Why Your Steam Game Isn’t Selling: A Diagnostic Guide

A Steam game that isn’t selling has exactly one of three problems: a traffic problem (not enough people see it), a conversion problem (people see it and don’t act), or a product/price problem (people buy it and bounce). Each one shows up as a specific number in your Steamworks dashboard, with a known healthy range and a different fix. Treating a conversion problem with more marketing, or a product problem with a new capsule, burns months and money on the wrong branch.

This post is the decision tree. Pull five numbers from Steamworks (about 20 minutes), compare them to the benchmarks below, and you’ll know which branch you’re on before you change anything. More than 19,000 games shipped on Steam in 2025 (SteamDB’s release stats count roughly 19,000-19,500; Chris Zukowski’s tally puts it at 20,282), and only 608 of them ever reached 1,000 reviews. Most of those developers never diagnosed which problem they had. You’re going to.

Everything you need is in Steamworks under App Admin > Marketing & Visibility > Store & Steam Platform Traffic Breakdown, plus your sales and refund reports. The official traffic reporting docs define the terms: an impression is your capsule being shown to someone, a visit is a unique load of your store page.

Write down, for the last 30 days:

One warning before you read your own numbers: impressions are not unique people. The same shopper can generate an impression in the Discovery Queue, again in search, again on a tag page. And traffic sources behave differently, so look at each source’s line, not just the aggregate. The rest of this post walks the branches in order, because the funnel multiplies: impressions x CTR x page conversion = wishlists and sales. Diagnose top-down.

The number to check: total impressions in the Traffic Breakdown, last 30 days.

The benchmark: there is no single official number, but Zukowski’s benchmark data gives a useful post-launch yardstick: even games that finish in his Bronze tier (under $10K lifetime revenue) typically pull around 22,000 Discovery Queue impressions in their first 10 days. Games that end up earning $1M+ pull around 665,000 in the same window.

If your impressions are a rounding error - hundreds per day, not thousands - Steam’s recommendation systems have decided your game isn’t worth showing. That happens for three reasons:

Your tags are wrong or thin. Tags decide which queues, tag pages, and “more like this” slots your capsule can even appear in. A game with 5 vague tags competes everywhere and wins nowhere. You want 15-20 accurate tags, with the most specific genre tags first. Run your set through the Tag Optimizer and compare against the games that actually share your audience; our best Steam tags for 2026 breakdown covers which tags carry traffic.

Your algorithm state is cold. Steam’s visibility systems feed games that show momentum: wishlist velocity before launch, sales velocity after. A page with no external traffic gives the algorithm nothing to amplify. This is the engine-and-fuel model from our Steam algorithm explainer - the store multiplies interest, it does not create it. If you launched quiet, the fix is an external spike (a content update with a visibility round, a festival, a streamer hit, a discount) that gives Steam a signal to work with.

You haven’t given Steam a reason to re-look. Post-launch games get re-surfaced by events: discounts (every 30 days is the cooldown), seasonal sales, major updates. A game sitting untouched for six months has near-zero organic impressions and that is normal, not a verdict.

The fix, in order: fix tags first (free, one afternoon), then plan an external traffic push tied to a discount or update. Do not spend on ads yet - if the next two branches are broken, paid traffic converts at a loss. The working ceiling for paid acquisition is $1-2 per wishlist; above $3, stop and rework the page first.

The number to check: visits divided by impressions, from the same Traffic Breakdown table.

The benchmark: from our Steam CTR benchmarks, built from aggregated developer data: below 2% is poor and self-reinforcing (Steam shows low-CTR capsules less), 2-4% is average, 4-7% is good, and 7%+ is excellent. Compute it per traffic source; Discovery Queue CTR and search CTR are different animals.

Honesty note: Zukowski himself argues developers worry about CTR too much, because impressions and page views predict outcomes better than the ratio between them. He has a point for healthy games. But a sub-2% CTR with decent impressions is still a diagnosis: people are seeing your capsule and choosing not to click, and that specific failure has exactly two causes.

Cause one: the capsule. At thumbnail size, in a dark UI, next to 11 other games. The fixes with the best track record are brutal simplification: one clear focal point, high contrast, no marketing text, readable genre signal in half a second. I’ve seen 2-3 percentage point gains from stripping a busy capsule down. Check the technical side with the Capsule Validator.

Cause two: the audience. If your tags put a cozy farming sim in front of extraction-shooter players, no capsule saves you. Wrong-audience impressions look identical to bad-capsule impressions in the aggregate number, which is why you check CTR per source. Strong tag-page CTR plus weak queue CTR usually means the capsule is fine and the broad-audience targeting is off.

The fix: capsule redesign and tag correction, then wait. A capsule change needs 48-72 hours of new impression data before the numbers mean anything.

The number to check: pre-launch, wishlist additions divided by visits. Post-launch, (wishlists + purchases) divided by visits.

