TrackGap Research

How to read a competitor's App Store rating — the average hides everything

A 4.5-star average can sit on top of a pile of one-star reviews. We read 18,145 reviews, classify every negative one by complaint type, and publish what we find. The company-specific versions go to the teams; the category versions are public.

Research No. 1 · 17 Sep 2026

We classified 3,300 App Store complaints across 21 field-service apps. Crashes are the #1 complaint in 11 of them.

The full complaint-rate ranking, the category breakdown behind it, and the finding that matters most: the median app here shows 4.58 stars while more than half its recent written reviews are 1–2 stars.

3,300 complaints classified21 field-service appscomplaint ratescategory breakdown
Research No. 2 · 17 Sep 2026

Billing complaints are 6.6% of what users complain about. They are also the complaint where the largest share of reviews are one star (96.3%).

We isolated all 354 subscription and cancellation complaints in the sample. 96.3% are one star, the median runs 83 words — 1.7× a crash report — and the dominant word is refund, not cancel.

354 complaints96.3% one-starrefundsbilling
Free · 3 categories

What users complain about in a category — before you build in it.

One free page per category we track: the median complaint rate, the three complaints almost everyone there gets hit by, and how the sub-categories under one label differ. For people who haven't shipped yet — there is no row for your app in it, and that's the point.

Read the category summaries →
Live · 46 apps

Per-app complaint snapshots

A public page for each of the 46 apps we track: complaint rate, rank inside its own track, and where the complaints concentrate — with the private detail held back.

Browse the snapshots →
Research No. 3 · 17 Sep 2026

Two tracks, one method: why a 7.4% median and a 3.8% median must never share one ranking.

The same complaint-rate method run over two groups of apps that serve completely different people — and the structural reason their complaint profiles diverge. Includes the full category-by-category comparison.

46 apps3 tracksbenchmarkingmethod
Research No. 4 · In progress

When users name a competitor, what feature do they say won them over?

Rare in the data, and the most expensive sentences in it. Every review in our set that names a rival, and the specific feature each one cited.

competitor losschurn signals
How to read these. Sample sizes range from 26 to 100 reviews per app, so treat the tiers as meaningful and the exact ordering inside a tier as noise. We publish complaint counts and recurring functional phrases only — never review text, usernames or avatars.

Method in one paragraph

We pull the public customer-review feed for each app, take the most recent batch of written reviews, and classify every 1–2 star review into a fixed taxonomy of complaint types: stability, missing features, pricing, subscription and billing, support responsiveness, performance, and competitive loss. Classification is rule-based and then manually reviewed. Unclassified complaints are excluded rather than forced into a bucket, and we report that coverage number alongside every result. The complete walk-through — sources, storage, classification, limits — has its own page: how this is made.

Want this run on your app?

Free, no signup. Send us the app and we'll classify your reviews, then send back the breakdown plus a comparison against three named competitors using the same method in the same week.

How it works →