The goal of our statistical models is to obtain the best possible understanding of the download and revenue volume associated with any app store top chart across the different categories and countries in our database.
To that end, we:
1. Plot store ranks against actual download / revenue data on a daily basis
2. Use statistical analysis to create estimate models
3. Estimate models then predict downloads / revenues for every rank
That means that for apps that spend a lot of time outside of the ranks, our models are less effective, and the resulting estimates less accurate.
Our models are highly dynamic and improve over time as we continue to obtain more data partners and leverage our understanding of the app stores. When updates to our statistical models are provided, we do a full re-run of the historical data base to ensure consistency and quality throughout the platform.
For more on methodology, check out our Methodology page.