Microsoft's innovation disconnect

A few weeks ago, a former Microsoft colleague posted a link on Facebook to an article by Ina Freid that discussed Microsoft's recent reshuffle of its Mediaroom (formerly Microsoft IPTV), Media Center and Zune software divisions. I worked at Mediaroom / IPTV for around three years, and though the division doesn't get the kind of press that the revenue-critical Windows and Office divisions get, based on how it frequently pops up at various technology events, Microsoft clearly considers it important. Mediaroom, along with Media Center and XBOX, form the core of Microsoft TV efforts, which if you believe the organizational principles layed out by Ray Ozzie at technology conferences, forms one of the "three screens" that underpin its cross-device efforts throughout the company.

In Fried's article (which was derived from a post by Mary Jo Foley), it appeared that the move undid a previous reorganization that placed those three divisions into one group. I can understand the rationale for attempting to merge them together, organizationally-speaking. Microsoft's overall approach to televised entertainment should be cohesive and interoperable, both with each other and with other nodes within the larger Microsoft product catalog. The reality, however, is that trying to merge those three divisions was probably like trying to splice a zebra, an elephant and a parrot. Had Microsoft wanted to make those products work consistently and seamlessly with one another, they needed to start much further back in their respective evolutions than 11 months ago.

Should it be that difficult to merge three product divisions that, in theory, are a part of the same broad product category within the same company? In two words, "hell" and "no." Understanding why it is, in fact, so difficult relates to the peculiar way Microsoft organizes internally a company that has, as of June, 2009, 92,000 employees.

Back in the late 1990s, management books discussing the secrets of Microsoft's management principles were somewhat common. Back then, Microsoft was the undisputed (albeit much resented) king of the computing world. Apple looked like it might go out of business (an event Microsoft tried to prevent with a $100 million invest), and Microsoft's antitrust troubles really got rolling during that period (which largely explains its Apple investment). A core organizational principle was the cultivation of internal "mini-companies" that were supposed to compete with each other as intensely as outside companies competed with each other and Microsoft.

In theory, this "competition" would inject the kind of dynamism that is hard to achieve in a large company. Big companies have a difficult time achieving the vigor of smaller competitors. Über-blogger Robert Scoble once described working at Microsoft as like being an ant on the side of a very large anthill. Truth be told, the analogy would apply to any large company. You work on large teams where it is hard to be heard, but it is easier to "hide," the benefits are comfortable, and if the company is really successful, money is never an object. Those realities aren't conducive to cultivation of the intensity that drives smaller companies.

Mini-company organization helps when entering new markets early, as it facilitates experimentation in the various ways of achieving the same technology goal. New technology can always be implemented a hundred different ways. In early days, none is the obvious better option. If you have the money (and Microsoft clearly does and did), it can make sense to try multiple paths to see what works best. Often, multiple paths prove productive in their own way, resulting in a merged product that is better than if only one of the paths had been pursued. The Office team was a notable practitioner of this approach, starting various competing projects to attack the problems of document automation from different directions, only to merge them later into a common product.

This model may have worked fairly well in the past, at a time when Microsoft was involved in fewer market segments and it had the guiding hand of a powerful founder to beat warring interests into line behind a common goal. However, that founder has moved on to much more important work, and Microsoft is now involved in far more products than when Gates had his hand on the keel.

To understand the importance of a single person with both good ideas and the power to force people to implement them, think Steve Jobs. Apple wasn't the same company when he wasn't around. Clearly, Steve Jobs has more good ideas than most people have in several lifetimes. Besides founding Apple, one of his side projects was Pixar, a company that redefined the meaning of animated entertainment. Bringing him back was critical to Apple's success, in my opinion, because Apple wasn't going to find someone easily with the good ideas and natural power of a Steve Jobs. Founders of successful companies have already demonstrated their vision (otherwise their company wouldn't be a success), and as founder, they have power of a sort that is very difficult to duplicate.

But, strictly speaking, the visionary doesn't have to be a founder. Microsoft's recent success with revamping its Windows product line with the well-received release of Windows 7 owes a huge debt to Julie Larson-Greene and the person who recognized her insight, Steven Sinofsky. Sinofsky gave her the power to impose a new UI on Office (resulting in the "ribbon" UI concept inaugurated with Office 2007), and when he moved on to a position in charge of the Windows team, he made sure she had the power to work her same magic there. Ms. Larson-Green had the ability to force disparate teams to hew close to a centrally-controlled user interface concept. Granted, Windows 7 isn't merely a UI update, and if you listen to Mark Russinovich, the stuff under the covers was rather important to making the UI more responsive. Making the UI of central importance, however, helps to guide the kind of improvements upon which Microsoft's "internals" experts focused. Giving someone with good ideas the power to implement them was critical to its success.

