Who Does Steam Show Your Game To? Choosing and Ordering Tags in 2026, the Visibility Signals Valve Has Confirmed, and How to Audit Discovery in Steamworks Traffic Reports

Steam decides who sees your game in two stages. First, tags decide which players and which other games your game is placed next to. Then, whether people actually buy and play it decides how much more Steam shows it. In its Steamworks documentation, Valve names the strongest signal as "Are people buying and playing your game?", and states plainly that store page traffic and the conversion rate from visit to purchase are not visibility factors. That is why choosing tags well and driving what happens after a visit cannot be treated as separate jobs.

Many teams still set their tags once while building the store page and never revisit them, and treat the Discovery Queue as something that simply happens at launch. In practice, the Discovery Queue accounts for close to half of page visits in the first two weeks after launch for many games, and getting into it before launch requires a burst of visits and wishlists in a short window. Drawing on Valve's Steamworks documentation and published developer data as of November 2026, this guide covers what each Steam visibility surface responds to, how to choose and order your 20 tags, how games get into the Discovery Queue, and how to audit surface-level performance in Steamworks traffic reports.

How well your page converts visits into wishlists is covered in How to Turn Steam Store Page Visits Into Wishlists, and how capsule art turns impressions into clicks is covered in What Steam Capsule Art Has to Do. This article covers the step before both: where impressions come from in the first place.

Which Steam surfaces show your game, and what does each one look at?

Steam's visibility surfaces fall roughly into two groups: surfaces that classify and match games by tags, and surfaces that rank games by sales and play activity. Most of them are personalized per user. As Valve puts it, you may see a certain game on a surface while a friend of yours won't.

Based on Valve's documentation (Visibility on Steam, Steam Tags, Wishlists) and published analyses, the main surfaces and their inputs look like this.

As the table shows, before launch the inputs a developer can directly touch are essentially tags, launch timing, and wishlists. After launch, sales and play drive most surfaces. Think of tags as the ticket into a particular pool of users on a particular surface. How far the game rises inside that pool is decided by what happens after the visit.

As of November 2026, Valve's events page lists themed fests including Auto-Battler RPG Fest (November 16–23, 2026), Desktop Companion Fest (January 14–18, 2027), Couch Co-Op Fest (February 8–15, 2027), and Real-Time Strategy Fest (May 10–17, 2027). Valve describes eligibility for each fest by its thematic category and does not list tags as a criterion, so the decision to apply should rest on whether your game genuinely fits the theme, not on your tags.

What has Valve said it does and does not use as a visibility signal?

The strongest signal Valve names in its documentation is sales and play activity. Valve also states that store page traffic, the visit-to-purchase conversion rate, review scores above a certain floor, and paid placement are not algorithmic visibility factors. Knowing this split tells you where to spend budget and time.

The Wishlists page and the Visibility page phrase things slightly differently, and the difference is worth noting. The Wishlists page says wishlists "can be an important factor in determining where your game appears on Steam" and that, before launch, a game may appear on lists of upcoming games according to filters such as popularity. The Visibility page says wishlists are not an algorithmic factor except for Popular Upcoming. Read together, wishlists count toward ranking on upcoming lists before launch, and after launch their role is less about ranking and more about driving first purchases through notifications. Discount notifications require a discount of at least 20% on the lowest-priced package, running for more than eight hours, and are limited to one wishlist email per app every two weeks.

The most commonly misunderstood row is traffic. Valve states that neither the traffic landing on your store page nor visitors' purchase conversion rate is a visibility factor. External ads that raise visits grow your visibility inside Steam only when those visits turn into purchases and play. For the same budget, channels should be judged by whether the people they send buy and play, not by how many visits they generate.

How should you choose your 20 Steam tags, and in what order?

The baseline is to choose 20 tags that describe your game accurately, and as specifically as possible, with the five tags that best describe it at the top. According to Valve, a title needs at least five tags before launch, developers are encouraged to use up to 20, and while many tags can be applied, only the top 20 are shown and used for visibility.

The tag rules in Valve's documentation can be summarized as follows.

In practice, the most efficient way to choose tags is to start from comparable games. Pick 5 to 10 games whose players you hope to share, collect their top tags, and keep only the ones that genuinely apply to your game. Because overlap in the top 20 matters more than order for recommendations, the more accurate tags you share with those comparables, the more likely you are to land in the same pool for More Like This and personalized recommendations. Next, move near-universal tags such as "Action," "Adventure," and "Indie" toward the back, and bring forward the niche tags that narrow down what your game is. Finally, check whether someone reading only your top five could picture the game.

Avoid copying popular tags. Inaccurate tags bring in players who don't fit, and those players neither buy nor play. If the signal Valve weighs most heavily is sales and play, impressions earned in the wrong pool can end up working against your next impressions. That is also why Valve tells developers to remove tags like "Family Friendly" on a horror game or "2D" on a 3D game.