The benchmark: from our store page conversion benchmarks: the median visit-to-wishlist rate for Coming Soon pages is 8-15%. Below 8%, fixing the page beats buying more traffic, every time. Post-launch, full-price purchase conversion averages 3-5% of visits, and direct traffic (people who came specifically for your game) should convert at 15-25% - if it doesn’t, the page itself is failing even your warmest audience.

This is the branch where most “I have traffic but no sales” games live, and it is the most fixable. The usual suspects, roughly in order of damage:

• •A trailer that takes 20+ seconds to show gameplay
• •Screenshots that don’t communicate what you do in the game
• •A short description that describes lore instead of the loop
• •No genre anchor (“like X meets Y”) anywhere above the fold
• •Price out of line with the visible content

We keep a full ranked list in the 12 most common store page mistakes, and the complete Steam store page optimization guide for indie games is the systematic pass through every element. If you want the diagnosis automated, run your page through the free Steam Page Analyzer - it scores capsule, trailer, screenshots, description, and tags against the benchmarks in this post and tells you which element is dragging the number:

One more segmentation trick before you rebuild anything. Check conversion by traffic source in Steamworks. Discovery Queue traffic converting at 3-8% is normal (low-intent browsing); if your search traffic converts below 10%, you likely have a capsule-to-content mismatch - people clicked expecting one thing and your page delivered another. Different fix: alignment, not polish.

If impressions, CTR, and page conversion are all in range and revenue still disappoints, the store page is no longer the suspect. Two numbers tell you where the problem moved.

Refund rate. Steam’s refund policy grants refunds within 14 days of purchase at under 2 hours of playtime, nearly no questions asked. From our refund rates data: the median indie game loses 10-15% of gross unit sales to refunds. At 15-22% something specific is off - the game is shorter than the page implies, it crashes on common hardware, or the store page overpromises. Above 22%, you have a product problem, full stop, and the first two hours of the game are where it lives. Read your refund reasons in Steamworks; players tell you exactly why.

Review score. Below 10 reviews you have no label at all and every review-score filter on Steam skips you. The boundaries that move shopper behavior are 70% (Mixed becomes Mostly Positive) and 80% (Very Positive) - the full grid is in our review score thresholds table. A Mixed label actively deters purchases and dampens visibility at the same time, which is why a two-point move around a boundary matters more than a ten-point move in the middle of a band.

Price sits in this branch too. A refund spike concentrated at the 60-90 minute mark on a $25 game is players deciding the content doesn’t justify the price. The options are a price correction (note: a price increase blocks discounting for 30 days), a content update that changes the value math, or a discount strategy that finds the price the market already voted for.

Then you don’t have a broken funnel, you have a small one. All ratios in range, applied to too few impressions, produces exactly the quiet result you’re seeing. The math is unforgiving: 20,000 impressions x 3% CTR x 10% conversion is 60 wishlists. No ratio fix turns that into a living.

Two honest checks for this case:

Wishlist count at launch. Wishlists convert to first-week sales at rates that scale with the pile size:

If you launched with 800 wishlists, a healthy funnel was never going to produce more than a couple hundred first-week sales. Our wishlist conversion rates guide covers the day-1 and week-1 math (median day-1 conversion runs about 12%).

Market size. Sometimes the game is well-presented to an audience that is simply small at your price point. Model the realistic outcomes with the revenue calculator - plug in your genre, price, and wishlist count, and it applies conversion rates, the reviews-to-sales multiplier, refund rates, and Steam’s 30% cut in one pass. If the honest ceiling is $8K, that’s not a diagnosis you can fix on the store page; it’s information for the next game.

Run these in order. Stop at the first failure; that’s your branch, and fixing it re-opens the branches below.

1. 01Impressions, last 30 days. Thousands per day post-launch, or growing week-over-week pre-launch? If not: traffic problem. Fix tags, plan an external push.
2. 02Tags. 15-20 accurate tags, most specific first? Run the Tag Optimizer.
3. 03CTR overall. Visits / impressions at 2%+? Below: capsule or audience problem.
4. 04CTR by source. Tag pages fine but queue weak = targeting. Everything weak = capsule.
5. 05Visit-to-wishlist rate. At 8%+ (pre-launch) or combined action rate in range post-launch? Below: page conversion problem - start with the mistakes list.
6. 06Search-traffic conversion. Above 10%? Below: capsule-to-content mismatch.
7. 07Review score. 10+ reviews and 70%+ positive? Below either: review recovery is the priority.
8. 08Refund rate. Under 15%? Above: read refund reasons; fix the first two hours or the price.
9. 09Wishlist balance. Enough wishlists for the launch you expected (a first week of roughly 15% of the pile)? If not: the problem is audience size, not the funnel.
10. 10Modeled ceiling. Does the revenue calculator say your genre/price/wishlist combination can produce the number you want? If not, the fix is scope, price, or the next project - and that’s a real answer too.

Twenty minutes of Steamworks archaeology beats six months of guessing. Find the branch, fix the number, re-measure in two weeks.