San Francisco-based Jeff Scott is the founder of 148Apps.com, which was recently acquired by PocketGamer.biz publisher Steel Media.

You might have noticed that there was a bit of a kerfuffle last week when Apple went toe-to-toe with analytics specialist App Annie.

The two parties managed to clash over growth in the amount of revenue developers make from the App Store – App Annie's figure of 12.9 percent antagonising Apple to the extent that took to the web to openly contest the firm's numbers.

The summary is, everyone was right. Both Apple and App Annie were looking at the same data from different angles. Statistics can be fun that way.

But the issue raised the question of just how App Annie gathers its numbers. Where do the AppAnnie Intelligence numbers come from, and what exactly does App Annie do?

I spoke with Oliver Lo, VP of Marketing for App Annie about what they do and how the company does it.

Stat attack

App Annie is, at its heart, a statistics company.

The firm offers a sales analytics tool for developers that both gathers their sales numbers and processes them. Indeed, App Annie's analytics tool is used by over 150,000 apps and, in particular, over 40 percent of the top 100 publishers by revenue.

Once a developer connects their App Annie account with their iTunes Connect account, the sales data is automatically downloaded and processed.

While App Annie keeps the data from that tool completely anonymous, it does combine all the download and revenue data it has access to together to build a global model of sales.

Lo comments, "All of the data we collected, the download and revenue data is completely confidential. We take the anonymised data points that we get and we correlate them with the public rankings from Apple and Google."

That data is then correlated with the public ranking data in the top app lists to place sales figures

with ranks.

Filling in the blanks

"Imagine a Y axis that represents downloads and X axis that represents rankings," adds Lo.

"Essentially we create a model of those two data points, correlating them and growing a distribution of those points through advanced statistics.

"Through that statistical correlation we are able to generate market estimates for the downloads and the revenue of the entire App Store ecosystem."

That means that, over time, App Annie has been able to tune its statistical calculations to help fill in the gaps for any hard data the company doesn't have. This helps the company produce a very accurate picture of sales and in turn, overall revenue for the entire App Store.

That accuracy is checked regularly by current and potential customers by checking the estimates with the hard data from download reports. It's this App Store revenue estimate that serves as the data from which App Annie Intelligence is derived.

AppAnnie sells that data product as an enterprise-level product to app development companies that want hard data on the App Store ecosystem.

This data could be very valuable to app developers looking to expand or new developers looking to enter the market, but it's considered a very premium product and priced accordingly.

Indeed, if the App Annie-Apple debacle proves anything, it's just how valuable – and potentially contentious – such numbers are.
Thanks to Oliver for his time.