I considered the lying problem.
I don't think users have any incentive to lie.
If Tapas made it possible to filter any tab/search result to 'fan works only', and replace 'genre' with 'fandom' when publishing fan works, it would make fan works significantly easier to find. It would be within the interests of fan creators to tag their works properly, as then the people looking for them could find them much more easily.
As such, it would not be unreasonable to make incorrectly labelling fan content against Tapas' Ts+Cs.
It would also be easier for copyright holders to locate fan works of their IPs and request their removal. By 'burying' them as it does presently, Tapas succeeds in both doing a disservice to fan creators by displaying their work in places it isn't entirely welcome, and looking very shifty to IP holders.
I have coded some algorithms in my time, and I have developed apps for almost two years now. But I'd still consider myself mostly a novice.
By my knowledge an Undiscovered tab could be hard-coded, unlike a popular or trending tab, which is likely to make use of some machine learning. I'm not sure if Tapas' popularity algorithms rely on AI - they don't particularly need to since relatively speaking there isn't that much content compared with say, Reddit, Instagram, YouTube, tumblr etc. But, most algorithms of that sort do, making them all massive resource-hogs. An 'Undiscovered' algorithm would not need that.
An 'Undiscovered' tab would rely purely on metrics that would all be computationally very simple to calculate like (for example) the rate of change of the average length of time between updates, or how quickly a creator replies to comments, the average length of those replies etc. - data Tapas already has, smooshed together with some elementary maths.
Things like likes and subs all write into the database in as close to real time as possible. When I click like and reload the page, the like is there. The metrics that made up a consistency score wouldn't be like this, because:
- It would be possible to calculate some of the score/estimate the score in advance when a user schedules updates in advance. This would make it possible to predict the Undiscovered tab in advance. As one moderator to another... being able to predict problems ahead of time... that's the dream, right?
- It would change extremely slowly on account of relying on the trends in the changes of averages over the course of many weeks, meaning Tapas' servers could run many of the calculations as a weekly background job during their least active hours, as opposed to running them in as close to real time as possible like how they handle uploads and payments etc.
- It would only be necessary to calculate the score for series that are 'active' and have been active for a predetermined length of time, that fall within the least popular quartile of series, meaning that only a very small number of series would need have a full consistency score calculated at all each week. In fact, this score need only be calculated for some 100-200 series each week.
Overall we're talking a scheduled job in the back end consisting of a few thousand good, old fashioned database queries and some elementary mathematics. To design the new creator dashboard will have taken orders of magnitude more developer time. The servers cope with more in a couple of minutes than calculating this metric would take in a month.
The biggest challenges would come in fine-tuning it, which would take time, but then that's true for every new feature in a piece of code.
Championing indie comics is in Tapas' brand image (even if some folks in this thread feel y'all haven't lived up to that), so Undiscovered would look very good indeed from that perspective, and no other platform offers it, so even better. A decent developer could get a version of this feature up and running in a week - a reasonable price to pay considering everything it could offer.
What could it offer?
I've already hinted at the consistency score's ability to be predicted in advance because of Tapas' episode scheduling functionality. Additionally, it would directly cause attention to go towards good things without much attention.The 'undiscovered' tab then would surely draw a certain allure to it.
In the UK in the 80s and 90s there was a radio presenter called John Peel, who was notorious for playing unsolicited submissions. He became so influential that record labels offered deals to artists who were played by him. As such, listening to the 'Peel Sessions' became a tool for predicting who the music industry was about to discover, and so became extremely fashionable.
... I just want to put it out there that maybe having not only the only Undiscovered tab in the business but the episode release schedules that allow you to predict what it will come up with next might be very useful indeed.