Most of what Microsoft does is indeed bafflingly bad. With a few exceptions. The baseline for software from Microsoft appears to be slow, bad UX and very buggy. And it isn't like this is some image thing; Microsoft products are always worse than I could remember when I'm confronted with them after some time away from them.
Developers at Microsoft are obviously not rewarded for quality. You have to assume that this is because managers and leaders in Microsoft are not rewarded for quality. You would think that a company that has deep pockets would be in a great position to do more ground-up re-implementations. And to do so with quality, performance and correctness as the main focus.
For instance the office suite. The last 20 or so years have taught me that an Office suite can be a lot simpler and it will actually work better if it is simpler. Just in the last 5 years I have observed three different companies where people routinely perform most of their writing and editing in other tools and then insert what they have written in Word. Because it is far better than creating the content in Word itself. At my current consulting gig a lot of people write things in Google Docs and then import them into Word documents to produce the official versions of documents.
Word is a mess. It is packed with too many features you will never use. Those have a cost because they take up screen space, and make the features you do care about harder to find and use. Word constantly distracts you because it misbehaves and you have to somehow try to deal with its quirks and interruptions. It is slow, complex and resource intensive.
Word is objectively not a very good piece of software. I have never met anyone who loves it. Who feels that Word makes them more productive than any alternatives. It is software you have to cope with. Software that must be tolerated. Or not.
I do not understand why Microsoft, with its deep pockets, has made no attempt to reinvent, for instance, Word, to create a word processor from scratch. With focus on quality, correctness, performance, usability, and perhaps most importantly: easy extensibility.
They could draw some inspiration from Visual Studio Code. There are many things that are wrong with VS Code, but they got a few things right. The most important being that unlike other IDEs it is essentially just a skeletal platform that derives its value from extensions. Third party extensions. This means that VS Code can be adapted to fit your individual needs, or more importantly, the needs to segments of users. It means that people who want to make tools can build on VS Code rather than having to do a lot of work orthogonal to their goal to create tooling.
Yes, you can probably wrangle special functionality into Word. But nobody does. Not at any meaningful scale.
Word is rooted in a world that existed before many of you were born. A world that is long gone. There has been decades of technology evolution. If you were to develop a word processor today, you would be starting from a point that is completely different.
And let's not get started on Azure. I have to deal with it about every two years. And every two years I try to approach it with an open mind and with optimism. Surely they have fixed things now? I am always disappointed. Things look slick on the surface, but then you start to use them, and you are confronted with systems that are slow, slow, slow, ugly and buggy. AWS is certainly not the belle of the ball. Its constant complexity and the awkwardness and just overall badness of the tooling makes me limit how much of it I make myself dependent on AWS services.
But at least AWS isn't as bad as Azure.
I don't get why Microsoft can't seem to invest in quality. Yes, I get all the arguments that it just needs to be good enough for their customers to keep using them, but surely, at some point it has to hurt your pride.
If I were in Nadella's shoes I would invest heavily in quality. In stripping things down. In starting over. In making sure that I understand the required cultural change required to make products that are objectively speaking, good. If not great. And perhaps that requires getting rid of a lot of long-time leaders that just can't change gears. Perhaps it requires creating teams that are isolated to a greater degree from other teams so they don't drag each other down.