just aren't very many of them).  As shown in Fig. 4, most single-chapter fics are much shorter than most multi-chapter fics (unsurprisingly), which is driving the effect in Fig. 2.

Conclusions

  1. Longer works get more hits.
  2. There are more short fics with one chapter, and more long fics with multiple chapters.
  3. Controlling for length, there is not much difference in number of hits for single-chapter vs. multi-chapter fics.
  4. However, for very long fanfic, single-chapter fics have a higher median number of hits.
  5. I have too much time on my hands.

Edit/Future work:  I'm not fully satisfied with this analysis; I believe that many stories do indeed get a bump every time a chapter is posted.  It's possible that many multichapter fics are posted all at once, and that this makes such an effect hard to see -- or that something similar is going on.  I may do further investigation into the matter in the future.

Notes:

Comments welcome, but I’m in the middle of a massive fandom stats backup due to Tumblr purge, so I may be slow to respond.

Series this work belongs to: