Thoughts on the new Windows leadership

I’ve had a number of people question if  Julie Larson-Green is up to the task of running Windows Engineering.  No one has questioned Tami Reller’s expanded responsibilities because, well, Tami is pretty much doing the same job she had before except that the buck now stops with her instead of falling on the shoulders of a division President.  So I’ll focus this post on Julie and her new role.  And moreover on the experiment it represents.

Over the last few years you’ve probably heard a lot about Microsoft’s move to a “Functional” organization style.  The entire company isn’t there yet, but it has been the way of things in the Office organization for many years and was brought to Windows by Steven Sinofsky.  When an organization goes functional the major engineering functions of Development, Test, and Program Management all report up to a very senior level of management.  Usually a Corporate Vice President or above.  In some cases, like Windows, a business unit President.  One grows through the management ranks by taking on greater responsibilities within their discipline, but gains no experience in multi-discipline management nor ever has outright ownership of a product or technology space.

This contrasts with Microsoft’s historical style of having product units of (typically) 30-80 people with the functions reporting to a Product Unit Manager (PUM), multiple product units reporting to a General Manager (GM), multiple GMs reporting to a CVP, multiple GMs and CVPs reporting to an SVP, and then multiple CVPs and SVPs reporting to a direct report of the CEO.

Just looking at this from a management development perspective that means someone climbing the management ladder first learns to manage individual contributors (as a lead), then learns to manage managers (that is, be a manager of a set of leads) within their discipline, and then learns to manage managers across disciplines (as a Product Unit Manager).  PUMs also learn what it means to take overall responsibility for “a thing”.  Sometimes it is a product, sometimes it is a feature area, sometimes it is a particular technology.  But they own it and they own the overall commitments for that thing.  The PUM role is primarily an engineering role, and those that show excellent leadership, good people management skills, ability to handle a wider scope, business acumen, strategic thinking, etc. go on to populate the ranks of General Management (which includes both the General Manager and CVP titles).

My point in going through this is to explain a big part of the head scratching about Julie being named to run Windows Engineering.  Historically someone coming into this role would have proven their multi-discipline management skills and ability to completely own and guide a “thing” multiple times before taking on something as senior as running Windows.  They would have held one or more PUM positions, one or more GM positions, and one or more (general management) CVP positions.  Because she moved to the home of the functional organization model fairly early in her Microsoft career Julie skipped this path and instead took on increasing responsibilities within the Program Management discipline.  She is well proven as a (I would say THE) senior leader of the Program Management discipline within Microsoft, but unproven as a general manager.

This isn’t just an academic distinction.  One day around 2000 there was a knock on my office door and I turned around to see a couple of HR people who wanted to talk.  The topic was “Why do Group Program Managers fail as Product Unit Managers?”  It seems that HR had noted that successful PUMs come out of the Development discipline while unsuccessful ones came out of the Program Management discipline.  Without getting into that specific topic, the point is that great success within a discipline is not an indicator of ability to succeed in multi-discipline general management roles.   That’s why my personal preference is to discover someone’s abilities at general management earlier in their career than when you hand them the reins of a multi-billion dollar product.   By the way, one can now find plenty of examples of successful GPM to PUM/GM/CVP transitions so don’t focus on the specifics of HR’s dated 2000ish observation.

How much of a problem is the fact that Julie hasn’t been proven as a General Manager?  For a good leader it represents a challenge but certainly not an insurmountable one.  Assuming Julie is surrounded by discipline leaders that compliment her strengths and together form a strong team her lack of general management experience might represent little more than a speed bump on the road to success.

Was Julie the most obvious choice to take over leadership of Windows Engineering?  I guess it depends on the criteria you use.  Do you want your leader to be the most experienced manager available, or the visionary who will press forward with a change that is already underway?  Sometimes you can have both, but often those are different people.

There were choices besides Julie within the Windows organization that Steve Ballmer could have elevated.  Jon DeVaan is one of Microsoft’s most experienced executives and has held numerous CVP/SVP general management positions over the years.  Recall that on Windows 7 he was Steven’s peer, with Jon owning the Windows Core OS and Steven owning the Windows Experience (née, Windows Client) team.   Jon would have been an easy choice to take over Windows Engineering based on experience.  Chris Jones, the owner of Windows Live for many years now, used to be a VP in Windows Client.  Etc.  Without knowing anything about how these other executives are currently viewed it might be hard to say why he chose Julie over them, but it is very important to note that Ballmer did have choices.  Julie didn’t get the position by default, Steve obviously believes in her ability to lead Windows forward.

