tl;dr for those who didn't read the article. This data is from LinkedIn and based on their measures of:
1. Number of job applications
2. [LinkedIn] engagement
3. employee retention.
The 1st factor one may be biased by Amazon's larger employee base, which according to Google is currently >500k.
The 3rd factor is a big one and may be heavily biased by Amazon's exploding stock growth and horrible vesting schedule.
The article calls out Amazon's frugality as appealing to workers. If this is true, then human beings too easily act against their own interests. As a consumer, I appreciate their frugality toward their workers. As a worker, I would tell them to GFYS (in fact, I did). Their frugality does not translate to higher comp.
I'm a little dismayed that someone would be willing to make use those factors and then claim it's about the best place to work. It's so obviously off point that I can't even buy that it's the result of stupidity - the blog post is consciously dishonest.
I know that that's pretty much the norm in blogs/reporting, but I'd really like to see this sort of behavior discouraged. LinkedIn in particular has a lot of practices that seem particularly skeezy to me. I'm not sure exactly what should be done about this sort of thing, but at the least I want to draw as much attention to it as possible.
It's very strange that number of applications is used in the calculation. As you suggested, it tells you more about openings, brand value, and perhaps geography than it does about how nice it is to work there.
This honestly just seems like LinkedIn trying to synthesize its data into something that will be written up in other places. A serious attempt at answering this question would use a survey. GlassDoor might be somewhere closer to knowing the answer.