As a company grows, the presence of someone with both vision and power grows in importance. Big companies are, in many ways, like political constituencies in a country. Each group has its own interests, and each works hard to attract as much revenue as possible from the center. Combine that with the fact that big, well-funded companies are particularly attractive to people who like to be in charge, and meetings can end up every bit as fractious as the floor of the US Congress.

At software companies, code ownership is a critical component of power. In the absence of a guiding hand to ensure that development follows a path directly beneficial to the corporate whole, this has a tendency to create huge duplication of effort. I know of at least three Digital Video Recorder (DVR) implementations within Microsoft (I'm certain there are more), which to my mind is a direct outgrowth of a desire to OWN and CONTROL code. The longer teams are allowed to "compete" as mini-companies without any central vision imposed upon them, the more incompatible their creations become.

In fact, incompatibilities can actually end up being WORSE in a company organized around mini-companies. Outside companies, paradoxically enough, have some incentive to make things compatible with competitors (unless they are overwhelmingly dominant) if for no other reason than they are likely to operate in environments where products from different vendors must coexist. "Mini-companies" within a larger corporation are safely plugged into corporate revenue streams, and their interests align with minimizing opportunities for competing divisions to attract mindshare by leveraging parts of their infrastructure. Politics in big companies is often zero-sum.

This, in a nutshell, explains why Microsoft might have run into a few walls as it tried to make Media Center, Mediaroom and the Zune software platform (that latter of which, based on the XBOX team's past behavior, is least likely to be consistent with other technology at the company) seamlessly consistent. Microsoft was trying to impose consistency far too late in the development process to think that it would be achieved in a mere 11 months. Microsoft needed someone with the power to impose a TV-oriented vision across its various television-related divisions five or six years ago (at least).

The past 10 years clearly shows that Microsoft has a disconnect between good ideas and implementation. Microsoft was in the mobile phone space years before any of its software company competitors, yet Windows Mobile is widely perceived as an also-ran to much more interesting and dynamic iPhone and Android (though as I noted previously, I don't consider the situation unsalvageable). It was pushing the eReader concept back in 2000, but it was Amazon who brought it to life. Gates was pushing the Tablet concept years ago, and for whatever reason, they weren't able to make it simple enough (or cheap enough) for the general public to embrace (we'll see if Apple does any better). At Mediaroom, I ran across lots of very forward-looking technology that never saw the light of day for reasons that had more to do with politics than software design.

I think Microsoft's isn't doing a good enough job of imposing a sensible vision across product categories (as a general principle which has lots of exceptions; it clearly doesn't explain Windows Mobile's problems). That, I think, is the result of either not identifying (or misidentifying) the people with good ideas, and / or failing to give them power early on.

I don't have a ready-made solution to the problem. Microsoft has, however, solved it in the aging divisions that are critical to the survival of the company - Windows and Office. They need to do something similar in their newer product lines. Windows and Office concentrate executive minds in ways other divisions don't. Perhaps that is the reason other divisions haven't grown into the sources of alternate revenue Microsoft had hoped they would become.

Am I saying that Microsoft should become Apple? Definitely not. I think XBOX has cloned too much of the Apple business model, and jettisoned too much of the core identity that makes a Microsoft product a Microsoft product. Control should be looser than it is at Apple, but that doesn't mean Somalia is a vision of Capitalist perfection (just to make a weird analogy). Recognizing the power of visionaries doesn't require an Apple-style top-down control any more than Capitalism requires the absence of government.

It's a balancing act. Apple has gone too far down the path of centralized control, which works for an initial product bang but has problems as the market matures and includes new competitors (in my opinion...but that's a debate for another day). Microsoft, I think, has moved too far away from centralized control, at least internally. Historically, that wasn't how Microsoft was organized, and the central vision imposed by its founder imposed order on his fractious corporate citizens. That central influence is replaceable, I believe, provided Microsoft manages to find more of the kind of people who turned around the Windows and Office flagships.

Granted, that's not as easy to do as relying on the proven skills of a founder, but is achievable...as Microsoft has demonstrated in its success with Windows 7.

Footnote: After I wrote this article, Dick Brass, a person who had a lot more power at Microsoft than I ever did, wrote an op-ed for the New York Times that appeared last Thursday. It's well worth a read, and offers more specifics than I'm willing to give.