Estimates for apps are made by fitting a curve to each country-category rank chart. The scarcity of signals a data provider can collect from the app stores means Priori Data – like all other leading providers in the industry – rely on app store rank as the only reliable indicator of downloads or revenue. This implies that there is a natural limit to app level accuracy – whatever the claim from the provider may be. We believe it is important to communicate this clearly while helping our clients reach the most reliable conclusions possible through the appropriate use of data contextualization.
Apps: Our MAPE (Median Accuracy Percentage of Error) depends on the country/category combination. In Western markets our accuracy is about the best you can get. On average, all country/category combinations included, for downloads, we have a 20% margin of error. For revenue, we are around 30%.
Aggregates: Once app level estimates are aggregated (to the publisher, category, country, store level or over a longer time period etc.) the accuracy typically increases dramatically. For some markets, we observe median aggregate error levels of <5%.
For more on methodology, check out our Methodology page.