I’m not going to have a lot to say on the reorg for a few days as I want to let it sink in. Besides, my analysis in the previous two (pre-announce) blog entries hit pretty close to home. But I have to say my first reaction is “Wow!”
For the twenty years that I’ve been paying attention to Microsoft’s structure and management philosophy, often from the inside, it has been focused on product groups. The product groups had both engineering and marketing responsibilities. They were organized around clusters of products that were targeted at particular customer segments. Before that they were organized around two broad groupings, Apps and Systems; A notion whose tentacles continued to influence the product groups through the last decade. Microsoft moved products between groupings, moved the groupings around, shifted responsibilities between the product marketing and field subsidiary marketing groups, and made other modest (sometimes significant) changes. The 2005 reorg was the most dramatic I’d witnessed but retained the essentials other than creating President’s with greater authority over far more independent business units (nee, product groups).
From where Microsoft has been the last twenty years there are three directions one could have taken. The first would be to break up the company, a direction I don’t believe the leadership has ever given serious consideration except under duress (i.e., when it looked like the courts might force it). The second would be to shuffle things around once again to refocus the company, but maintain the basic philosophical underpinnings of product groups. The third would be to swing completely away from the idea of independent customer-segment focused product groups to a single integrated company. One could look at the 2005 reorg and see it on a path to the next step being breaking up the company. Well, for this reorg they went in the opposite direction. In fact, pretty much as far in the opposite direction as is possible.
This reorg is a risky proposition. Microsoft’s leadership knows how to navigate a product group-focused world. Much more modest changes, such as when marketing programs responsibility were moved from the product groups to the field subsidiaries (e.g., Germany) took over a year to figure out how to actually make them work smoothly. So conceptually the reorg is brilliant. In practice it could lead to some very painful ball-dropping mistakes in its first year.
Besides the structure itself I want to point out how much of a refresh of the senior leadership has taken place the last few years. Terry and Julie are long-time Microsoft veterans who have only recently broken into the top-tier of the management structure. Qi and Tony were external hires who came in at the top-tier, Tony very recently as part of the Skype acquisition. Tami is at this point a veteran of the company, but again is new blood at the Senior Leadership Team (SLT) level. Satya is another long-time veteran who is still relatively new at the SLT level. I left Microsoft at the end of 2010, and almost the entire SLT has turned over since my departure.
I’m pretty excited about the new SLT. It’s new blood, new dynamics, lots of experience both internally and externally. Some of the leaders are proven in roles of the scope they have, while others are stepping up to the next level in their career development and have a lot to prove. All will be challenged by the radically new (for Microsoft) structure. It’s going to be fun to watch.
As far as the engineering leaders go let me give one personal vote of confidence. If I were still at Microsoft I’d happily work for any of them.
Hal's (Im)Perfect Vision 
Here is my quite selfish concern: I work for an ISV that develops, among other things, desktop Windows applications. Over the last 10 years, I feel like Microsoft has decided it doesn’t care about ISV’s anymore; it cares far more about enterprise developers, which has more growth potential for sales of lucrative server products and Visual Studio/MSDN subscriptions. .NET and Azure are primarily enterprise development tools, if not exclusively. With Visual Studio being put under a division with “Enterprise” in the name, my fear is that that trend will continue unabated. All of the focus of the development tools group is going to be on enterprise developers. I wonder if you have a take on that.
Visual Studio is in exactly the place it has been for twenty years. They may have renamed STB to Enterprise, but little else has changed in terms of the collection of products and its target customer base. And a huge part of the energy the last few years has been away from the higher-value SKUs and towards free or low-cost offerings to enable Windows and Windows Phone development. So I don’t see the organization structure making Visual Studio any less appropriate for ISVs.
The question about ISVs is who has abandoned who. Microsoft watched as ISVs ran to JAVA or Open Source, got gobbled up (Oracle absorbed most of its previous strategic partners in the ERP/CRM space), or switched their desktop focus to the web. Microsoft-exclusive ISVs, a huge force in the 90s, are a small niche these days. At some point Microsoft decided it was better to switch than fight and has been working diligently to make sure multi-platform app developers, and particularly ISVs, can target the Windows platform without having to use a totally proprietary stack. The move to support HTML/JaveScript as native technology for writing Metro apps is an example of this. This isn’t a new trend but it has been accelerating. For example, I started the effort to add a JDBC driver to SQL Server back in 2000 because lots of the newer ISVs were adopting J2EE app servers and they were having trouble achieving quality access to SQL Server. Many other efforts have come under the “Breadth Developer” heading over the years.
Most apps in the Windows/Windows Phone stores are from ISVs, so its hard to see how Microsoft has abandoned that market. But the market and way one engages with it have certainly changed. I think that’s more driven by industry trends far more than any desire from Microsoft to move away from ISVs. Just the opposite. Microsoft execs have long despaired the shrinking of the ISV base and looked for ways to revitalize it.
Hal, it is interesting you felt this way about ISVs. It isn’t true though. The people who fled were refugees because MS’s products were no longer competitive.
1) vb6 got killed, and edit and continue didn’t come back to .NET for many years. In the meantime, scripting languages gave near instantaneous response and turnaround.
2) deploying software never got easier. OneClick got close, while Silverlight flew too close to the sun, and melted.
3) writing good looking apps in Windows is still too hard compared to html+css. If you want to tweak anything in the XAML based technologies, you’ll need to do a rebuild while hitting F5 in the browser gave immediate feedback.
