Methodology: How We Estimate Steam Game Revenue

Steam Estimate calculates a game's estimated sales by multiplying its total Steam review count by a base multiplier, then adjusting that multiplier for price, review sentiment, game age, genre, and any active discount. The result is optionally cross-checked against the game's current concurrent player count. Revenue is then derived from estimated sales, average sale price over the game's lifetime, and Steam's revenue share.

The sales formula

estimated_sales = total_reviews
                 × base_multiplier      (22x–30x, by review count)
                 × price_factor         (0.50x–1.15x)
                 × sentiment_factor     (0.75x–1.25x)
                 × age_factor           (0.75x–1.40x)
                 × genre_factor         (0.40x–1.45x)
                 × discount_correction  (1.00x–1.30x)

If concurrent player count data is available, the result is blended toward the range implied by that CCU (using an assumed active-player ratio of 0.2%–8%, depending on game age) whenever the review-based estimate falls far outside it.

Base multiplier (by review count)

Bigger games attract more casual buyers who are less likely to leave a review, so the multiplier drops as review count grows.

Under 100 reviews22x
100 – 49925x
500 – 1,99928x
2,000 – 9,99930x
10,000 – 49,99927x
50,000 – 199,99922x
200,000+18x

Price factor

Cheaper, more impulsive purchases are reviewed less often than premium purchases, so lower-priced games get a higher multiplier.

Free to play0.50x
Under $50.72x
$5 – $9.990.85x
$10 – $14.990.92x
$15 – $19.990.97x
$20 – $29.99 (baseline)1.00x
$30 – $39.991.05x
$40 – $59.991.08x
$60+1.15x

Review sentiment factor

Better-reviewed games tend to have more engaged players, who review at a higher rate.

Over 95% positive1.25x
90 – 95%1.15x
85 – 90%1.08x
80 – 85%1.03x
70 – 80% (baseline)1.00x
60 – 70%0.92x
50 – 60%0.85x
Under 50%0.75x

Age factor

Older games have accumulated sales that were never reviewed, so each existing review represents more sales the longer a game has been out. This also accounts for Steam lowering its review-prompt frequency before 2019.

Under 1 month old1.40x
1 – 3 months1.20x
3 – 6 months1.10x
6 – 12 months (baseline)1.00x
1 – 2 years0.95x
2 – 3 years0.90x
3 – 5 years0.85x
5 – 7 years0.80x
7+ years0.75x

Genre factor

Niche, engaged audiences (strategy, city builders) review at a higher rate than casual, broad audiences (puzzle, match-3), so the multiplier runs from 0.40x up to 1.45x depending on genre. A few examples:

Free to Play0.40x
4X0.70x
RPG (baseline)1.00x
Action1.10x
Horror1.15x
Puzzle1.35x
Match 31.45x

Discount correction

Heavy discounts bring in impulse buyers who review less often than full-price buyers, so review counts gathered during a sale understate real sales unless corrected.

No discount1.00x
1 – 24% off1.05x
25 – 49% off1.12x
50 – 74% off1.20x
75%+ off1.30x

From sales to revenue

Average sale price is modeled with a sigmoid discount curve — discounts ramp up quickly in a game’s first year and then plateau — and a further 15% reduction for regional pricing. Steam’s tiered revenue share is then applied: 30% up to $10M in gross revenue, 25% between $10M and $50M, and 20% above $50M.

Frequently asked questions

How does Steam Estimate calculate a game’s sales?

We multiply a game’s total Steam review count by a base multiplier (22x–30x, chosen by review count) and then adjust that multiplier for price, review sentiment, game age, genre, and any active discount. The result is an estimated total sales figure, which is optionally blended with an estimate derived from the game’s concurrent player count to catch outliers.

How accurate are the estimates?

The algorithm is calibrated against publicly known sales figures for reference titles such as Hades, Stardew Valley, and Baldur’s Gate 3, and cross-checked against each game’s concurrent player count. Accuracy depends heavily on review volume: games with 10,000+ reviews (High confidence) are the most reliable, while games under 1,000 reviews (Low confidence) should be treated as rough approximations.

Why might these numbers differ from SteamSpy or VGInsights?

Steam Estimate does not use SteamSpy or VGInsights data or methodology. It is built independently from Steam’s own public review, pricing, and player-count endpoints, using multiplier values we calibrated ourselves. Different methodologies will naturally produce different numbers for the same game.

How often is the data updated?

A scraper runs three times a day and cycles through all tracked genres and tags over about 3.3 days, plus refreshes the current top-sellers list on every run. Each game page shows its own “Last Updated” timestamp for the most recent scrape.

Are these verified sales figures?

No. Every figure is a statistical estimate. Where a developer has publicly disclosed real sales numbers, we show that verified figure alongside the estimate instead of replacing it.

All figures produced by this methodology are estimates only and are not confirmed sales or revenue data from Valve or any developer.