We all know that in the coming few weeks there will be many more details available on Microsoft’s 2012 product launches. Final Windows 8 details, Surface details, the launch of Windows Phone 8, the launch of Office 2013, the Xbox Dashboard update, etc. But what comes next? Forget everything you think you know about Microsoft, because it seems likely that 2012 was just the tip of the iceberg in terms of change.
A lot of the news I expect to hear between now and the end of the year is around reorganizations. The Server and Tools Business (STB) recently reorganized many of its divisions into a functional organization reporting to President Satya Nadella. Soon we’ll here of more reorganizations, including changes in Steve Ballmer’s direct reports. That this is an ideal time for reorganizations should not come as a surprise. Those will be in the details. Today’s Microsoft organizational structure largely reflects the “old Microsoft” and now it needs to reflect the “new Microsoft” and what it needs to accomplish.
There may be some huge surprises in the upcoming reorganizations but the only thing I’ll speculate on is what happens to Windows Phone. It makes no sense as a standalone business reporting to Steve, so where does it go? Speculation, even before the Entertainment and Devices business unit was split apart with Robbie Bach’s departure two years ago, was that Windows Phone would eventually become part of the Windows business unit reporting to Steven Sinofsky. This still seems likely. However, Microsoft has also made previous attempts to combine various telecommunications offerings together into a single business unit and it is possible they’ll try that approach again. Recall that the smartphone business was once part of a broader Mobile Communications Business (MCB), but as part of gaining focus on Windows Phone MCB was consumed by its child and non-WP efforts abandoned. However, with the acquisition of Skype one could imagine combining Windows Phone, Skype, Lync, and perhaps other telecom-oriented assets (e.g., Mediaroom) into a single Telecom Business Unit. It’s a long-shot, but it makes sense.
It should also be no surprise that Microsoft has already started some planning work on the next version of Windows. Early disclosures about Windows 8 included comments about work done prior to Windows 7 RTM. Rumors are calling this Project Blue or Windows Blue and wondering why it isn’t called Windows 9. Some are speculating that’s because it is something between Windows 8 and Windows 9. The explanation for this apparent change in project naming schemes can run from extremely simple to wildly speculative.
Let’s start with the simple. Sometime (relatively late) in the Windows 8 cycle it was decided to give Windows on ARM a different name and some different characteristics (no Win32 apps, no joining to Domains, etc.). The “Windows 8 Project” thus produced two products , Windows 8 and Windows RT. Calling the next project “Windows 9” thus makes no sense, even if the Windows on x86 product is called Windows 9. You may now just have a project called “Windows Blue” that will produce two products, Windows RT and Windows 8.1/9/whatever. That would be the simple explanation.
The more complex explanations run a broad gamut. Microsoft could be signaling that it is considering a naming scheme change for the x86 product and thus using a number in the project name makes no sense. Or the name “Project Blue” may encompass more than just Windows, a topic I’ll get to later. Most likely it is also an acknowledgment that everything we know about Windows (and other product) release cycles is outdated.
I’ve blogged previously about how Microsoft might change its release cycles to match modern expectations, but we won’t really see what they’ve decided to do until about 12 months from now. I believe Microsoft is going to standardize on 6 month releases for Cloud offerings and 12 month releases for packaged, or on-premises, products. They’ll do this for four reasons. First is that technology is in a rapid change phase and the 24-36 month cycles that had evolved over the last couple of decades no longer accommodate that change. Second is that this is the expectation that major competitors (Apple, Google) have set, particularly for consumers. Third is that as Microsoft provides common products for Cloud and On-Premise use they can’t allow the two to deviate to far from one another. A six month cloud release cycle and 24-36 month on-premise release cycle means that customers would have to implement hybrid solutions using products that are as many as six releases apart. Fourth, with IOS dominant and Android ascendant, Microsoft has to become far more agile if it wants to see it’s “re-imagined” Windows bet pay off in the tablet space.
Before all the enterprise types freak out about 12 month release cycles keep in mind that back in the 80s this was more typical for enterprise products. At one point DEC’s Rdb/VMS was releasing every 6 months! The key here is to keep the change from release to release very well controlled so that neither apps, nor user training, are broken. It is more of a slow evolutionary approach. From an IT perspective, these releases (despite considerably more functional change), should look closer to what is required for deploying a Service Pack than what is required today for a new release.
