Vendor lock-in, non-interoperable software, most data stored in formats that are not easily machine-readable (images, pdfs, etc), and most computer users unaware of the basics of computing.

I didn't notice how they measured productivity.

If it was simply the company is more profitable after an IT investment then there are possibly confounding factors.

If it is more complex than that, I would like to read the very lengthy research investigating all those confounding factors.

So you looked up the paper cited in the article, right? Since you understand there are confounding factors right? Since you're a competent and knowlegeable human, right?