How did you deduce this?
HTML is fast to develop relative to native apps therefore if the app is ideally used infrequently, a native experience isn't worthwhile. There are exceptions, of course.
The last option is just a default last option.
This is why this article is BS, because most of the data used in it is extremely irrelevant for the Android ecosystem right now. This article should've been written in 2011. Not at the end of 2012. The Android ecosystem has grown and improved a lot since then.
Here's an article from a few weeks ago supporting my claim: http://m.guardiannews.com/technology/appsblog/2012/dec/04/io...
Edit: to clarify, Android is doing better, but it is still no match for iOS.
I know how to write HTML as well as Android XML or Dalvik (Java) code. I do not see where pounding out HTML code is faster than say pounding out Android XML code.
The only thing fast about it is theoretically you can use it on both Android and iOS. Most companies go native. Facebook tried doing this before running into massive problems and going native.
I also do not see what the frequency of app usage has to do with anything. I might have two app ideas - one used once a month by 60 million people, another used daily by 2 million people. If they are ad based, they will both probably bring in the same amount of money. Why should I go native on the 2 million app but not the 60 million app?
Platform independence was the point.
If people use the app daily, it's worth going native for the better performance.
Your example about 60m vs 2m is intentionally engineered to be an exception and I said there are exceptions.