Itch.io staff replied

We managed to finally get a response from Itch.io directly. Essentially, it was confirmed to us that this change to the algorithm was not a bug, and that their intention was to break up what they see as a group of games that are "popular for being popular." According to Itch, they see views and popularity as zero-sum. For games to be seen, other games must necessarily not be seen--the "free views" (their words not mine) they were giving to games that were doing well are now being given to games that are, apparently, more deserving.

According to the staff member who spoke to us, high quality games with well-designed pages, where the devs actively promote their work, will now be rewarded. The implication of this seems to be that games like mine that are dropping off are not doing these things. I'm a little salty about this if you can't tell, but if the idea is "let's give games that aren't as popular a chance," then that is good. But to do that the pain isn't being spread out evenly, and instead of new initiatives to help beginners and smaller devs, it has been determined that other devs must be unseated to make way for the less fortunate.

Ultimately what's happened is the permanent top X games of any given tag will remain as untouchable as ever, but the second stringer or third stringer games like mine that are popular in a few tags but generally middling in broader categories, are being cannibalized by less popular games to facilitate those games' success. There also seems to be a belief underpinning this change that these less popular games aren't less popular for any other reason than something arbitrary like luck, which is probably true in some, but definitely not all, cases.

Likewise, there is a presupposition that games outside the very tip top of the top are actually undeserving of whatever success they've garnered--that it's some sort of snowball effect, what the staff calls "popular for being popular" and attributes to their own algorithm granting these games "free" exposure. It's not that games like mine steadily climbed the list on the back of quality updates and constant pavement pounding, but that Itch.io arbitrarily granted us a stroke of success that fed into further, even less deserved success.

So, are thing really zero sum? There is an economy of attention in a sense, certainly. People can only play so many games. Is the success of middling games undeserved? Success is almost always built on some part of luck, certainly. Should games that fail to break through be given more chances to succeed? There are many games that are amazing that never see recognition, of course. But my opinion is that all of this is misguided.

For one, I think a better way to surface new games is to feature them. If it truly is a zero sum game, having a list of featured games appear prominently, curated to be the types of games Itch wants to surface, spreads the pain evenly instead of specifically cannibalizing the second and third stringer games. Itch does feature games already: I'm saying to expand and tailor this system to the stated goals. Second, I think rather than making the popular game ranking more volatile, the "new and popular" ranking could be made less volatile and be set as the default sorting method when browsing itch. New and popular is already intended to be a sort of "trending" sorting method, where games are surfaced according to popularity but with a much stronger bias towards recency. Making the Popular sorting method have this same bias, but somewhere between where it was before and where "new and popular" is now, seems like a worse move than moving and prioritizing the latter.

Finally, one of the main things Itch has changed, if I am reading between the lines correctly, is that external traffic is now weighed much more heavily than internal traffic. This is based on the idea that external traffic "proves" devs are promoting their game, whereas internal traffic from within Itch must be proof of the popularity snowball effect they want to avoid. All this does is cement a permanent upper class of games. The most popular games on Itch are now games that haven't been updated in two years like GGGB because there's a lot of chatter and external traffic driving toward that game. Hell, Depression Quest has resurfaced as a top popular game, because there's like fifteen million articles on the Internet about it. On top of that, this ensures that the zero sum game comes only to cannibalize one group of devs. If you have the money to advertise, or the cultural penetration to have tons of articles written about your game, now Itch will reward your insurmountable external popularity with increased internal popularity. Why, it's almost like a snowball effect!

My intention is not to criticize Itch for wanting to change the algorithm, and any change they made that hurts my games would have made me upset on some level. But this change is made specifically at the expense of a certain group of people (though unintentionally) at a certain level of popularity, and that it was made based on a series of assumptions that I don't think are fair or even true. On top of that, I do not believe these changes will be good, or help surface better games; instead, I think they'll make the algorithm less useful. Of course, I don't really know how the algorithm works, I can only base my analysis on the outcomes I see and on reading between the lines of the response we've received from on high, so I don't speak with certainty about any of this. But to me, this is an uncharacteristic mistake on Itch's part, and it seems to be borne of staring at spreadsheets instead of analyzing the actual games it will effect.

Regardless of the state of Itch's algorithm, or my place in it, it remains a good service that helps indie devs across the world and treats it's devs like rock stars. I truly love Itch, and believe the service it provides is invaluable. So no matter what, Itch.io will never have my scorn. Still, this is a disappointing turn, and I hope it doesn't make Itch a less vibrant, less useful, or less fun place.