Just wanna add that I don't think adding a bunch of fetishes is the answer anymore. Due to changes with the latest page, one can now block specific tags. Not in the mood for incest? Block it! Now, all incest games are hidden. Simply adding tags for view-ability no longer works.
That said, the single most powerful weapon under anyone's disposal is without a doubt repetitiveness. The more often someone notices a game, the more likely they are to try out the game. It doesn't matter what the game looks like, or many other factors. If I see a game enough times on a feed, I WILL click on it. If I click on it, there is a 100% larger chance that I'll download and play it when compared to not even bothering to click it.
Many games have actually used this strategy to their advantage, rather they meant to or not. I'll use
Twisted World as an example. This is not a game I'd normally play under most circumstances. It contains tags I generally don't enjoy, the banner is the ultimate balloon tit character I generally dislike, and yet it kept popping up while I was searching for new games. Eventually, I just said "fuck it" and gave it a chance. I actually like it, despite my initial skepticism and general dislike of some of its tags.
Another example would be
Grandmas House, while it does contain tags I generally enjoy, the banner doesn't do it much justice. Furthermore, it's name is a massive turn off for me. I have 0 interest in GMilfs, and this name screams "its all Gmilfs!" however, due to it's update schedule being insane, it was impossible to ignore. I simply had to click on it and now it's become one of my favorite games.
While TW kept appearing because of it's growing popularity at the time, and GH appears not only due to popularity, but its insane update schedule, they both show the power of visibility and repetitiveness. While gaining the cult-like following of TW or the update schedule of GH may be difficult, this may be a line of thought worth considering. How can one manipulate the algorithm to show your game more often?
While I may not have the answer, I do believe this to be the most powerful potential tool any developer can utilize, if they can figure out how.