There is another cultural aspect of placing Julie at the head of Windows Engineering that is probably at play in many people’s initial reaction.  When I joined Microsoft in 1994 there were two very distinct cultures at play, named for the (earlier) divisional structure of Apps and Systems.  In the Apps (what we now know as Office) culture Program Management was the leading discipline.  By way of example the developer who coded PivotTable in Excel told me that he really hadn’t understood what they were for or how customers would use them, he’d just implemented the spec that the Program Manager had written.  That would never have happened in Systems, where the Development discipline reigned supreme.  In Systems the Program Managers were mostly process people.  Technical discussions of nearly any kind were always with Developers.

Read the book Showstopper! if you really want to get an idea of the Systems culture, particularly as the influx of Dave Cutler and others from Digital Equipment Corporation mutated it.  A few months before the book came out I had my one 1990s technical discussion with a Program Manager in Windows, Bob Muglia.  After that I think it was Developers all the way.  When Steven Sinofsky and Jon DeVaan took over Windows they shifted the culture more towards Apps’ Program Management-centric one.  That set up the possibility of Julie running Windows Engineering.  But for those groomed in the Systems culture, or who idolize some of its most prominent adherents (e.g., Dave Cutler), Julie seems like an odd choice to run Windows Engineering.

So is Julie a good choice?  On a strategic level I think there was no one better positioned to finish the job of re-imagining Windows that started with Windows 8.  I have some evidence that Julie is indeed easier to collaborate with than Steven was.  And she’s inheriting from Steven a well-functioning engineering organization that, of course, she helped create.  She doesn’t have to fix anything (major) that I know of on the organizational or engineering process fronts.  That means she has time for her multi-discipline general management skills to mature while focusing most of her energy on completing the Windows re-invention.  Plus, by splitting the business and engineering responsibilities across two executives (and taking on the President responsibilities himself) Steve has kept Julie’s new role from being too much of a stretch.  So yes, I think Julie is a good choice.  Hopefully we’ll be able to look back in a few years and say that she was a great choice.

This entry was posted in Computer and Internet, Microsoft, Windows and tagged , , , , , , , . Bookmark the permalink.

9 Responses to Thoughts on the new Windows leadership

  1. dave says:

    From what I read, building bridges between units is going to be a key skill required to help MSFT to evolve from its siloed software structure. Is there any indication one way or another that Julie will be able pull together teams that way?

    • halberenson says:

      There are signs that Julie will be better at collaboration and more responsive to internal partners. Things are also structurally different, which is important. Because Julie doesn’t have business responsibility it isn’t a simple argument for her to make that she can’t, for example, do something for Windows Server because it detracts from the Windows Client business. It would have been hard for Ballmer to overrule Sinofsky on such a decision and simultaneously hold him responsible for the Windows Client Business. The new structure doesn’t have that dynamic. If a tradeoff is significant enough that the parties can’t agree, Ballmer now gets to arbitrate without risking undercutting a Windows Division President. And since you really don’t want to be at that level and need Steve to step in, Julie has a very strong incentive to work things out without needing to involve Steve.

  2. Paul says:

    Interesting. Thanks.

    A question though. If W8 sales are coming in well under MS’s forecast, as Thurrott is reporting today from one of his internal contacts, and WinRT/Surface is also coming up short (as several media sites are claiming), couldn’t JLG find herself having to make some major fixes and rather immediately? I’m also wondering if a major factor in her promotion was to try and mitigate the obvious concern that Sinofsky’s departure caused? It was still a $10B haircut. But had she not been promoted I think it’s reasonable to think it would have been even worse. Not saying she’s not the right person. Just saying it’s not clear to me that was the main factor in Ballmer’s decision.

    • halberenson says:

      There are three different topic here.

      1) The claims about Surface seem all related to taking something Ballmer said out of context. The issue with Surface appears to purely be its limited distribution.

      2) Windows 8 sales. Given some of the channel things I’ve seen, like lack of Windows 8 Tablets and Convertibles at retail, it isn’t clear to me that this is an engineering problem.

      3) Julie has nothing to do with marketing and sales and so putting her in charge of Engineering doesn’t seem like it would be dictated by short term concerns. Yes I am sure that Steve wanted the message to be one of continuity, but that doesn’t help if 6 months from now he has to name yet another new person. His intent would have had to be that Julie is “the one” running engineering through at least the next major release.

      The stock market reaction to news Sinofsky leaving was not something that could have been mitigated by any specific person being assigned a role. It could only have been mitigated by a process that suggested a more orderly transition. It is the abruptness of the change, not change itself, that has observers and investors so unnerved.

  3. Pingback: A Sudden Departure | Geoff Coupe's Blog

  4. Pingback: Microsoft inside baseball: Are all collaborations created equal? | News Technology or Trading

  5. Pingback: Microsoft inside baseball: Are all collaborations created equal?

  6. Pingback: Microsoft inside baseball: Are all collaborations created equal? | Mon Web News

Comments are closed.