I can't believe it took me this long to read Kurt Eichenwald's long article on Steve Ballmer and the fall of Microsoft. It's a great piece full of telling detail. But I do think it's worth offering a more analytic take. There's a reason people in Ballmer's position tend to fail.
The basic issue facing Microsoft over the past ten years has been this—innovating is really hard.
The company reached a point where Office and Windows were so popular that wasn't much you could do to increase their popularity by improving the product. They continued to work on improving the product, and kept these divisions very healthy and profitable, but there simply wasn't an explosive growth opportunity left to be had because the previous successes had been so enormous. So you create a situation where the company as a whole is basically a venture capital firm. It has this huge stream of Office/Windows profits and needs to figure out how to invest those profits in exciting new products. But successful venture capitalists are really rare, and for all we know most of them are just getting lucky. The average financial returns from the venture capital sector as a whole are terrible. But Microsoft qua venture capitalist faces the additional burden that the top management of the company has to be good at running the giant existing Office/Windows businesses. It's as if you were trying to hire a tax attorney who could also perform open heart surgery.
One alternative strategy could have been to just give up. Pay huge dividends, keep focusing on incremental improvements to the core products, and basically don't worry if other firms dominate mobile and online services. But not only is that psychologically unappealing to managers, it'd be weirdly demoralizing to the staff. Windows and Office need to be able to hire talented engineers—the kind of people who are going to want to work for a company that aspires to be forever on the cutting edge, not a company that's resigned itself to operating as a boring dividend machine.
So what are you supposed to do? Obviously "kill an early promising e-reader project and bury the team working on it in your office division," "kill morale with a creepy evaluation system," and "lose billions on developing a search engine" seem in retrospect like misguided ideas. But this is genuinely hard stuff. And it's at least possible that Windows 8 will be a huge hit and people will turn in drove to buy smartphones and tablets that seemlessly integrate with the still-dominant desktop PC platform and we'll all look back on the ten-year Apple Bubble and laugh. Financial markets are betting against that, but it's not a crazy story. Anyway, read Eichenwald's piece. Laugh at Ballmer's errors. But just remember that basic reality. Innovation is really hard. Staying on top once your core business has saturated the market is really hard. Having the third-highest market capitalization in America is pretty damn impressive.