Methodology
How we score app ideas
A niche is one job people go to the stores for — scan a receipt, identify a plant, track a habit — measured in one store and one country. Every score on SideRadar is built from what those stores show publicly, the same way for every niche, every night.
01Where the data comes from
Public App Store and Google Play data: search results and rankings, ratings and how fast they grow, written reviews, in-app purchase prices, install bands on Google Play, and when each app launched and last shipped an update.
Niches are found by our own discovery: the searches the stores suggest, the top charts, the neighbours of apps that already rank, and the words people use in reviews. Nothing is bought from a panel or estimated from ad spend.
Every night the readings are refreshed, every niche is scored again and new niches are added, across the App Store and Google Play and more than 170 storefronts.
02Six dimensions
Each niche gets six scores from 0 to 100. The headline score weighs them; you can change the weights on the list.
- Trend
- How fast the leaders are gaining ratings, how many apps entered the niche recently, how high a young app can climb, and how often the leaders ship.
- LTV
- The genre's revenue profile, how often the app is used again, and evidence that people pay.
- Content ease
- A high score means the product does not have to keep producing content.
- Ops ease
- A high score means light moderation, support and compliance load, and modest compute cost.
- Conversion
- How transactional and how urgent the search intent is. A ranking between niches, not a percentage.
- Rankability
- The gap between demand and ranking difficulty. Can you actually take the traffic?
03What holds a score back
Some niches look strong on paper for the wrong reason. Rather than hide them, the score is held back and says why:
- The term is one app's name. When nearly every review among the apps carrying a term belongs to one of them, people are searching for that app, not the job. Held back by a fifth.
- Almost no reviews yet. When the most-reviewed app in the top ten has fewer than 200, demand is unproven. Held back by a tenth.
- A licensed business. Lending, cash advances, gambling and pharmacy apps need a licence in most markets. Held back by a quarter.
Brand names and terms that are not a product at all are removed from the list entirely: a high score for a niche you cannot enter is worse than no score.
04One job, many countries
The same job is measured in every country where it ranks, against that country's own rivals. The list shows it once: from the United States when it was measured there, otherwise from the country with the median score — never the best one, which on a small market is mostly luck. Filter by country to see each country's own reading.
05Movement, money and complaints
Movement comes from a daily record of each niche's top ten: who entered, who left, whose ratings are growing. Time windows from 7 days to 6 months open only once the record reaches back that far, so a window never answers with numbers it did not measure.
Revenue is an estimate and shown as one: a range, with its basis and its assumptions, never a lone number.
Complaints are repeated phrases from low-star reviews of the leading apps, each with a sentence someone actually wrote, so a theme can be checked rather than trusted.
06What the numbers cannot tell you
How hard a product is to build, how much content it needs forever, what a support inbox will look like — the stores do not publish these, and the scores for content and operations are read from the kind of app, not measured. A score is where to look first, not a decision.