Every studio releasing on Steam still asks one familiar question: how many copies did that game really sell? Despite over a decade of better analytics, official store data remains private. Developers and analysts continue to rely on review-to-sales ratios as a guiding formula.
This estimation culture, shaped by the community itself, persists because transparency from digital stores continues to be minimal.
The review-based estimation model was made popular years ago by Mike Boxleiter, whose formula compared the number of public Steam reviews to assumed total sales.
Jake Birkett refined it in 2018, suggesting around 80 sales per review as a working average. Fast-forward to 2025, and that figure has shifted dramatically, according to the data collected by Simon Carless’s GameDiscoverCo newsletter.
Carless introduced what he called the “NB number,” short for New Boxleiter, showing how review ratios have evolved alongside player behavior and Steam UI changes.
Also read: Paramount Skydance: CEO David Ellison’s $2 Billion Restructuring Cuts 2,000 Job
The NB number, now averaging about 63 sales per review, helps creators estimate a game’s commercial success using transparent, crowd-sourced insight rather than guesswork.
Simon Carless and the Modern Game Data Mindset
Simon Carless, a longtime analyst and gaming industry veteran, has spent years demystifying how games gain traction. Through his GameDiscoverCo newsletter, he translates invisible patterns, player behavior, the timing of releases, and store algorithms into digestible knowledge for developers.
His tone is practical, driven by numbers but rooted in empathy for small studios struggling to gauge success.
What makes his work relevant is how it blends human insight with platform data. Instead of strict formulas, Carless emphasizes statistical reasoning, a mindset where creators understand the context behind ratios. He reminds developers that while reviews can signal engagement, they are not direct proof of revenue.
Steam’s introduction of the “Would you like to review this?” button in late 2019 marked a notable behavioral shift.
Players who once ignored reviews started participating, dramatically increasing feedback rates. Consequently, games released after 2020 began showing ratios closer to 40–50 sales per review, compared to 70–80 in earlier years.
Through his analysis of 237 participating developers, Carless uncovered that newer game releases settle around 20–60 sales per review. This compression of ratios is not just statistical noise; it reflects a cultural shift where community feedback became more normalized and incentivized by Steam’s evolving interface.
From Boxleiter to NB: How the Ratios Evolved
Originally, the Boxleiter model functioned as something of an industry myth. Developers needed a way to guess sales without access to official figures, and the review multiplier became shorthand for internal forecasting. It worked because user reviews were public, visible, and consistent across every game page.
Over time, however, this method needed recalibration. Jake Birkett’s 2018 survey refined the original number to around 82 sales per review on average, with medians in the mid-70s.
Then came Carless’s massive update, which introduced the NB number. By collecting input from hundreds of developers, he identified that review behavior wasn’t static. Game genre, release date, pricing model, and discount frequency all altered ratios.
Free-to-play titles, for example, distort averages since their download counts outpace paid conversions. Carless proactively removed outliers like those from the final dataset, strengthening the reliability of the finding.
What emerged is a modern ratio more suited to post-2020 Steam, smaller, tighter, and reflective of how players now engage with store ecosystems that prompt feedback more often than ever.
Why These Ratios Still Matter
Even with limitations, the NB number remains a valuable compass. Until platforms like Steam, Epic, or even the PlayStation Store begin sharing standardized sales visibility, ratios guide everything from marketing budgets to investor pitch decks.
Many developers, especially indie teams, use them as informal yardsticks for understanding where their titles stand in the sales hierarchy.

However, Simon Carless consistently warns readers about overestimating net revenue. A developer might look at an NB number, multiply by review count, and assume full-price earnings.
Reality is less optimistic. Variable discounts, VAT, refunds, and Steam’s standard 30% platform fee can reduce the actual net return to less than half the gross estimate.
He frames this as an educational opportunity rather than a discouragement. Transparency is not just about publishing sales numbers but building a smarter developer culture that knows what those numbers mean.
By adjusting assumptions around pricing and revenue share, creators can forecast more realistically. It’s a way of turning the NB ratio from rumor into responsible financial modeling.
The Broader Impact on Digital Store Ecosystems
The influence of review-to-sales modeling extends beyond Steam. As new digital marketplaces expand, from Epic Games Store to Game Pass integrations, developers increasingly want cross-platform compatibility.
Yet without public-facing numbers, review counts remain the one universal breadcrumb trail left to analyze performance.
Console platforms, however, rarely show player reviews, creating an imbalance in perceived transparency. Analysts predict that platforms like Xbox and PlayStation may eventually need to address these data gaps if they want to attract self-publishing developers who rely on these open, review-based insights for benchmarking.
Furthermore, public discussion around the NB number aligns with the modern trend of data storytelling, where community-shared science replaces secrecy.
Just as social media engagement metrics evolved from vanity figures into business intelligence, Steam review multipliers form a grassroots data framework powering business decisions today.
Practical Uses for Developers in 2025
For creators planning a release this year, applying the NB number responsibly can mean understanding both potential and risk.
If a new indie title receives around 1,000 reviews within a few months, the NB benchmark suggests approximately 40,000–60,000 copies sold. But those numbers fluctuate depending on genre, price tier, and review positivity.
According to Carless’s data, games averaging 84–89% positive reviews tend to maintain stronger long-tail momentum.
The correlation between user sentiment and continued sales proves that community trust still converts directly into commercial staying power. The NB number may reveal immediate sales, but review quality and frequency indicate sustained relevance.
Developers analyzing their own performance should view reviews as more than sales markers. They are evidence of engagement depth, showing how many players felt compelled to share their experience. As such, the modern NB ratio is not simply an economic equation but a behavioral reflection of post-2019 player culture.
Toward a More Transparent Future
Game industry progress often depends on pioneers who share knowledge instead of guarding it. Figures like Simon Carless represent a shift away from competitive secrecy toward collective education. His NB number survey demonstrates how goodwill, statistics, and collaboration can reframe industry understanding.
As we move deeper into 2025, discussions about revenue transparency are expanding beyond estimation formulas. Storefronts are increasingly pressured by both players and creators to demystify success metrics.
Even if Steam never reveals complete data, community-driven analysis will fill the gap, continuously refining these unofficial but powerful ratios.
Developers worldwide still gather around shared spreadsheets and newsletters to make sense of it all. Their motivation remains consistent: understanding real performance beyond marketing hype.
Until public APIs unlock direct data access, the NB number stands as both a mathematical constant and a cultural symbol, proof that when the ecosystem hides answers, the creative community learns to measure itself.

























