The study is garbage.

The metric they're using (average tenure) is not useful for comparing employee retention across companies. E.g. a company that's growing will automatically have a lower average tenure than one that's not growing.

Now, it's understandable why they're using an inappropriate metric. It's the only thing they can even remotely reliably derive from their data set of scraped Linkedin profiles. But when your data is bad, the right thing to do is to just not do the study. Not to write a content marketinb blog post that's so blatantly incorrect that it has to be intentionally dishonest.