Hi everyone,
With itch.io (and my own personal account in it) about to celebrate its 9 year anniversary in March, I wanted to voice some tangents and thoughts regarding the site’s design and approach. These thoughts eventually emerged as an essay I wrote about the site over the past two days or so.
Gotta warn you - it’s overlong, has some weird analogies and mediocre attempts at dry humor, but hopefully some of it will resonate and reach the right people.
PASSWORD: goodbadugly
Thanks for writing that, but:
I fear you're describing the itch.io you'd build if it was yours. Or rather the one you think you want. Are you really looking at the real one?
I understand the rationale behind reviews, in the sense that they should not completely govern how the public sees a game, especially in this era of review-bombing and public brigading. That’s fine and acceptable. But you can’t have it both ways; you can’t have a rating system and then hide half of it. You can’t have the site ask people to rate games when they don’t understand the impact of their rating. Hell, even you agree that they should be redesigned, too. So we both agree there.
The search system is definitely not the worst, but in a world where auto-complete and suggestion repos exist, “this is how it has always been” is not a good counter-argument. You can have a full-title search system while also accepting semi-complete titles - Bandcamp is a good example.
Now, the algorithm. The fact that it’s obfuscated to prevent exploitation is not the problem - it’s something I expect and respect. I also fully understand the problems with balancing such a system, and I outlined them in the review. My point is that the system has already been unintentionally “gamed” by a positive reinforcement loop, and it should be fixed or at least tampered. One can keep a trade secret while trying to fix it.
Thank you very much for reading through, though I think I might have been misunderstood. The main thesis of the article is that itch is great for devs and kinda awkward for anyone else to use, as it was built that way from day one. Nowadays it is in a weird limbo where it is neither a standardized storefront nor a Newgrounds-esque community space. This base assumption fuels every single criticism. If you disagree with that assumption - that’s perfectly fine! We’ll have to agree to disagree about the issues and the way to solve them. I love itch, otherwise I wouldn’t use it for the better part of nine years.
Now, the algorithm. The fact that it’s obfuscated to prevent exploitation is not the problem - it’s something I expect and respect. I also fully understand the problems with balancing such a system, and I outlined them in the review. My point is that the system has already been unintentionally “gamed” by a positive reinforcement loop, and it should be fixed or at least tampered. One can keep a trade secret while trying to fix it.
Is that really a problem that Itch can solve, though? Exposure begets exposure, and popular games are really all we can expect to see on the "popular" tab. Most of us will never make the kinds of games that attract the fickle attention of major platforms, regardless of how Itch organizes its content. What would you do differently?
I agree it’s a complex issue, and maybe I bit off a bit more than I could chew with that segment - I will accept that.
To answer your question however: assuming I had the knowhow and the means, I think I would take a note from Steam’s playbook. The tag system is already a well-established and well-known system here on itch, and the front page already uses it to give users personalized recommendations. I’d expand the tagging options and make them more robust, and sharpen the personalization methods.
I wouldn’t get rid of the “most popular” page so much as I would make the “games for you” page a default option. Give the people more of what they love alongside some products mildly outside of the users’ niche to prevent an echo chamber and keep people open for more things. If that user follows other creators we can factor in their reviews in as well.
I’m not vain or pretentious enough to claim this is what itch should do. This is obviously not a perfect solution, and it has its own share of problems. For one, it leads to “tag wars” as everyone tries to mix and match the various metadata elements to get the most page views and clicks and hit that discovery sweet spot. Personalization algorithms also have a bad habit of digging people deeper and deeper into their per-established “rabbit holes”. But itch is not as competitive as Steam, not by a long shot, and it may (or may not) solve the issue of perceived monotony and sameness.
Adding the "recommended for you" tab to the browse page is probably a good idea. I tend to forget that it's there, since it is buried halfway down the front page, and because on the front page it always shows the same few projects that I don't want to see but can't get rid of. I don't know how well it theoretically works, though, because it is based on ratings, and there are several reasons why I intentionally don't rate projects in many cases. Looking at my current recommendations, 12 of the 40 have fewer than 10 ratings, so at least it's not all games that are already widely popular. None that have ratings have less than a 4-star average, though. I expect that games with few ratings are easily tanked by one or two downratings, both here and elsewhere.
I have to wonder if a lot of people never browse beyond the "popular" tab, for whatever reason. People often complain about an overwhelming glut of horror games on Itch, but I hardly ever see them, and that is probably because I don't look at the popular section.
Good stuff. I also don’t think itch should adapt a full “superindie Steam” push, but that ties into its internal identity crisis where it’s unclear if it’s Newgrounds, Bandcamp, or, well, Steam.
As you point out with community profiles, itch’s UI/UX is generally a mixed bag. Took me a while to understand that people use collections as a sort of backlog, too. They actually have all sort of uses: as wishlists, as a way of featuring games you liked, as a way to organize certain things you want to play on stream, as a way to sort out your “library”, and even as a way to organize and design your own creator profile. It’s a versatile system and I understand itch’s hesitation to put in a “proper” wishlist collection when collections would suffice, but it does harm the site’s “flow”, if it makes any sense. It’s like the pieces are there but they’re all just slightly off.