The last topic I wanted to cover is a change which many may think has already happened, but it is just starting. Although it looks like Microsoft did a very good job of coordinating and planning its 2012 releases the reality is that it was simply serendipity. For example, the Office team did not start out with Windows 8 support as the high-order bit in its planning process. It was focused on the Cloud, beating Google, doing more around Information Protection, it’s perennial stalking horse of Collaboration, etc. Oh, and we have to support touch. Oh and there is Windows 8 so we better have a release somewhere around the same timeframe to support it. This is why, for example, there are no pure Windows Runtime versions of the Office client apps other than OneNote. Besides the teams reporting to Steven Sinofsky (Windows, Windows Live, IE) the only Microsoft organization to truly align its planning and priorities with Windows 8 was the Developer Division. Everything else was serendipity and/or late plan changes to take advantage of the momentum building behind Windows 8.
I believe this is changing, with Microsoft moving to more broadly align planning across its product family. As an example, if Windows RT is a demonstration of where Microsoft wants to go with Windows than allowing desktop-free environments is a priority and requiring it for Office is an embarrassment (to say the least). Joint planning would identify elimination of the desktop requirement as a goal and the need for the Office team to produce pure “Metro” versions of its products on the same timeline as Windows releases. Likewise on the Enterprise side the various Server products have never coordinated their planning and are now likely to move to do so.
Note that coordination is a lot easier when product cycles are 12 months long than when they are 24-36 months long. The shorter cycles eliminate much of the schedule risk of taking dependencies on one another, and even more importantly eliminates much of the strategic problem of having to meet your own customer requirements but not being able to wait for the other product’s next 24-36 month cycle (which in the worst case could mean 4-6 years for you to address a customer requirement).
Which brings me to one very speculative question, what if “Project Blue” isn’t the name for the next release of Windows but rather the name for the next wave of client products that includes a release of Windows?
Over the next few months we will see Microsoft reorganize, change its planning processes, and replace its historical release paradigm, to adapt to the changing world. It’s a new Microsoft, and seeing how it changes to create and release software is almost as important as knowing what it will produce.
Certainly could’ve fooled me.
Sinofsky’s Office background, plus the fact that he took some of the most prominent Ribbon people with him, plus the Surface thing, plus the launch of Outlook.com with Metro-inspired UI features, etc. WinRT itself seemed like a minor miracle, considering the bad blood between Windows and DevDiv over .NET and Longhorn, WPF’s performance issues, etc.
It sure looked like Microsoft’s divisions had already begun to coordinate more closely. But you’re speculating that this was just the tip of the iceberg — that the coordinated efforts had barely begun. Certainly, we live in interesting times.
Everything in your second paragraph was under Sinofsky’s control.
DevDiv really had no choice in the matter. Had they not gone along Sinofsky would have built his own competing development platform and tools. Or more likely, DevDiv would simply have been taken out of STB and given to Sinofsky. The reason for it being inside STB is itself historical. Most of the server products, other than Windows Server, got their start inside DevDiv. In the late 90s they outgrew it and became peers, with DevDiv returning to its narrower focus of building dev tools and then, when Java became a competitive threat, a platform (.NET) of its own. So DevDiv has lived inside the overall STB umbrella for some time. But with Windows 8 introducing a new application model Sinofsky would have been well justified in calling for DevDiv to either make that application model and Windows 8 a priority, or move DevDiv to Windows. Although I do think it would have been amusing had Windows 8 adopted Eclipse as its IDE 🙂
Coordination (or lack thereof) was always the fatal flaw at Microsoft while I worked there. The degree to which different teams were at war with each others shocked my naive self. I think it’s a VERY good thing that Microsoft is starting to provide a more top-down unifying vision of what types of products the company creates. It was what Gates used to provide to the company, to a certain extent, and was what, more recently, differentiated centrally-controlled Apple (under Jobs watchful eye) from fractious Microsoft.
Letting the XBOX team (as an example) chart its own path was probably necessary under the old fractious Microsoft. Pulling it back into a unified vision of the company is a sign that Microsoft has changed.
Microsoft’s breadth of interests have always been its undoing in this regard. For example, the Exchange team has always been focused on Enterprises. If you tell them to produce a Consumer Email system then it is going to detract from the heavy demands (and then rapid growth) of the Enterprise side. So you buy Hotmail and end up with two email offerings and ecosystems. Bad idea? Think back to 1997 and that environment. Lotus Notes was considered a near existential threat to Microsoft. One could not imagine defocusing Exchange in that environment.
The mistake with Xbox was not in allowing it to chart its own path initially, it was in paying lip service for so long in how it would relate to the rest of the product family. The Xbox team was too busy in reaction mode competing with Sony and Nintendo to focus on the bigger picture. In fact it was too busy in hardcorps gaming to focus on casual gaming and, even more importantly, other forms of entertainment. It paid them lip service until just the last 18-24 months. But you couldn’t have taken the lead away from Sony, and countered Nintendo’s resurgence, without letting the team spend most of its existence really focused on that problem.
