Skip to content

Stress-test your Steam launch budget.

Shipping a Steam game in the next 90 days? You get a revenue floor to budget against, not a single guess. A failed indie launch typically wastes $5,000–$50,000 in marketing spend; the P10–P90 range lets you size that budget against the floor as well as the median. Each forecast comes with five real comp launches, causal estimates of marketing-lever effects with confidence intervals, and a Total Lift Attribution tool that recovers the campaign wishlists Steam’s UTM dashboard misses (~75% under-reported).

Free on every game: the storefront-only cone (better-than-4-in-5 realized coverage, n=6,422 held-out launches, validation run 2026-06-11), Boxleiter cross-check, top reasons, comp set, and a sample what-if lever. The $299 one-time launch report unlocks the wishlist-aware cone (81–86% realized per wishlist tier), every what-if decision lever, and Total Lift Attribution. One launch, no subscription.

14-day full refund if you haven’t uploaded Steamworks data yet. 50% refundable thereafter. Terms →

4 in 5
held-out launches land inside the cone — and if yours doesn’t, you don’t pay
Better than 4 in 5 on n=6,422 held-out launches · both the 2026-06-11 and 2026-09-08 validation runs clear it · exact figures
050100

What’s in the $299 single launch report

One launch, one purchase, no subscription. Everything below is delivered through your authenticated dashboard the moment payment clears.

Calibrated revenue cone

P10 / P50 / P90 lifetime revenue estimates with validated coverage. When your game is unlike anything in our 77K-app calibration corpus the cone gets wider, and the page says so.

Boxleiter cross-check

A second estimate from your wishlist count using the Boxleiter formula, with a divergence flag when our cone disagrees with the rule of thumb.

Marketing-lever causal estimates

Causal effect estimates with confidence intervals for festival appearances, demo releases and trailer revisions. When the interval crosses zero the row says so.

Comp-set evidence

5+ real launches similar to yours by genre + tag overlap, with their actual revenue figures and how their wishlist trajectory mapped to outcome. The cone is grounded in specific games you can name.

Total Lift Attribution

Upload your Steamworks “Wishlists by day” CSV and we recover the campaign wishlists Steam’s UTM dashboard misses (~75% under-reported). You get a per-campaign cost per wishlist that includes the users utm_source lost.

Re-runnable through launch

As wishlists grow, re-run the forecast: the cone narrows and the Boxleiter check and attribution update with it. The report stays live through your launch window.

Retrospective demo: 5 indie launches

Five launches that show how wide the spread is, including one that turned 1.5K wishlists into 200K+ copies. Run your own game through the forecaster to see where it lands.

The Binding of Isaac: Rebirth
Action
Full retrospective →
Hades
Action, Indie, RPG
Full retrospective →
Dyson Sphere Program
Indie, Simulation, Strategy, Early Access
Full retrospective →
Vampire Survivors
Action, Casual, Indie, RPG
Full retrospective →
Nubby's Number Factory
Strategy
Full retrospective →

Common questions

Why $299 one-time and not a subscription?

A launch is an event. You buy the report once, run the forecast as often as you want through that launch window, and that’s it. Nothing recurs and there is nothing to cancel.

What if my game is unlike anything in your data?

The cone gets wider, and we tell you so. Calibrated coverage is a statement about the reference distribution: the interval contains the true outcome with frequency at least 1−α on average, not for every individual game. The forecast shows comp-set similarity so you can judge how much the cone applies to your game.

How is this different from rule-of-thumb formulas like Boxleiter?

Boxleiter is a single-number heuristic (revenue ≈ wishlists × ~$5). It’s a starting point, not a budget plan. We give you a cone with a P10 floor and a P90 ceiling, a Boxleiter cross-check beside it, and 5 named launches similar to yours so you can check the model against games you know. Methodology →

What does Total Lift Attribution actually do?

Steam’s UTM dashboard only counts wishlists where the user clicked your campaign link AND completed the wishlist add in the same session. Real users discover via your campaign, leave, come back later. Steam attributes those to "direct/none." Total Lift Attribution compares your “Wishlists by day” CSV against a 14-day median baseline to recover the lift Steam misses (~75% under-reported in our internal benchmarks), and reports cost per wishlist per campaign on the recovered number.

What if I don’t get value out of it?

14-day full refund if you haven’t uploaded any Steamworks data yet. 50% refundable thereafter. Email [email protected] from the same address you bought with; refund processes within 14 days. Full terms here.

Who built this?

Greg C., solo. I spent 10+ years building production ML for sports prediction, where the payout distribution is heavy-tailed and a badly calibrated interval costs real money. Steam launch revenue has the same shape, so the same calibration checks apply. Reach me at [email protected].

Free guides: the numbers behind the cone

Public-data benchmarks, no signup. Every guide states its sources and its uncertainty.

Stay in the loop

Short notes every couple of weeks: calibration runs, new features, real launches we’re testing against. No spam.