Methodology

MRRBeam tracks App Store apps, ranks the most successful ones and explains where there is room for a new app. This page describes exactly how: which data we use, how each score is computed and what we do not claim. Everything is deterministic: the same inputs always give the same scores.

Data sources

• App metadata: name, developer, category, price, version, release and update dates, average rating and rating count, from Apple's public Search and Lookup APIs.
• Chart positions: the top free and top paid charts, overall and for all 25 App Store categories, read every day.
• Complaint themes: recent 1–2★ public reviews, grouped into themes by keyword. Only theme counts are stored, never review text.

Every value is labelled by provenance: Observed (read from the store), Estimated (a modelled range from MRRBeam's model, below) or AI-derived (analysis of observed data).

Revenue and download estimates

MRRBeam estimates monthly downloads and revenue for every app that is on a top chart or gaining ratings. The model is public, and every app page shows the working behind its numbers.

• Downloads from chart rank. Daily downloads fall off as a power law of chart position, anchored on published US thresholds (about 156,000 a day for #1 overall and 52,000 for #10). Category charts use a curve scaled to the size of the category, paid charts sit far below free ones, and other storefronts are scaled down from the US.
• Downloads from ratings. Off the charts, new ratings per day are multiplied by a typical number of downloads per rating (about 70).
• Revenue from downloads. Paid apps earn their price per download. Free apps earn a revenue per download set by how their category makes money (subscriptions, mixed, or ads and commerce) and by their subscription price when it is known.

Estimates are always ranges with a stated confidence, kept low on purpose. Use them to compare apps and spot patterns, not as anyone's real sales figures. Rankings on Top apps use observed data only.

Update schedule

Data is gathered every day for the US App Store. The top charts are captured and newly charting apps are added if they pass a discovery threshold. Each run re-reads 100 tracked apps: the best-charting half first, so chart leaders are fresh every day, then a rotation through the rest of the catalog. Every figure carries the date it was observed. Growth signals need snapshots at least 7 days apart, so they appear as history builds up.

Chart positions

“Overall” is the better of an app's positions on the top free and top paid charts. The overall charts and the charts of all 25 App Store categories are read every day, and every tracked app's current position is recorded with its date, independent of the rotation that refreshes ratings.

Success score (Top apps rankings)

The Success score answers “which apps are winning?”. It orders the Top apps rankings and the “Most successful” sort.

• Traction, 50%: rating count on a log scale; 2 million ratings is the ceiling.
• Chart position, 30%: the better of the overall and category top-chart rank (category ranks count 85%). Apps not on a chart score zero here.
• Satisfaction, 20%: average rating, shrunk toward 3.8★ by 25 pseudo-ratings so a handful of 5★ ratings cannot outrank a well-loved app.

Tiers: 75+ market leader, 55–74 established, 35–54 rising, below 35 early stage.

Opportunity Score

The Opportunity Score answers “is there room for a new app here?”. Higher means more opportunity, so a dominant, well-loved app usually scores lower than a popular app with unhappy users.

• Demand, 20%: Is there proven, paying demand? Estimated revenue and downloads (observed proxies when no estimates).
• Competition, 22%: How open is the market? Higher = less entrenched incumbents and a thinner category.
• Monetization, 12%: Do users pay, and how much? Price level, pricing model and estimated revenue per download.
• User pain, 26%: Are users unhappy in ways a new entrant could fix? Complaint clusters and rating weakness.
• Momentum, 20%: Is demand accelerating? Revenue/download growth, rank movement and category momentum.

Each dimension is a weighted average of normalized features. When a feature is missing (for example revenue estimates), a declared observed proxy is used or the dimension is scored from what is available, and confidence drops accordingly.

What we do not claim

• Revenue and downloads are never fabricated for real apps.
• Keyword demand is a proxy from the apps that rank for a keyword; Apple publishes no search volume.
• Complaint themes come from a fixed keyword dictionary and recent reviews only.

Questions