So what Microsoft has been lacking is a strong hand to decide when to coordinate and when to allow people to follow their own course, and when to reintegrate efforts. Even Bill was rarely as decisive on this as someone like Steve Jobs. But at least Bill was making well informed decisions on where to allow what behavior.
Disappointed to read your assessment that the product portfolio is still a function of serendipity vs planning. The need for coordinated planning should have been evident and in place years ago, at least imo. But it’s hard to argue with your perspective given some of the incomplete product offerings (i.e. Office for WinRT). Agree that WP makes no sense as a standalone business. An alternative to being subsumed into Windows or put into a telecom unit would be for MS to drop the OEM model entirely [in the phone segment] and go direct (potentially by buying Nokia, etc). That seems like a more viable option that saddling an already growth challenged business like Windows with what will be significant ongoing costs keeping a phone OS current or ideally making it leading edge.
One change you didn’t mention was Steve Ballmer himself. With the recent terrible sales numbers out of DELL/HP and obvious evidence that tablets are making a big dent in PC sales (despite MS’s earlier claims to the contrary), MS is likely looking at somewhere between 3-0% growth for FY 13 and a raft of downgrades after it reports numbers this quarter. The stock price is currently far too high for that growth profile, so either the stock plunges until equilibrium is restored or MS will have to cut expenses dramatically. I have friends at HP and know they’ve already begun offering widespread early retirement packages (presumably as a precursor to layoffs if those aren’t fully enough subscribed). Will MS take similar steps? I doubt it given all the product launches this year. So how much more pressure will investors place on Ballmer, particularly with Apple stock up another 60% this year alone, and Google poised to overtake MS next, possibly even before the shareholder’s meeting? I realize speculation about Ballmer being fired has become a national pastime, but in light of the above this year is truly different from others. There’s now undeniable evidence that MS is being disrupted and its strategy for growth isn’t working. WP has failed to be anything more than a placeholder in mobile. Bing hasn’t gained any share against Google despite losing $8 billion trying. ERP/CRM is sitting at around $2b vs the $10b by 2010 originally called for by Raikes when billions were spent buying Great Plains, Navision, etc. Xbox is still underwater as an investment after thirteen years and even back losing money last quarter. And worst of all, Windows is being disrupted by tablets and smartphones, two areas that MS pioneered for a decade before being blown out by Apple and Google.
The Longhorn disaster, which was partially about lining up everyone around the Windows release, left nearly every product years behind where they needed to be. Only Office, which refused to go along with locking it’s fate to Longhorn, escaped the disaster. So Microsoft went through a period of avoiding locking different products efforts together.
As for the business concerns those are all backwards looking and I think the market gets it. They will wait to see how the new products do before there is any renewed momentum around booting Ballmer.
Seems like the wrong lesson for MS to take away from that failure. Understandable, perhaps, but the wrong one nonetheless.
Disagree about those business concerns being backwards looking. All have direct implications for future growth, which is the main thing Wall St. cares about. I see Ballmer coming under a lot of pressure this fall, and forced out completely by this time next year.
One thing I wish someone would address: Is Microsoft actually mothballing the Desktop? Can we expect no significant technology investment on the desktop side going forward? I mean, I can understand the pendulum swinging to mobile for Windows 8, but are we seriously at a point where the pendulum is not going to swing at least partially back to the desktop? Apple is the market leader in the mobile space, but they show no signs of folding up their desktop tent. The desktop is the one place left where Microsoft has a stranglehold on market share. What business sense does it make to shunt their area of strength aside and essentially take the market share battle to Apple’s turf? If Microsoft is telling ISV’s, “Spend the millions of dollars necessary to port your desktop apps to Metro!”, a lot of those ISV’s are going to pass. It is hard to see how a lot of desktop apps fit in the Metro environment. And while I’m attracted to Metro/RT as a mobile platform, when I think about having to do all my work in that environment, I get claustrophobic. Are desktop applications that unimportant now? And yet Office is still a desktop app. Microsoft seems extremely uninterested in providing any clarity about it’s vision of the future along these lines. I guess we’ll just have to read the tea leaves from the coming re-orgs.
What Microsoft does with the desktop depends on many factors, not the least of which is what IT and enterprise ISVs tell Microsoft they want. But the classic desktop computing market is shrinking, and it has probably run its course in terms of major innovation possibilities. So just what should Microsoft really do with it?
Personally I would probably create a Windows Workstation product that was desktop focused and sell it at a premium for those usage scenarios that the new paradigm isn’t very good for. Without that (or a major turn of the crank that makes “Metro” more suitable for those environments, Microsoft will lose a chunk of its business to Linux.
