Hits and AO3 feedback
In the last four months, I've started posting written works on AO3 in a new fandom and have been tracking hits and other feedback over time to better understand reader behaviour. The following is a comparison of engagement over weeks and months on fics with contrasting properties:
Gen vs R rated short fic
Fest vs non-fest release
Serialisation effects
In terms of work quality, these are fine, not stellar or terrible, so the numbers are more a reflection of interest from the reader pool. Hits show how many want to take a chance on these tags instead of author recognition or strong recommendations. But while the exact numbers won't match other fandoms, the trends should scale.
Hits over time: New one-shots
New works and updated works get a big boost of visibility on the AO3 platform because the default display for searches is reverse-chronological. Lots of potential readers are looking at what was put up recently. Correspondingly, the rate of hits naturally goes down with time, with fewer new eyes seeing the work as it slips out of the top few pages within a fandom tag or any more specific searches performed.
I'm using a power series factor to model the accumulation of hits because it's one of the options on google spreadsheets, but also it does a pretty good job. I think its probably better to model hits over time with two terms, one for the initial burst of hits on new works and a second for the attractiveness of the tags that would better estimate the long tail, but programming that will have to wait.
The two oldest works, the G rated and R rated rarepair fics, both have slowed to 0-3 hits per day, and in a year, they should both be well under a hit a day. The gap between them from the initial rush of interest at a new R rated work will continue on unless disrupted by external forces.
Although hits slow down as a work ages, the feedback rates per hit doesn't necessarily change that much. The G rated fic accumulated Kudos and Comments faster (per hit) in the first two weeks of being released, but after that the rates stays around 0.1 and 0.01 respectively. The R rated fic shows persistently lower Kudos rates, but seem to stay around 0.05 Kudos per Hit months after being posted, and comments continue to creep in. Both fic show fewer bookmarks after 2 months, perhaps a reflection of saturation: those who want to keep track of these kinds of works have already found it. Mean while more casual readers are coming across them whenever they dig deep enough.
Hits over time: Fest reveal vs Recent Work
The purple work above, R VHope (8.5k) was initially submitted to a fic fest. Fests are fanfic writer events that prompt and promote new works within some theme, in this case A/B/O fic featuring JHope. Fests also help foster community between fic writers with similar interest and readers looking for new niche content.
Fest fics are published on AO3 for submission but remain hidden until they are revealed as part of the fest. As such, by default, newly revealed fest fics can be found within the fest collection however they are often too old (weeks old) to show up amongst the most recent works. So that means their initial audience is only the Fest readership.
After my fic had been revealed in the fest for a little over a week, I decided to change it's structure from 1 to 2 chapters without modifying the text at the same time as adding a few tags. This change resulted in the fic suddenly showing up in the Recently Updated Works for standard searches of AO3. This natural experiment makes it possible to see how the fic faired under fest and open conditions.
The difference in hits between the initial days as a fest fic and the hits since being visible as a recent work is pretty substantial, with almost twice as many hits accumulated after being amongst the recently updated (232) than only within the fest (133).
Of course, more hits doesn't always mean more rich feedback for fic authors. The feedback rates differed in different directions between the conditions. The Kudos rate per hit was the same as a fest reveal and as a recent work, around 0.1, while comments were more common during the fest reveal conditions (0.023 vs 0.013) as were bookmarks (0.05 vs 0.03).
For this fic, the benefit of releasing as part of a fest collection is not so clear, but I should add that this work turned out to be a poor fit for the collection:
It was the only work released with this particular rareship in this round of reveals
it was at the bottom end for total hits in the collection before the update (bottom 3 of 18)
It received the least engagement from other participating authors
When a fic better fits the fest readership's interest, the increased comment rate is a substantial boost in engagement relative to posting just as a recent work. But they have to be willing to try reading it at all for that to have an effect. So now I know to lower expectations when posting against the grain.
Serialised Fic: WIP vs Complete
Over 5 weeks, I posted a 9 chapter fic with a total length of 58.6k words. Updates were initially twice weekly, and I noted statistics after the first 24hrs and at the next update as a work in progress. Stats have been recorded daily after the work was completed.
The biggest difference is in engagement per update is after the work was completed. While some of that boost was from subscribed users and others anticipating the completion of the work before reading, I think we can also claim that incomplete works are avoided by a lot of readers. There is a WIP penalty for hits received as a recently updated work.
Just looking at the hits gained per chapter while a WIP, we can generalise around some natural noise to a rate of engagement for the work itself. The work eventually had 44 subscribers but that later gain didn't produce consistently more hits for every chapter. Instead we can see between 54-97 hits in the first 24 hrs (avg 83 hits) and then decreasingly many over time. Note: the clustering isn't tied to day of week, more likely it is differences in social media mentions.
But as has been described above, not all hits are created equal.
While the fic was a work in progress, the only form of feedback that difference substantially in the first 24hrs from subsequent days is the guest kudo rate. Some AO3 users log out to leave guest kudos when they want to show appreciation for a work they have already kudoed, and I would guess this boost is from a few of those following the WIP instead of people without accounts diving in repeatedly after each update.
The first 24 hrs of the fic being complete also shows a surprising story, with a lower Kudos rate and a surge in bookmarks, public and private. After this surge, Kudos and comment rates per hit have settled back to values similar to the rate on the WIP. Presummably they will also drift down a bit like the trends on the older One shot fics. If they do something else, I'll have to come back to explore it.
In a few months I might be back anyway with estimates on how many people have actually read these works, because the hits to feedback ratios leave a lot of room on all of these works.
Sorry for writing so much and if you found this interesting or different from you experience, I'd be happy to hear about it!
