Steam Puts a Review Label on Every Game, but Valve Publishes Only One Number Behind It: 40 Percent Positive

Watch what happens when you open a Steam store page. Up top sit two review scores: one covers the last 30 days, and one covers the game's whole life on the store. Each score boils down to a thumbs-up or thumbs-down call from players, and Steam turns the counted votes into a percentage and a short label you can read at a glance. Here's the part most threshold charts skip: Valve's own Steamworks documentation, checked in full for this piece, publishes no table that maps a percentage and a review count to each label. The documentation gives you the machinery — who can write a review, whose review counts, how the scores get displayed — and it gives you exactly one number. That number's 40 percent.

So read every label the way you'd read a scoreboard with one published line on it. The label tells you where a game landed. The documentation tells you how the counting works, and it draws its one bright line where Steam's featuring decisions start to care.

Steam's one published number is 40 percent positive, and it decides whether Steam features a game

Valve states the rule in its Visibility documentation, and the User Reviews page repeats it in a FAQ answer: as long as a game's reviews are Mixed or above, at 40 percent positive or better, review score isn't a factor in algorithmic visibility. Dip below 40 percent, into Mostly Negative, and the game will be less likely to be featured. That's the whole published threshold — one line, aimed at the store's machinery, not at bragging rights.

Translate the number and it gets practical fast. Forty percent means 4 counted thumbs up out of every 10 counted reviews: 40 out of 100, or 200 out of 500, holds a game exactly on the line. One counted thumbs down past that split drops the share under 40 percent — and that's the moment featuring gets harder. Above the line, a higher score still works on people, because shoppers read labels before they read reviews. It just doesn't buy extra push from Steam's algorithm, because Valve says score stops counting as a visibility factor once a game clears Mixed.

Steam shows two scores on a store page, one for the last 30 days and one for the game's whole life

The aggregate of positive and negative reviews is used to calculate a review score — that's the User Reviews page's own summary of the system, and the page is careful about what aggregate means. Only reviews from accounts that bought the game on Steam count toward the score, which the next section pulls apart. The two scores sit at the top of the store page, and they answer two different questions. The lifetime score tells you how the game has landed across its whole run. The 30-day score tells you how the game in front of you today is landing right now — after the latest patch, the latest price, the latest server wobble.

Watch what happens when the two scores disagree, because that gap is the most useful reading on the page. A strong lifetime label beside a weak recent one usually means something changed late: a rough update, a monetization shift. A weak lifetime label beside a strong recent one usually means a game climbed out of a bad launch. Neither score is a verdict on its own. Together, they're a before-and-after photo, and the two windows — 30 days and lifetime — are the frames Valve built for that comparison.

A review you can read on Steam doesn't always count toward a game's score

Purchase isn't required to write a review. Any customer who has launched a product through Steam and logged playtime can write one: a player who activated a key bought elsewhere, a player who tried the game on a free weekend, a player who borrowed it through Family Library Sharing. The review posts, the playtime shows beside it, and other shoppers can read every word. Steam just doesn't count that review toward the score. Only reviews posted from an account that purchased the game on Steam and played it move the percentage, so the count under a label is smaller than the pile of reviews you can scroll through.

The display rules add two more tells. A reviewer who got the game for free can flag that fact, and the documentation recommends the disclosure for press and promotional copies. A review written during Early Access keeps its mark after full release, and a pre-release review carries its own mark — so a glowing verdict from a half-finished build announces its vintage on the page. Prominence fades, too: Steam surfaces recent reviews other players found helpful, and older reviews phase out over time.

Valve bans developers from paying for reviews or asking for them inside the game

The User Reviews page sets three rules for developers, and they're blunt. Don't attempt to abuse or artificially manipulate the review system. Don't solicit reviews in exchange for games, DLC, money, or other rewards — with one exception carved out for sending a copy to press or internet personalities for previews or reviews. And don't ask customers to review your product from within your application. No in-game pop-up begging for a thumbs up, ever; Valve wrote that one down in plain words.

Customers play by a rulebook, too: no abuse, no threats, no commercial use, and no artificial score games — multiple accounts, coerced reviews, and paid reviews all sit on the banned list. A copy received for free should be disclosed inside the review itself. Anyone can flag a review that breaks the rules, and a confirmed violation gets banned, which removes it from Steam and removes its weight from the score. A review flagged off-topic works differently: its visibility drops, but it still contributes to the score.

Review bombs get their own machinery, the strongest exception in the document. A review bomb — a crowd using reviews to drag a score down and provoke a response — triggers a Valve review when the volume and sentiment swing looks anomalous. If the surge is unrelated to the product, Steam buckets that abnormal period and the reviews inside it don't contribute to the overall score. If the surge traces back to a broken feature or a fresh bug, the advice is equally plain: fix the product, and weather the storm.

A developer can answer a review on Steam, but a developer can't delete one

Developers can't pull a review down directly. They can flag one, and the review stays listed until a moderator deletes it or clears the flag. What developers can do is answer: a reply through the provided response area appears right under the review, marked as an official developer response. The documentation then spends real space talking developers out of overusing the button. Not every review needs a response, a response draws more attention than the original post, and arguing with a reviewer reads as an attempt to silence the crowd — the page's phrase for the trap is that it's a battle you won't win.

The best-practices list reads like hard-won patch notes. Limit the time you spend reading reviews. Pull the high-level feedback out, then get back to the product. Let your community answer each other's questions. And if your game is getting bombed, protect the person on your team the crowd has picked as its target — the documentation names that cost, the mental and emotional toll on staff, as potentially the biggest one in the whole event.

So here's your next run at any store page. Read the recent score first and the lifetime score second, because the gap tells you whether you're buying the game that launched or the game that exists now. Check the review count under the label before you trust the word on it, and remember whose thumbs built that percentage: Steam purchasers only, with key activations and free-weekend verdicts readable but uncounted. And open the recent reviews Steam surfaces, since the documentation says those reflect the game's current state. For the other letters on that page, the age rating in the corner, this desk's explainer on how the ESRB rates a game covers who assigns those. Grab the second controller. Watch what happens when you play it.