ASOClimb
Get Started

Methodology · Modeled figures

How revenue and download estimates are made

ASOClimb models monthly store revenue and downloads from public chart positions. Every figure is shown as a range with a point estimate and a confidence level, and is labelled Modeled wherever it appears.

What the numbers mean

Est. revenue (net). Store revenue from paid downloads, in-app purchases and subscriptions in this storefront, after the store commission (15–30%). Excludes ad revenue, web and other off-store payments, taxes and refunds the model cannot see.

Est. downloads. First-time downloads in this storefront, modeled from free and paid chart positions. Re-downloads and updates are not counted separately; the anchors the curves are fitted to do not distinguish them.

Store fees. The grossing curve is fitted to gross consumer spend. Net revenue applies a 30% store fee on the low side, 15% on the high side and 22.5% for the point estimate.

How the model works

Daily value ≈ a · rank^(−b), per store and chart (free/paid → downloads, grossing → revenue), calibrated on public anchors; category charts are calibrated against the overall chart through apps listed in both; daily values are averaged over the last 30 captured days.

The point estimate is the central curve (geometric mean scale, mid slope). It is a sorting aid inside the range, not a more precise answer.

Ranges

The range is the envelope of every curve the public anchors allow (scale and slope varied across them), widened where anchors are dated, cross-vendor or out of scope, then multiplied by a country factor. It is a heuristic band, not a statistical confidence interval: no backtest against reported revenue has been published, so ASOClimb makes no accuracy claim. For low-confidence rows the high end can be 100 times the low end or more, and the product shows that width instead of hiding it.

Confidence levels

  • medium: A storefront with its own anchor for that metric (the US for both; Great Britain and Japan for downloads only), an anchored curve, an overall-chart position, and at least 7 days in that metric's chart in the last 30. Medium is the highest level this model claims: its anchors are published vendor estimates.
  • low: Anything else: a country without its own anchor, a curve without an independent anchor (Google Play paid downloads, Google Play revenue), fewer than 7 charted days for that metric, positions only from a category chart, or public rating counts that contradict the rank signal.

Revenue gets a point estimate only after 3 days in the grossing chart. Below that the range runs from what the observed days alone are worth to that pace held for the whole month. An app never seen in a captured grossing chart gets only an upper bound.

Known weak spots

  • New apps and apps that chart only briefly: few days in the window means a wide, low-confidence range.
  • Every country except the United States: no public absolute anchor, so a wide prior scaled from the US curve.
  • Google Play revenue: the market size is a difference between two vendors’ published figures.
  • Apps with heavy off-store revenue (web subscriptions, ads): none of it is visible to a store chart.
  • Paid apps: chart rows carry no price, so price × downloads is not computed.

Update cadence

Each storefront is re-modeled at most every 12 hours from the public charts captured over the previous 30 days; a weekly history point is kept for 400 days. Estimates not refreshed for 45 days are removed.

Public calibration anchors

Model changelog

  • rank-power-law-v2: Downloads and revenue each carry their own days observed and confidence; revenue needs 3 grossing days for a point estimate; net-of-fees revenue added.
  • rank-power-law-v1: First release: rank power-law curves calibrated on public anchors, one shared days-observed count and confidence per estimate.

Estimates are modeled and provided for information only. They are not financial or investment advice.