The tag list itself changed in 2026. According to GamingOnLinux's report of May 19, 2026, Valve added 17 tags and removed 28. Clicker was renamed Incremental and Conversation became Dialogue Heavy, while Unforgiving was merged into Difficult and Jet into Flight. If your tags were set before this update, it is worth opening the Tag Wizard to check whether any of your top 20 were removed or merged, and whether any new tags (such as Bullet Heaven or Desktop Companion) describe your game more accurately. Tag publishing is handled separately from store page publishing, so you can fix tags without affecting pending store page changes.

How do games get into the Discovery Queue?

The Discovery Queue recommends 12 games to each user. In the two weeks after launch, most games receive some Discovery Queue exposure, but the amount can differ by more than an order of magnitude depending on pre-launch preparation and early performance. Before launch, only games with a concentrated burst of visits and wishlists make it in. Valve does not describe Discovery Queue selection in detail in its documentation, so the figures below come from published analyses of developer data.

Start with the post-launch Discovery Queue. In a February 2025 analysis of data from 75 games, Chris Zukowski of How To Market A Game found that for most games, the Discovery Queue accounted for almost half of total page visits in the two weeks after launch. Exposure varied widely by revenue tier, however.

Top-tier games received roughly 24,000 to 35,000 impressions per day on days one and two, and for the $1M-plus tier the median rose to 69,690 on day three. Games under $30,000 generally did not get those initial 25,000 to 35,000 impressions at all, and tended to share the same traits: fewer than 5,000 wishlists at launch, a short marketing period, a slow climb to 10 reviews, low review ratings, and low concurrent users. The lower tier's higher Discovery Queue share does not mean it got more Discovery Queue traffic; it means it had less traffic from everywhere else. The divergence from day three onward reads like Valve's "buying and playing" signal expanding or shrinking exposure in real time.

The pre-launch Discovery Queue is a much narrower door. In an October 2023 analysis, the same author looked at 14 games with significant pre-release Discovery Queue showings. Their numbers in the seven days before the peak were as follows.

Most games got in at around 10,000 visits, and wishlist counts clustered between 1,500 and 4,000. There was no visible correlation between wishlist count and the height of the Discovery Queue peak, and the exposure typically lasted about two weeks. The triggers developers reported were launching a Coming Soon page with a trailer, streamer coverage, a viral social post, Kickstarter, and festival appearances. One game saw 56,460 visits and 15,356 wishlists over two weeks after a TikTok video passed 715,000 views. Conversely, the author had not seen paid ads alone trigger a Discovery Queue appearance. In short, the pre-launch Discovery Queue appears to respond to a short burst of attention rather than to accumulated wishlists, with the caveat that the sample is 14 games.

The Discovery Queue's format has also evolved. In July 2022, Valve's Steam Labs Experiment 14 tested a queue that players could browse without leaving the page, that showed why each game was recommended, and that allowed one-click wishlisting. Valve also said titles deemed unsuited to the player would be hidden. When the reason for a recommendation is shown on screen, tags also do the job of explaining to the player why this game appeared.

How much do More Like This and Steam's recommendation algorithms rely on tags?

More Like This and tag pages depend heavily on tags, while the recommendation model trained on play history barely looks at them. Tags can decide which games you stand next to, but your position inside the recommendation model is ultimately decided by what the people who played your game actually do.

More Like This is the box of 12 games below a game's details page. Valve's tag documentation says the top 20 tags determine which similar games appear there, but adds the caveat that its tag comparison view is "a strictly tag-driven comparison" and not a direct reflection of the More Like This algorithm. According to GameDiscoverCo's Simon Carless, writing in March 2021, the "See All" page for More Like This then listed 27 games: 9 with particularly similar tags, 6 upcoming releases, 3 new releases, and 9 top sellers. He saw a shift after October 2018 toward the top 12 favoring popular, already-released games. The composition may have changed since, but the idea that More Like This weighs both tag similarity and popularity is still a useful frame. For an unreleased game to appear in a hit game's More Like This, tag overlap alone is not enough; it needs enough interest to earn an upcoming-release slot.

The recommendation model works differently. Valve's Interactive Recommender, released through Steam Labs in 2019, looks at how much you have played each game in your library and uses machine learning to make recommendations, according to Valve's own page, with tags offered alongside release date as filters to narrow the results. How much weight this recommender carries across Steam today has not been disclosed, but it points in the same direction as Valve's description of the main capsule as driven by what players spend time and money on.

Put the two together and the division of labor is clear. Tags decide which pool you enter on More Like This, tag pages, browse views, and filters. Playtime-based recommendations show your game to people similar to those who actually played it for a long time. Accurate tags mean the first people in are the right people, and when the right people play for a long time, the recommendation model finds the next ones. Inaccurate tags break that chain at the first step.

How do you audit tag and Discovery Queue performance in Steamworks traffic reports?

