With the recent announcement of the (probably forced) retirement of Microsoft CEO Steve Ballmer, respectable publications are rolling out another series of “Where Did Microsoft Go Wrong?” pieces that almost all trace the same narrative arc:
In 2000, Ballmer inherited a software juggernaut so powerful that it was known to many as the “Evil Empire”. How could he possibly have mismanaged it to the point that, despite its continuing record of raking in cash, it is now almost the industry’s comedic afterthought?
We like our classic tragedies, where one man’s hubris brings down everyone around him. We like stories where there’s someone we can point at and blame, especially when that villain is easy to dislike. And by all accounts Ballmer wasn’t a good leader for Microsoft; neither great himself, nor the kind of person to inspire those under him to greatness.
Blaming Ballmer for the woes of Microsoft, though, misses the fact that every problem the company is experiencing today was written into its DNA in the 1980s.
In 1980, IBM were looking for an operating system for the IBM PC, their entry into the nascent personal computing market. When attempts to license CP/M from Digital Research stalled, IBM went to Microsoft. Microsoft didn't have an operating system so they licensed a CP/M work-alike originally called QDOS (literally “Quick and Dirty Operating System”), filed down some of the more jagged corners and licensed it on to IBM.
Not long after, seeing the success of the IBM PC, a company called Compaq realised that there was a market for “PC Compatible” computers. The only two parts of the PC that weren't off-the-shelf hardware were IBM's proprietary BIOS (which Compaq engineers reverse-engineered), and the PC-DOS operating system, which Microsoft was happy to start selling to third parties as MS-DOS.
DOS wasn't the best PC operating system. It wasn't even a particularly good one. All that mattered is that it was the operating system on the IBM PC, and then on the “PC Compatible” clones.
Microsoft did not make its mark as a builder of great things, but as a very successful bundler of good-enough things.
Microsoft built its empire making products that complemented DOS, and that would eventually become symbiotic with DOS so that each would ensure the others dominance. Microsoft then used its monopoly position to exert absolute control over the channel that delivered those products to customers. I’ve heard a first-hand accounts of Microsoft’s ruthless dealings with OEMs in the DOS era, and they’re even worse than you think.
Eventually by the mid-90s an argument could be made that Microsoft was making the best PC software in its class. Not because it had suddenly found the ability to develop cool, innovative products, but because everyone else was dead. This was the ‘golden era’ that most modern-day pundits seem to be hearkening back to: the two great serpents of Windows and Office, each with its jaws eternally locked on the other’s tail; Microsoft the victorious monopolist.
Still, outside that monopoly, there is a clear distinction between Microsoft products that can trace their market dominance back to DOS, and those that can't. The former make money, the latter lose it. Lacklustre success online and in mobile shows just how quickly the brand loses its attractiveness once compatibility and integration with Microsoft on the desktop stops being a point of distinction.
Microsoft’s losses in the tablet market must be particularly galling. They spent the decade before the iPad trying to make a compelling tablet product, only to have the whole thing taken away from them overnight. It’s also no coincidence that the first thing Microsoft fell back on to try to fight the iPad was promises of delivering the “full Windows experience”, complete with Office.
The Xbox is an interesting, possibly the only interesting exception. Originally one of the big selling points of the console was that the first generation was just a PC inside, and game developers could use familiar DirectX APIs to program to it. Realistically though, this didn’t have nearly as much influence on the success of the console as Microsoft’s purchase of a small gaming studio called Bungie, redirecting their game Halo from a planned simultaneous Macintosh/PC release to an Xbox exclusive.
Even Microsoft’s attempt to use their PC dominance to muscle in on the Internet failed; ironically not thanks to the US Department of Justice and its legally successful but practically ineffectual antitrust suit, but thanks to that unlikeliest of heroes, America Online. For more than five years AOL kept Netscape alive almost entirely as an escape hatch in case licensing Internet Explorer became too expensive, long enough for the Mozilla project to survive the Netscape 6 disaster, bring out Firefox and re-ignite the browser wars.
Then the 2000s happened. The Internet led us into the post-PC era, breaking DOS’s twenty-year grip. New generations of PC hardware no longer provided the noticeable real-world improvements in things people actually wanted to do (besides play games, which is why the 2000s also saw the final victory of the consoles). And with Windows XP, Microsoft had finally produced an operating system that was stable and functional enough that it could compete meaningfully with its own successor.
The market for new PCs and for new versions of Windows both went into steep decline, and Microsoft went along with it.
Software isn’t an industry where the monster company selling the last generation’s product gets to stay being the monster for the next generation. It’s the industry where a thousand hungry small companies are waiting for a shift in the market that will allow them to slay the monster, carve it up and eat it for breakfast.
Apple is the obvious outlier here. Or maybe it isn’t: that’s a topic for an entirely different thousand words. Either way, if your plan to save your business is “First, hire Steve Jobs”, I’ve got some bad news for you.
This goes doubly when the monster in question is a company with no record to speak of for building products that succeeded in their own right, just in bundling together stuff that was “good enough” to keep customers locked in to the orbit of a cold, dead, thirty-year old command line operating system.
Maybe Ballmer being CEO didn't help, but I find it hard to think what would have saved Microsoft. Microsoft was the ship that DOS built, and a belated mutiny isn’t going to stop it from sinking.