In Soviet Microsoft, Sinofsky Plans You

If nothing else, Steven Sinofsky’s remaking of Windows has provided us with barrels of entertainment. Jay Greene’s thorough profile of the man and his mission describes how Sinofsky quickly rose through the ranks, earning his reputation by shipping Office on time and eventually repairing the Vista disaster. Along the way he restructured Microsoft’s work environment, alienating and even “chasing away” talented employees who wouldn’t go along, according to Greene’s interview partners. Sinofsky’s internal strategy is outlined in his 2009 book One Strategy: Organization, Planning, and Decision Making:

At the heart of the book is Sinofsky’s case for a “functional organization.” That’s management-speak for reporting lines that are structured around job functions – such as product management, development, software testing – as opposed to a “product organization,” where multi-disciplinary teams work on specific feature sets together. Sinofsky and [co-author Marco] Iansiti argue that functional organizations create clearer road maps for workers to march toward a final goal.

It’s dense stuff, but the functional organization structure championed by Sinofsky is a flashpoint for his critics. Managers beneath Sinofsky say they had greater control over product development, working across groups with engineers, product managers, and software testers. Now, they say they feel more like cogs in the machinery, marching toward a final pre-determined goal, without the authority to shift course if they believe there’s a more innovative approach to product design.

“You are told what to do now,” said a current Microsoft executive. “It puts more directional control in the hands of the leadership.”
One former senior executive referred to the approach as “Soviet central-planning.” In a blog post quoted in “One Strategy,” Sinofsky acknowledged that the organizational model is “controversial” but said it was a necessary structure to ensure workers could focus on their areas of expertise to meet product planning goals.

It’s hard to argue with a business strategy that ships well-received products on time. Vista was late and disappointing because the original ambitious plan to rebuild Windows on top of .NET had been an utter failure. Clearly someone should have been in charge to evaluate and reject infeasible plans at an early stage. On the other hand, a “product czar” who can reject bad ideas may also ignore valid criticism – why can’t we have a Start button, again?

Moreover, while Sinofsky’s “deep grasp of technology” is widely acknowledged, public relations do not appear to be among his strong points. There was the embarrassing Metro branding trainwreck, of course. I blame the dearth of Windows Store apps at least partly on an unnecessary alienation of loyal .NET developers. As for end users, confusion reigns right before the launch of Windows 8. Last week Paul Thurrott complained about an inundation with e-mails by people who don’t understand the difference between Windows RT and Windows 8. Later he wrote a long clarification on the difference between the Skydrive apps for each system. Windows 8 may succeed in the long run, but it seems to me that Microsoft has largely failed to make people understand and look forward to its new products.

2012-11-13: Rather sooner than anyone had expected, Sinofsky has left Microsoft, apparently due to his abrasive personality making too many enemies.

7 thoughts on “In Soviet Microsoft, Sinofsky Plans You”

  1. Not all his decisions are correct and when they don’t, they affect the whole company, badly. For example, he (and others) tried to shift focus from .NET to C++, and still, the majority of Windows 8 apps are written in C#…How he believe that after 10 years of using this masterpiece (C#) anyone would get back to C++ ??

    1. Do you have a source for the language use in Windows 8 apps? I haven’t seen any data but I would have expected the majority to be written in C++, given how hard MS pushed that, or else JavaScript for web / social media stuff.

        1. Thanks, interesting find. I wonder how many .NET apps will be there on release day, and if users will be able to spot differences in performance or functionality.

  2. No problem. From what i have read, i mean, from people that write Metro (sorry, i meant “modern”…) apps, the performance difference between C++/CX and C# is extremely small. The WinRT projections-magic helps with that, that’s for sure…

    1. Thanks for the link! That’s a nice demonstration of the popularity of .NET among Microsoft developers. I wonder how many actually end up putting .NET apps on the Windows Store, though.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.