Earlier quoted context omitted.
It tests every app, including this one. Think about what really happened here. Edit: Apparently my downvoters can’t think about what really happened here so I’ll explain it. In App Review this app worked fine. Oh, the keyboard was likely lame and not useful, but the scam screens were no where to be seen. Then the app is approved and placed on the store. Now the scam screens appear. It’s trivial to do, is done all the…
I wonder if it is trivially hard to detect or not. For example, if the app reviewer’s touch interaction with an app were recorded and the resulting screens diffed with the same pattern after publishing, there should be no changes. This is a method used as part of some UI testing with selenium. There could be apps this does not work for, such as when content changes or conversion type UI is shown. But I imagine there…
There are lots of valid reasons for apps to change based on outside events (location, time, online content). For example apps displaying weather forecast or current news.