We’ve worked with indie game developer Lars Doucet of Level Up Labs to bring his novel Diving Bell prototype to Steam Labs in the form of the Deep Dive experiment, where it now directly leverages Steam APIs to serve up recommendations and related game information. The new experiment offers an exploratory interface to discover new games based on their similarity to familiar ones, plus the ability to use these recommendations themselves as jumping-off points to dive even deeper into what Steam has to offer. 🌊
Deep Dive leverages tags provided by each game’s creators and the Steam Community of users to identify games similar to the ones you know and love. In the interest of helping users explore a breadth of games, Deep Dive displays very similar, somewhat similar, and little-known but well-loved similar games among each set of recommendations it displays. Explore Steam’s popular new releases as a jumping off point, or log in to begin exploration from one of your own recently played games. How many degrees of separation do you find between No Thing and Everything? We count six, which checks out with Bacon’s Law.
Deep Dive’s variation from the original prototype is based on some things we learned by playtesting during development. For example, tags the base game shares in common with each recommendation are more prominently displayed to help inform both what the game is about and why it was recommended.
We’ve tried to strike the right balance between providing relevant details without overwhelming users in the process of browsing and we hope you’ll let us know what you think. We can’t help but add a few items to our own wishlists each time we test Deep Dive. (Well hello, Circa Infinity!) We suspect this is a good sign of its potential, and we’d love your thoughts on whether Deep Dive offers a compelling alternative to other forms of content discovery.
Experiment 006: Community Reviews
Your Steam Friends list and activity feeds are great ways to see what games your buddies are currently enjoying, but what about the rest of the Steam Community? Top sellers are one way to keep your finger on the pulse, but positive reviews can be an even stronger signal of what the Community is enjoying on Steam. With this in mind, our Community Reviews experiment surfaces the games people are actively recommending to one another on Steam.
Update - Experiment 002: The Interactive Recommender
The Interactive Recommender doesn’t need you to interact with it before proving itself useful, so we’ve given it a chance to earn its keep right on the Steam Store home page. Simply log into the Steam store and you’ll see games our machine learning server has lovingly featured just for you based on your recent gameplay. Plus, click through to Steam Labs where you can explore titles new or old, popular or niche, with the machine’s bespoke take on your tastes, informed by thousands of other players... a lot like you.
As always, we hope you’ll check out these latest and upcoming additions to Steam Labs and let us know what you think in the discussions. Your feedback shapes our experimentation and informs the ideas which become a part Steam for keeps.