We once purchased estimated telecom subscriber data for many regions, then when we started to analyze it (growth rates, variances, etc) it turned out that the d

We once purchased estimated telecom subscriber data for many regions, then when we started to analyze it (growth rates, variances, etc) it turned out that the data was not as granular as expected - for example, growth rates were clearly estimated for several regions at once (i. e. all regions in the same macroregion had the same growth rate) So make sure you're buying real data, not researcher's assumptions which you can assume yourself.