On the pitfalls of mistaking correlation with causation.

This blog post was originally intended to be focused only on Press Kits, but I realized that it applies to other stuff as well.

Say you want to increase downloads for your apps. So you select one of the most downloaded on the App Store and create similar screenshots to drive more downloads. It makes sense, right?

The problem with this approach is that you cannot be sure that these screenshots are one of the reasons the app is top-rated. Maybe it is popular despite its screenshots and with better ones, it would have even more downloads.

I am pretty certain that most of the people downloading Spotify (or any other popular app) don't care about the screenshots or description. So although Spotify's screenshots might be excellent, the high popularity doesn't guarantee that.

Similarly, with your Press Kit. Big and already popular apps will get a lot of press, regardless of whether the press kit is good one or exists at all. This means that checking out press kits for super-popular apps isn't the best idea.

To close, we can extend this idea to app features as well. If you see what looks like a fantastic onboarding in one popular app, don't rush immediately to implement something similar yourself. From this single observation, you cannot know (without perhaps seeing their internal data) if the onboarding is successful or not.

correlation != causation.