Steam Tag Optimizer

Search for the best Steam tags and you get a ranking of the most-applied tags on the store. Given what Valve says about rarity, that list is close to backwards: the tags at the top are the ones carrying the least weight in recommendations, precisely because everyone has them. A global popularity ranking cannot tell you anything, because it does not know what your game is.

A comparable set can. Pick six games a buyer would consider instead of yours and look at what their tag lists have in common, and you have described a neighbourhood: the browse pages those buyers actually open, the filters they actually tick, the recommendation clusters those games sit in. A tag five of your six comparables carry is load-bearing for that neighbourhood no matter how rare it is store-wide. Roguelike Deckbuilder is on a rounding error of the Steam catalog and on three of the four biggest deckbuilders on it. That asymmetry is the whole method.

It cuts both ways. A tag none of your comparables carry is spending one of 20 finite slots to put you somewhere your buyers are not. Sometimes that is a deliberate differentiator and worth it. Often it is a leftover from an earlier version of the game, or an aspiration. The tool names them and leaves the call to you.

How common is this problem

Across 405 Steam pages we ran through our analyzer between February and July 2026, a missing-discovery-tags finding was one of the most frequent results. Two things to say honestly about that number. First, our issue generator changed part-way through the period, so the rate is a range, not a figure: roughly 57% to 66% of coming-soon pages and 34% to 49% of released pages depending on which half of the period you look at. Second, that sample skews pre-launch -- 56% of it was coming soon -- and pages people bring to an analyzer are not a random sample of Steam. Both cuts point the same way: unreleased pages are under-tagged relative to released ones, which fits, because tags accumulate community weight after launch and a pre-launch page has none of that yet.

Keep the expectation honest, though. Simon Carless, who has looked at more Steam performance data than most, put it this way: there is convincing evidence that launching with no tags can badly hurt sales, but no compelling evidence that tag optimisation on its own makes a large difference. Treat this as removing an obvious deficit rather than as a growth lever. It is an hour of work that stops you being invisible on pages you belong on.

Steam sorts the tags on your page by weight: "Tags are ordered on your game page by the amount of weight each has on your game, as a result of your own sorting in the Tag Wizard. This can be modified over time as a number of players apply particular tags to your game." Your ordering sets the opening position. The community moves it from there. You do not get the last word, but you get the first one, and on a page with no reviews yet that is most of the word.

Valve makes exactly one numeric claim tying order to a discovery surface: "Some store filters prioritize the first 15 tags." It does not name the filters. Separately it says your top 5 "should paint a fairly clear picture of your game", which is framed as description rather than ranking, and that Steam "generally uses all of your 20 visible tags for generating recommendations", adding that for recommendations "the order is less important than the overlap between multiple tags".

So the practical split is: which tags you carry is what drives the algorithm, and what order they sit in is mostly about the human reading your page, plus whatever those unnamed first-15 filters are. Two very widely repeated claims are worth naming as unverified, because we could not find either in Valve's documentation: that the top 5 tags get extra algorithmic weighting, and that developer-applied tags outweigh player-applied ones. The second is contradicted by Valve's own line that player application increases a tag's weight. Treat both as folklore.

What this tool recommends, then: put the tags your comparable set agrees on in front, put the tags only some of them use behind those, and put anything none of them use at the back where you can see it and decide. Within each group the broadest tags on Steam sink to the bottom, because the recommender discounts them anyway. That is a defensible default, not a formula -- if a tag genuinely describes the first thing a buyer should know about your game, lead with it regardless.

On 18 May 2026 Valve rebuilt part of the tag catalog: 17 tags added, 28 removed, and several merged or renamed. Added were Bullet Heaven, Desktop Companion, Organizing, Cleaning, Decorating, Wuxia, Xianxia, Falling Blocks, Espionage, Samurai, Zoo, Wolves, Capybaras, Animals, Cult, Poker and Language Learning. Removed were a long list including NSFW, Mature, Kickstarter, Crowdfunded, RPGMaker, GameMaker, Masterpiece, Well-Written, Roguevania, Cult Classic and several licensed-IP tags -- Valve's stated reasoning being that community-sourced tags are a poor way to attribute intellectual property. Clicker became Incremental, Conversation became Dialogue Heavy, Pool became Billiards, Unforgiving merged into Difficult, and a handful of animal tags were pluralised.

If you have not opened the Tag Wizard since spring 2026, some of your slots may be holding tags that no longer exist. This tool checks your list against Valve's published catalog -- all 446 tags, read from Valve's own tag endpoint on 2026-07-30 -- and flags anything retired, naming the replacement where Valve provided one. Tags the announcement described as merged but that Valve still serves are deliberately left alone, so the check never cries wolf.

1. 1. Paste your own tags in the order Steam shows them. Copy them off your live store page, or out of the Tag Wizard if you have not published yet. Order matters to the analysis, so keep it.
2. 2. Add three to eight comparable games. Not the biggest games in your genre -- the games a buyer would actually consider instead of yours. Moderately successful ones are better comparables than megahits, whose tag lists are shaped by an audience you do not have yet. Paste their tags, or paste their AppIDs and let us read the store pages.
3. 3. Read the gap first. Tags at least half your set carries that you do not are the fastest fix on the page. Add the ones that honestly describe your game. Skip the ones that do not -- a wrong tag is worse than a missing one, because it puts you in front of people who will bounce.
4. 4. Then look at what you are spending slots on. Every tag no comparable carries is one of 20 finite slots. Keep the ones that are real differentiators. Drop the rest to make room.
5. 5. Apply the order in the Tag Wizard and publish tags. Tag publishing is a separate flow from your store-page changes, so it will not disturb anything else you have pending. Then leave it alone -- community weight takes time to settle, and re-shuffling weekly tells you nothing.