MS should hire a couple of rails or php developers on a small project and do some competitive research and see how productive these people are, then go back think about what features they need to work on.
Chui,
I haven’t worked as an ISV, but I have worked for companies where our software was part of a product and/or service we sold.
I developed with VB1 – 6. It needed to die. It couldn’t interoperate with code written in other languages unless the other languages went to extra trouble to support the special VB COM interface. Yeah, I missed edit and continue, but then I remember all the stuff I didn’t have to write because of the .NET BCL and I’m inclined to consider it a fair trade off.
As far as deploying is concerned, I don’t miss DLL Hell one bit. I remember how at one company, our software wouldn’t run on one of our QA machines and it delayed the release to our customers by several days while I tracked down a DLL that got replaced with an incompatible version on the QA machine by some other software.
As far as tweaking stuff in XAML, yes, there is definitely room for improvement there, but that true for any XAML developer, ISV or not.
Bob, it took HTML/JS/CSS 20 years to get to where we are today. I have to concede though that Google’s billion dollar business model depends on it distributing its runtime widely, and it has been relentless in throwing some serious engineering chops rather than just treating it as a cash cow.
Thanks for the insights. Though I have to say it’s hard for me to think of those kids living in their mothers’ basements writing Store apps as “ISV’s” ;-).
Almost all the great tech companies started in a basement, or a garage!
Eric,
10 years is about when .NET came out. (Yes, it was actually 2002.) What makes you feel that Microsoft has in effect abandoned ISVs? Are the tools lacking in some manner? Are they too expensive?
Hey, Bob,
First, let me say that my perspective is skewed by the fact that I develop desktop applications. I found .NET to be geared more for enterprise developers than ISVs. I love C# as a language, but .NET takes the wrong side of the safety/performance tradeoff for ISVs. Garbage collection is for babies. The original .NET library for desktop applications, WinForms, was a thin layer over the Windows SDK that just made it slower. Then came WPF, which was a great innovation in many ways, but again because of the .NET underpinnings and the retained-mode-only graphics system (catastrophic decision, retained mode is also for babies), was of limited use for ISVs. We ported the UI of our very large native C++ application from MFC to WPF, but our graphics had to stay as GDI because of the performance/memory overhead of retained mode. Based on conference calls, I think Microsoft realized after WPF (i.e., too late) that the desktop app ISVs that were left all had massive native code bases that were not going to be moved to .NET. Now we see a resurgence of interest in C++. And we also see that Windows Phone developers have been given the absolute dream development toolkit: the ability to use XAML with C++ and native code, along with the ability to use Direct2D to draw directly to the native buffer of XAML widgets. This is a set of tools that I would never have even dreamed Microsoft would have produced, and then they did, but it’s only for Store apps. These kids living in their mothers’ basements churning out these largely useless Windows Store apps (okay, it’s getting better) have tools that are orders of magnitude better than the ISVs (those few of us who are left) that built the desktop applications that made Microsoft what it is. The tools ISV’s have to work with today are older than the Internet (MFC). WPF has been abandoned. There has been a resurgence of native C++ as a language, but for desktop application developers, there is no application framework in which to put that new C++ to work other than MFC.
Microsoft has hastened the trend that they stand to be most damaged by, the trend away from desktop applications, by failing to provide state-of-the-art tools for ISVs.
I’m not asking for much. Just provide the same XAML/C++/DirectX support for desktop application developers that they have provided for the Metro app devs, and I’ll shut up.
Eric,
Thanks for your reply and explanation. It sounds like you are in an engineering or other graphics intensive field. In the past year I’ve done desktop development for a very large organization that has thousands of customers using this application, but it is a business oriented application. The main UI requirement there is simply don’t hang the UI thread. As such, a lot of the work is dumped into a queue with an event delegate back to the UI thread to keep the UI responsive. We don’t do anything like real time rotation/manipulation of 3D objects. If I am correct about the industry your products cover, I understand your frustration and believe you have a valid point.
I think this narrows the discussion tremendously. It isn’t about ISVs in any general sense, it is about a specific and fairly narrow class of ISVs. And yes, I think that Microsoft has paid less attention to this narrow class the last few years than they deserve. But do keep in mind that this is probably 1-2% of the ISVs out there.
Take heart in having the guy who put that dream environment on Windows Phone now in charge of Windows.
Pingback: Sources: Steve Ballmer's Reorg Takes 'A Big Risk' With Microsoft's Corporate Culture | Business Insider Australia
Pingback: Sources: Steve Ballmer’s Reorg Takes ‘A Big Risk’ With Microsoft’s Corporate Culture (MSFT) | This Is Jah Smith DOT com
Pingback: Sources: Steve Ballmer’s Reorg Takes ‘A Big Risk’ With Microsoft’s Corporate Culture (MSFT) //
Pingback: Microsoft’s risky reboot – can Ballmer make it work? | Seattle News
So basically Microsoft tried startup venture capital outside subsidiary
to create products when that didn’t work.
It has copied straight from Apple’s playbook being “Functional Organization”.
except Apple has 1/5 the engineering employees of Microsoft with small teams.
So real hammer is coming in a quarter or two.
This question comes is Microsoft going to have single person responsibility
or is it just gong to be pass the buck and have one more meeting to decide
on more meeting mode.
Did they get rid of stack ranking. No?
Not to talk about getting into device business and first thing Microsoft
does is first hawks reduced price to education and then to general public.
What will happen when 7 inch RT has to be reduced in price because expectation
is already set with 10 inch and The 10 inch RT will be EOLed.