How to estimate a Steam game's sales and revenue

Estimate a Steam game's sales by multiplying its total review count by a review multiplier: research suggests about 20 to 55 sales per review for games released since 2020, with 30 as a central case, and higher for older games. To estimate revenue, multiply units by a realistic average price after discounts, regional pricing, VAT and refunds, then remove Steam's 30% cut. Always report a low, central and high range.

Can you see how many copies a Steam game sold?

No. Valve does not publish sales figures for other people's games, and most developers do not share theirs. Every public number you see, including on ReviewPulse, is an estimate built from signals Steam does show: review counts, prices, player counts, followers and chart positions. The honest output of any method is a range, not a number.

The most widely used shortcut is the review multiplier, often called the Boxleiter method or Boxleiter number. It works because a fairly stable share of buyers leave a review, so reviews act as a proxy for units sold.

What is the Boxleiter method?

The method is simple: estimated units sold = total Steam reviews × a multiplier. The difficulty is choosing the multiplier, because the ratio of buyers to reviewers has changed over time and differs between games.

Published research gives these ranges:

Two points stand out. First, the multiplier has fallen: Steam now prompts players to review more often and developers ask for reviews more actively, so recent games get more reviews per sale. Second, even a good model has wide error bars. Video Game Insights reported that its ranges held for over 80% of games in its dataset, which means roughly one game in five fell outside them.

The same research notes some directional patterns: higher-priced games tend to have lower multiples, MMOs have higher multiples, and sports and racing games lower ones. Birkett found no meaningful link between review score and the multiplier.

How do you estimate a game's units sold, step by step?

1. Get the total review count. Use the "All reviews" figure on the store page, not "Recent reviews". Check whether the count includes reviews from key activations; Steam lets you filter by purchase type.
2. Pick a multiplier range by release year. For games released in 2020 or later, start with 20x to 55x and 30x as the central case. For older games, move the range up.
3. Adjust for the game's profile. Move towards the low end for higher-priced, story-driven or very community-active games. Move towards the high end for cheap, casual or viral games that many people buy without engaging further.
4. Multiply and keep the range. Report low, central and high figures. Do not collapse them into one number.
5. Cross-check with other signals (see below) and move within the range if they agree.

Example: Hades. As of 31 October 2026, Hades (Steam app 1145360) shows about 308,000 Steam reviews across all purchase types. At 20x that is about 6.2 million units, at 30x about 9.2 million and at 55x about 16.9 million. That spread is the point: the method tells you Hades is a multi-million seller on Steam, not exactly how many millions.

How do you turn units into revenue?

Multiplying units by the US launch price overstates revenue a lot. Four things sit between the sticker price and what the developer receives.

1. Discounts

Over a game's life, many copies sell during seasonal sales, launch discounts and bundles. Valve's own rules set some limits: a game generally can't be discounted within 30 days of release (other than a launch discount) or within 30 days of a price increase, according to Steamworks pricing documentation. After that, discounts are common.

2. Regional pricing

Steam supports prices in 37 currencies across 4 region groups, and most developers charge much less in lower-income markets. On 31 October 2026, Hades cost $24.99 in the US, 24,50€ in Germany, 114,99 zł in Poland, R$ 73,99 in Brazil, ₹1,100 in India and $8.25 in Turkey, where Steam has priced games in US dollars since November 2023. A sale in India or Turkey earns a fraction of a US sale.

3. Taxes and refunds

Prices in the EU, UK and many other markets include VAT or similar taxes, which do not go to the developer. Steam also refunds purchases requested within 14 days with under two hours of playtime. Valve does not publish average refund rates, and they vary a lot by game length and genre.

4. Steam's cut

Valve keeps 30% of a game's Steam revenue up to $10 million, 25% between $10 million and $50 million, and 20% above $50 million, under tiers it introduced in late 2018 (TechCrunch). Almost every indie game stays in the 30% tier. If the game has a publisher, the developer's share comes after that.

A worked revenue estimate

Take a hypothetical indie game released in 2024 with 1,000 Steam reviews and a $14.99 US price. Because there is no public data on average realised prices, the table below uses an explicit assumption that the average copy earns 50% to 70% of the US list price after discounts, regional pricing, VAT and refunds. Change that assumption if you know the game sells mostly at full price or mostly in sales.

A range of roughly $105,000 to $404,000 for the developer, before any publisher share, is the honest answer. Birkett made the same point in 2018: because discounts are unknown, review counts alone cannot pin down revenue.

Why can review-based estimates be badly wrong?

• Free-to-play games break the method. Reviews track players, not purchases, and revenue comes from in-game spending.
• Keys, bundles and giveaways add players (and sometimes reviews) without matching revenue.
• Viral hits often sell to many casual buyers who never review, pushing the real multiplier up.
• Review bombing or campaigns spike review counts without extra sales.
• Early Access games may have reviews from several versions of the game and very different pricing over time.
• Other platforms are invisible. A game selling well on Switch, PlayStation or Xbox does not show it in Steam reviews.
• Subscription deals such as Game Pass can bring in revenue that never appears as Steam units.

What other signals help you check the estimate?

• Concurrent players. Steam shows current players, and third-party trackers record peaks. A sudden jump in players usually means a sale or update drove purchases.
• SteamSpy owner ranges. SteamSpy gives broad owner bands that can confirm whether your estimate is in the right order of magnitude.
• Followers and wishlist signals. Follower counts and appearances in Steam's top wishlist and top seller charts show demand before and after launch.
• Public statements. Developers sometimes announce milestones ("1 million copies sold"), though these often combine all platforms.
• Paid trackers. Services such as GameDiscoverCo and Video Game Insights run their own sales models. Compare their ranges with yours rather than treating any single figure as fact.

ReviewPulse combines review counts, live Steam data and SteamSpy into a sales and revenue estimate shown as a range on every game overview, and the investor view adds momentum and comparable games. Use it as a starting point for the same reasoning described here, not as a replacement for it.

How should you use these estimates?

Use them to compare games and to size markets, not to value a single game precisely. They are good for questions like "do games like ours usually sell 5,000 or 50,000 copies?", "which competitor is growing fastest?" or "is this genre big enough to justify a two-year build?" For your own forecasts, combine them with first-party data: your wishlists, your demo numbers and, if you have one, your publisher's sales history.