Steam Popular Upcoming Algorithm — What Signals the List Actually Cares About

Last updated: 2026-09-08 · Reading time: 5 min

Steam doesn’t publish the Popular Upcoming formula. What Valve does publish is the boundary: its visibility documentation states that with a few exceptions like the Popular Upcoming tab, wishlists are not a factor in a game’s algorithmic visibility on Steam — so this list is 1 of the few places your wishlist count does the work directly. Below is what we’ve been able to infer from observing unreleased-game rankings against their public wishlist counts, velocities, and genre clusters across 18 months of pre-launch data. These thresholds are ours, not Valve’s.

Signal 1: Total wishlist count (soft floor)

The strongest single signal. Empirically:

• Below ~5,000 wishlists: almost never surfaces on Popular Upcoming. The list isn’t a hard cutoff (some genre niches surface games at lower counts), but in practice, sub-5K games rely on direct-search and external traffic.
• 5,000-10,000: intermittent surfacing during low-competition windows (typically January-February or mid-summer when fewer launches compete for the list).
• 10,000-25,000: consistent surfacing on genre-filtered views; less reliable on the global Popular Upcoming front page.
• 25,000+: reliable global surfacing in the days before launch.
• 100,000+: top-20 Popular Upcoming positioning likely for several days surrounding launch.

Caveat: regional restrictions and price tier affect surfacing. A game priced under $5 may surface less prominently on US/EU Popular Upcoming than its wishlist count alone would predict.

Signal 2: Wishlist velocity (heavily weighted)

Recent wishlist additions count more than dormant ones. A game with 10K wishlists that grew 80% in the last 90 days outperforms a 30K-count game that’s been flat for 6 months. The algorithm seems to weight roughly the last 30-day window most heavily, with diminishing weight back to ~180 days.

Practical implication: the announcement-spike-then-silence pattern is the worst trajectory shape for Popular Upcoming. Games that announce, get a Twitter spike of 5K wishlists, then go dormant for 6 months consistently underperform games that grew slower but kept compounding. Steam reads dormancy as "this game has lost momentum."

Signal 3: Tag + genre cluster match

Popular Upcoming has both a global front-page view and genre-filtered subviews. A roguelike with 8K wishlists is more likely to surface on the Roguelike Popular Upcoming subview than on the global front page. Implications:

• Tag accurately. Cargo-culting popular tags that don’t fit your game gets the game surfaced to wrong audiences who bounce, hurting downstream conversion signals.
• Pick the right primary genre. "Strategy" and "Roguelike" are crowded; niche tags like "Auto Battler" or "Survival Crafting" are easier to surface in but smaller TAM.
• Cluster with breakout neighbors. If your game shares 3+ tags with a recent breakout (Vampire Survivors, Dredge), you may get tag-cluster surfacing benefits when users browse "more like X."

Signal 4: Day-1 sales rate (post-launch reweighting)

Once you launch, Popular Upcoming becomes Recently Released. The new list reweights based on early sales conversion: high wishlist + low Day-1 conversion gets demoted; the opposite gets boosted. The wishlist-to-sale conversion benchmarks show what a healthy launch-week rate looks like by genre, so you can tell whether your count is likely to convert or get demoted. This is the algorithm’s defense against wishlist farms: lots of wishlists with no buyers is a stronger negative signal than a low absolute sales count.

What you can actually do

1. Hit the wishlist floor before launch. Aim for ≥10K wishlists with a non-flat trajectory in the 90 days before launch. Per-genre benchmarks →
2. Maintain wishlist velocity. Schedule announcements, demos, and content releases to keep the trailing-30-day wishlist additions positive. Dormancy hurts more than slow growth.
3. Tag accurately and prune. Use the tags users actually click on for your game in playtest feedback, not just the highest-traffic tags.
4. Don’t announce too early. An announcement-to-launch window over 18-24 months invites the dormancy trap. The "1-year ahead with frequent content beats" pattern from successful indies is what the algorithm rewards.
5. Plan demo timing carefully. Demos drive a wishlist spike followed by a velocity dip if there’s no content to follow. Either run multiple demos or release one demo close enough to launch that there’s no dip.

What we still don’t know

Steam doesn’t publish the formula and they update it without notice. The signals above are stable observations across the last 18 months but should be treated as "what works empirically" not "what Steam guarantees." Specifically unclear:

• Whether and how follower count (a separate metric from wishlists) factors in.
• How regional weighting works: whether a game popular in CIS but not US/EU gets penalized on the global front page.
• Whether negative reviews on existing titles by the same developer suppress the new game’s ranking.
• Whether Steam Next Fest participation adds a temporary boost or is purely a wishlist-velocity vehicle.

Frequently asked questions

Built by Greg C., senior software engineer with production ML experience in calibrated prediction. Steam Launch Forecaster trains a CQR-calibrated model on a 77K-app Steam corpus. See the methodology →