I don’t really expect major innovation, but I do wish they would give native application developers (like me) a decent alternative to MFC (like the XAML+Native Code option that Metro app devs have) before they completely stop investing in it. Unless, like you said, they have something in mind for WinRT that would make desktop scenarios fit better.
Lots of great points made in your blog posting. Synchronizing release cycles and enhancing inter-product functionality is critical if MSFT wants to become consumer-focused.
MSFT has to find a universally accepted measure of success in the phone and tablet space, or it is at risk of loosing everything.
What cam be done to recapture the hearts and mind of the millions of commercial and corporate app / application developers?
Who will own this problem under various reorganizations?
I’m still trying to figure out what you are trying to recapture, their love for using Windows as a development environment or for writing Windows desktop apps? On the latter the issue is that all the client apps growth is in the mobile space, even for corporate and commercial apps. Developers aren’t running away from Microsoft, they are running to IOS and Android development because that’s where the demand is. The purpose of Windows 8 here is to give developers a reason to stay focused on Windows, albeit the new app model, because their users will have Windows 8 devices rather than iPads et al.
So the real problem with your first question is that it is a bit like asking “What is IBM going to do to recapture the hearts and minds of the CICS developers?” back in 1992. Well, whatever they tried it was irrelevent becuase CICS development, which continues to this day, was on its way to becoming a niche. Developers willing to learn new things, or desiring to follow the gold rush, moved on. Microsoft will likely do a few things for desktop development now that the initial push for the new app model is out of the way, and Native XAML makes a lot of sense as one of those. But in the end consider that desktop app development, which the web had already greatly reduced, is a shrinking niche. Nothing Microsoft could do would change that trend.
As for your last question, well it points to one of my dislikes of the functional org structure. It is ownership by committee. Some triad somewhere in the organization will own some piece of it. It isn’t clear that any triad will actually own the problem outright. Single point of ownership is only at the President level, who will not likely have time to pay attention to it. This is why I dread the day they take SQL Server and have it report functionally up to a President. I can’t imagine being successful in the database space with such a structure.
Yes, I formulated the question poorly, and your IBM / CICS example is good.
As a software marketer, it seems almost all the young developers I see are nowadays working form MacBook Airs. Almost all of them. The guy carrying the ThinkPad is the unusual rare and odd man out.
Yes, the problem is multifaceted. And yes it is the tip delivering the poison that Mac /iOS and Android are into the beast of Windows.
That’s because they are mostly developing for IOS, which requires an OS X machine, or for the web which can be done on any platform. And anyone under 30 seems to have a revulsion to Windows and a love of anything Apple.
My brother (who is in his late 40s) now runs Ubuntu as his OS because the technologies he chose for his startup to use are primarily Open Source and most native on Linux. He could use Windows for development with these technologies, but he can’t find a reason to do so. There are three applications he does occasionally boot into Windows to run (though honestly I don’t know why he doesn’t just run Windows in a VM for that), but occasionally means “rarely”. And because so much has moved to the Web, he really doesn’t need Windows for personal productivity either. Now if he were a marketing or other business guy he might care more about having an environment that was optimized for Microsoft Office, but he’s an engineer.
If Windows Runtime et al catches on then we’ll see a shift back towards Windows-based development, but otherwise I think we see continued drift towards OS X and more likely Linux as development platforms for both Client and Server apps.
I see four problems with your scenarios:
1. The natural aversion of teams to become too dependent on other teams (Office is the prime example, and they were proved right about the time of the Longhorn reset). And again, thinking of Longhorn, if the dependency graph gets too complex (or circular), well, then you get Longhorn.
2. The problem of doing “big things” in short cycles. Given Windows dev/build/test/package cycles, trying to release a version every 6 months or a year means that the time allowed for “greenfield” dev (working on brand new (often distinguishing) features) gets measured in single digit weeks. It may, however, make sense to have a two track approach – one for “major versions” and one for “minor versions”.
3. Enterprise customers will hate this. There are a lot of major customers just finishing up (or in some cases, just starting) their move to Windows 7. Vista took way too long, but 6 months or a year is way too short.
4. Support. Microsoft supports enterprise software for 10 years (though with non-security hotfixes only for 5 years). Service packs don’t get that. Service pack N only gets support for 2 years after SP-N+1 comes out (for the OS). If fully supported versions of Windows are released every year, the number of versions in support will grow to a large number. Supporting that costs a lot of money (Apple doesn’t have this problem – they only support version N and version N-1 at any one time (and they quickly leave the laggards behind).
You are making the classic arguments, but there are ways around all of them.