The primary objective of this analysis is to establish a modern benchmark for estimating Steam game sales based on publicly available user review counts. By introducing the "NB number" (New Boxleiter number), the research provides developers with a more accurate tool for market estimation, accounting for significant shifts in user behavior following Steam’s late 2019 implementation of automated review prompts.
The findings are derived from a survey of 237 non-outlying Steam titles. The data indicates a clear downward trend in the sales-to-review ratio over time. While historical estimates from 2018 suggested an average of 82 sales per review, the current aggregate average across the sample is 63, with a median of 58. This ratio is highly dependent on the year of release: games launched before 2017 maintain an average of 81 sales per review, whereas titles released in 2019 dropped to an average of 57. For games launched in 2020, the average further declined to 41 sales per review, with a typical range of 20 to 60.
The analysis concludes that Steam’s UI changes have successfully encouraged more reviews, thereby compressing the ratio of sales to reviews. It cautions that these figures should be used as a general estimation lens rather than a precise financial calculator. When projecting revenue, developers must account for additional variables, including average store pricing, the frequency and depth of discounts, platform commission fees, taxes, and refund rates, which collectively can reduce gross estimates by 50% or more.