The Steam Labs are still relatively recent and have been running for the past couple of months, coming out around June of 2019. Along with the slew of other improvements that Valve have successfully pushed this year, they exist to promote discoverability beyond what genres people are already used to. Also, they can help find hidden gems in preferred genres that may have been overlooked at the time of release due to other, shinier titles overshadowing them. \n \nThere are six Active Experiments being run at this time, and we take a look at all of them. \n \nhttps:\/\/twitter.com\/Steam\/status\/1208105148448247808 \nMicrotrailers \nThe phrase 'don't judge a book by its cover' rings a bit more hollow for games; title art can make or break a game before the game can even boot up, and it should do a decent job of conveying what users should expect. Videos are an even stronger tool for developers and their audiences alike, as it allows the devs to display their games core gameplay loop, and consumers to temper their expectations of the title. \n \nThe Microtrailers experiment brings this full blast, with six-second trailers being shown for a slew of games. Not every title is currently supporting for this, and it feels like a bit of an ADHD dream with the flashing videos that loop every six seconds, but it's an interesting way to see titles at a glance without loading their store page, browsing for three minutes, and then finding a new title to explore. There's more than one way to explore the Microtrailers, as well. Steam currently offers the ability to browse titles by individually hovering over them, play entire rows of microtrailers at once, or browse based on curators. \nInteractive Recommender \nAnyone with money to burn in their Steam wallets has likely ended up searching through Steam's overwhelming amount of titles for sale. You typically select what you're in the mood for, sort by popularity, and start exploring. The Interactive Recommender checks what games you've played recently, for extended periods of time, and then matches you to other games based on what you may also like. By reading title tags, and offering tag filters and exclusions, along with popularity niches and age, you can do some sizeable searching for gems that you would have otherwise missed. \n \nIf you're really feeling froggy, bring that popularity slider all the way to niche and start searching some gems that are unheard of. \nAutomatic Show \nThe idea behind the Automatic Show is that Steam will autonomously build a half-hour show giving viewers a peek of trailers of current games that are currently popular, trending, positively reviewed, or segregated by genre. By allowing it to be built autonomously, it's easier to iterate and deploy as it won't require active developers working on the show when they could otherwise be adding hats to Team Fortress 2. \n \nLike the Microtrailers experiment, there are multiple builds for you to explore. The primary show is a 25-minute long bot-generated show that brings a huge wealth of games front and center in brief trailers. Like the Microtrailers idea, but back-to-back. Each title is briefly shown for a few seconds, and then it's on to the next one. Keep your mouse over the 'Now Playing' icon, and click when you see something interesting; it'll pop up in a new window, allowing you to keep watching the original window. It's almost unhealthy for financially savvy consumers; be wary of your wallet in this mode, because you'll end up with a few gems. \n \nThere's also a Top Releases for June 2019 video, and a three minute VR show. Finally, there's a Rapid-Fire Horror video that plays horror games in three seconds, half the Microtrailer time. It's frankly just slightly too fast for you to pull titles from unless you give it a second watch, or are constantly pausing it. \n \nAs stated, this is only really viable if it can be done entirely autonomous. For a first push, it already looks great; here's hoping it's added to Steam in lieu of daily queues. \nSearch \nSteam already has a Search feature so this may fly under your radar at first glance. It beefs up the Search experience by selecting the price you're looking for, and you can narrow it by the tags and by electing to keep items already ignored to not appear in the search query. It's reasonable, functional, and quick, even if the glitz and glamor are left to the other experiments. \nDeep Dive \nIn the Deep Dive experiment, users select one title that they enjoy, and are greeted by a slew of titles that mirror the tags of the title that is used to query the search. \n \nMultiple titles can be aggregated as a search query to further define what you're looking for. The inclusion of Microtrailers can give you a quick at-a-glance of the titles that the Deep Dive system comes up with, and the query can be refreshed for a new list. It can also be hit and miss; querying the Witcher 3 ended up giving me a handful of bizarre visual-novels that I've never shown interest in before. Of all of the experiments, this is the one that was most interesting, so it's a shame the functionality can be spotty. \nCommunity Recommendations \nThe final experiment that is active for the moment is Community Recommendations; the system finds reviews that have been marked helpful, and displays the view along with the title the review is for. It's a backward approach to finding games, but an interesting one. Typically users find an interesting game, then seek reviews to figure out what they may be getting into, for better or worse. This brings users out of their comfort zones and typical genre-stomping grounds, and the results are both surprising and refreshing. \n \nYou can further define what is shown to you, to select precisely what you're seeing in this game mode. From reviewer playtime contingencies that you can set, to date written, it brings a scrollable list of games that you'd otherwise likely bypass. \n \nOf all six experiments, the Community Recommendations and Automatic Show (coupled with Microtrailers) brought me to the vaguest of doorsteps on the Steam platform, areas that haven't been fully celebrated by the masses yet, and resulting in purchases. That's arguably a great thing for developers and users alike. Here's hoping more experiments are inbound.