Why We Don't Forecast App Store Rank or Install Volume

August 18th, 2015

by David Bell

CEO at Gummicube, Inc.

A common request from our mobile marketing, agency and app publishing clients is for a forecasted app store rank and/or expected install volume increase (or decrease) metric. This desire for data is understandable and frequent enough that I felt a short post here could add some insight into how we view providing these types of metrics. We have opted against forecasting the increase of rank or install volume inside our software. There were several reasons for this decision, but the most important one was that this kind of representation is almost always disingenuous. It caters to the desire of a marketer to be data driven but does so in a way that doesn’t reflect the way organic marketing really works. We approach ASO by looking at the whole funnel. The top of the funnel focuses on keyword ranking and visibility for the application. We look at things like keyword volume, relevance and also phrase matching in this process with a goal of giving apps the biggest possible footprint within search. Further down the funnel is conversion — once you have visibility, users have to choose to click your app vs. competitors who also appear in search results.

The conversion rate from app to app on a keyword is almost never the same.

Two different apps can have dramatically different results on the same keyword in similar positions. We incorporate A/B testing and work on icons and screenshots in our ASO process because this helps improve conversion on your search visibility in a very data driven way. In the world of SEO, which is a much more mature market, the top analytics software available also doesn’t provide this kind of projection for similar reasons. To illustrate, this would be like forecasting the number of users who would convert on a website based on the traffic that the keyword planner indicates a particular keyword has. It doesn’t take into account the actual conversion rate — which is different for every product. As a philosophy, Gummicube always tries to provide accurate data and representations of how the process works. We usually avoid pursuing features that in our best determination may misrepresent/mislead partners — even if they may provide short term satisfaction.