Recommendation Systems Need To Change
I may not be good at basic things like driving or cooking, but I do know one or two things when it comes to video games. I know my ARPGs from my JRPGS, Dota 2 from League of Legends and what it means to be molyneuxian. Yet I mostly acquired this familiarity over the last few years of my life, a relatively short span of time compared to my entire gaming career. Most of my game-infused adolescence was spent brainlessly plowing through dull rounds of Team Fortress 2 and grinding in World of Warcraft – awareness of what was happening in the rest of the blooming gaming industry would come way later. I actually feel pretty fortunate for not having turned away from the medium completely after having realized how much time I “wasted” doing things that were only marginally less boring than real life, I know enough people who have.
Likewise, I believe that a large part of the reason why people who’ve never touched a controller recoil from anything related to gaming is the relative obscurity of the non-AAA world. The games that have the largest marketing budgets and therefore the ability to reach the most people tend to be pretty bombastic, they tend to be “gamers’ games”, the Uncharteds and The Witchers, and I can see why someone completely new to the medium can be turned away after getting the impression that that’s all there is. These two problems, of keeping people immersed in the medium despite burnout and getting the uninitiated interested can both be solved, I believe, by shaking up the seemingly abandoned recommendation systems embedded in digital distribution platforms.
Steam recommends its wares using a simple tag system. Every game gets a bunch of key terms associated with it like “story-rich” or “roguelike” and comparing how alike two games are boils down to seeing how many terms they share. The algorithm might be improved by taking playtime and wishlists into account, but it basically means getting a bunch of games similar to game “x” you’ve played. This is fine in principle, but when was the last time you found your next gaming obsession based on an algorithm instead of after reading about it on some gaming website? As far as I can tell, no one really uses it, at least not on Steam, where the recommendations section is hidden at the very bottom of the marketplace, a desolation of dull rotating lists no one ever uses.
There’s a good reason for the dusty, abandoned existence of game-recommender systems compared with the vibrancy and the buzz around their equivalents in the video and music industries. Services like Netflix rely on subscription fees and therefore need to keep people hooked for as long as possible. This involves creating a culture around their content, exposing people not only to shows and movies similar to what they had watched before, but constantly trying to get them into new, unexplored areas, things they might never have thought they would spend their next week binge-watching. Steam relies on sales for revenue, which means featuring the newest and freshest games, highlighting massive sales that provoke spending-sprees and catering to AAA powerhouses with sizeable marketing budgets. Promoting the whole culture is only financially beneficial to Steam in the long term, selling more $60 shoot-em-up’s profits them right away. The way I see it, the ideal game recommendations system should be an educational device. If you’re already familiar with the medium enough to follow the discussions on major news outlets, Reddit threads and Youtube channels then you should have no real problem of finding exciting games to play. For people new to gaming, finding games to play must be pretty overwhelming. We talk a lot about how the medium is getting more and more accessible to non-gamers, with titles like That Dragon, Cancer and Firewatch being a few of the latest I could completely see both my grandparents and too-grown-up-for-games peers enjoying. Their existence, however, continues to be shrouded, with the broader public only getting exposed to titles that can afford littering movie screenings with trailers.
A keen sense of the progression from curious non-gamer to gamer should be a key design concept for recommendation systems. They are meant to help you find the game you will enjoy the most, but recommending Heroes of the Storm to a Dota 2 player or Overwatch to a Team Fortress 2 player shouldn’t really be the aim. Taking you from Gone Home to Civilization 6 should be. Tutorials for the game of gaming, not intricate networks cataloguing similarities. Of course this invites all kinds of intertwined hard questions about how one would go about building a definitive tutorial for gaming, which games would make it into the canon and how personalized recommendations would be dealt out with regards to the different backgrounds of people using platforms like Steam. However, this, I believe, is a worthwhile discussion to be had as it could not only get way more people versed in the language of video games, but would also bolster interest in small, underfunded indies, which usually require less familiarity with game mechanics to pick up than AAA titles.
I know how promoting games on platforms like Itch.io sounds like it would not be in the immediate financial interest of Steam, but growing the gaming community by exposing people to the whole spectrum of games can’t hurt them too much. Is this me just being salty and secretly lobbying for a system which would recommend Chrono Trigger to everyone? Maybe, but I think that would be at least marginally better than what we have right now.