Audit tags and discovery in Steamworks' "Store & Steam Platform Traffic Breakdown" by separating impressions, visits, and click-through rate by surface, and by comparing the two weeks before and after a tag change. Steam has no tag A/B testing, so changing one thing at a time and comparing matched periods is the only way to verify an effect.

Start with the report definitions. Per Valve, impressions are tracked when your game is displayed on screen to a player, which does not guarantee the player saw it. Visits are unique page loads of your store page, and click-through rate is the share of impressions that led to a store page visit. Some impressions, such as friend-in-game notifications, do not link to the store page, so their click-through rate cannot be measured. Since March 23, 2021, impression and visit reports exclude users who opted out of the applicable cookies, so be careful when comparing against older data. The report lives per app under Marketing Tools & Data → Marketing & Visibility → "Store & Steam Platform Traffic Breakdown," and wishlist, purchase, and activation conversions from external campaigns are tracked separately with UTM links.

Here is how to read each surface and what to fix first.

If you change tags, verify over a four-week cycle. In week one, record surface-level impressions, visits, click-through rate, and wishlists for the two weeks before the change, and list the top tags of your comparable games. In week two, change only the tags, and leave the capsule, trailer, discounts, and external campaigns untouched during that period. In weeks three and four, compare the two weeks after the change on the same metrics, and hold off on conclusions if a discount, festival, or major update overlapped and shook traffic. For games earning fewer than 10 wishlists a day, a two-week comparison rarely shows a readable difference, so time tag changes just before traffic peaks such as a Coming Soon page launch or trailer release.

If you hand this work to a Steam strategist or marketing partner, a few questions reveal their experience. Which comparable games do they use to set tag order, and why? Which Steamworks metric do they use to judge the effect of a tag change? When aiming for a Discovery Queue appearance, when do they think external traffic should be concentrated? Can the contract use wishlists, purchases, and play as success criteria rather than visits? And do they keep traffic metrics, which Valve says are not visibility factors, out of the center of their reporting? If the answer to the last question is "we'll raise visits to push the algorithm," that explanation does not match Valve's documentation.

What happens after the visit creates the next impression: Playio's perspective

Reduced to one sentence, Steam's visibility works like this: tags show your game to the right people, and the more those people buy and play, the more people Steam shows it to. Pre- and post-launch marketing budgets should therefore be judged less by how many visits they bring and more by whether the people who arrive wishlist the game, buy it on launch day, and actually play it. Since Valve says traffic and conversion rate are not visibility factors, a channel that only raises visits cannot grow your visibility inside Steam.

Playio is a community of 5 million gamers that connects players with games that match their tastes, based on genre preferences, play history, and in-game behavior data. Its Steam product for PC Steam games is built so that each action in the funnel can be a campaign goal: wishlist and follow, add to cart, purchase, playtime milestones, and achievement completion. The default pricing model is CPI, and CPE based on specific actions is also supported. Set wishlists and follows as the goal before launch and purchases and playtime in the first two weeks after launch, and what the advertiser buys is not a page visit but the "buying and playing" behavior Valve says it weighs most.

This approach has a principle of its own. If you bring in players who don't match the game your tags describe, wishlists may pile up without turning into launch-day purchases or sustained play. That is why players are selected by genre preference, and why, when designing a campaign, sharing your top five tags and comparable games first lets us propose a matching player segment and target actions together. If your launch date and wishlist goal are set, we recommend starting with a short pilot timed to your Coming Soon or demo launch, and confirming the effect with wishlist and purchase numbers in your Steamworks reports.

More details are available here. (https://playioadsen.oopy.io/bizdeck)

For the full set of wishlist channels and launch timing, see Steam Wishlists Decide Your Launch, and for demos and Next Fest, see Steam Next Fest Marketing and How to Run a Steam Playtest Before Launch.

Key takeaways

Steam uses tags to decide which pool a game enters, and sales and play to decide how much more it shows the game within that pool. Valve names sales and play activity as the strongest signal, and states that store page traffic, visit-to-purchase conversion rate, review scores at 40% positive or above, and paid placement are not visibility factors. Wishlists count toward ranking only on upcoming lists and otherwise work through launch and discount notifications. Only the top 20 tags are used, the top five describe the game, some filters prioritize the first 15, and for recommendations, overlap with comparable games and niche tags matter more than order. Tags set before the May 2026 update (17 added, 28 removed, several renamed or merged) are worth reviewing. Published analyses show the Discovery Queue providing close to half of visits in the two weeks after launch, with exposure differing by more than an order of magnitude based on early performance, while the pre-launch Discovery Queue responded to short bursts of attention, at a median of about 11,000 visits and 3,600 wishlists in the seven days before the peak. Verify tag changes by comparing surface-level impressions, visits, and click-through rate for two weeks before and after in Steamworks traffic reports, and judge external channels by wishlists, purchases, and play rather than visits.

For inquiries about Playio's advertising solutions, reach out at: [email protected]