There has always been a significant divergence in PC Software buying behavior between Consumer and Enterprises. Today we see lines at Apple stores to purchase the latest ithingy when they go on sale at midnight, but in 1995 that was people lining up to buy a Windows 95 upgrade. Even today, when Windows 7 came out the consumer PC market shifted to it immediately. But the Enterprise story is completely different.
Enterprises tend to shift operating systems, and other software, very slowly. First they have to evaluate the impact on their installed base of hardware. Some will be upgradeable, others will need replacement. Since we are talking about large numbers of PCs, often tens of thousands (and in a some cases 100s of thousands), it can take considerable time (i.e., years) to obtain enough budget to refresh an organizations PCs. Next they have to make sure their applications run properly on the new operating system. This itself requires dedicated staffing for a considerable period of time, particularly if applications need to be modified. And they may have to wait for a 3rd party to support the new OS, and then look at the impacts of upgrading to the 3rd party’s latest version. Then they need to manage the logistics of rolling out the operating system upgrade, potentially with new hardware, and potentially concurrent with new applications. And potentially with new training for employees. And before they start any of this they need to be convinced that the new operating system offers enough benefit for the cost and effort it will take (which, for example, is why so few businesses deployed Windows Vista) and that it is stable. Put all of this on a timeline and it’s a multi-year process, and often a many-year process. Which has led to an interesting dynamic: New Operating Systems come out before Enterprises finish rolling out the previous version, and thus they usually skip a version. Enterprises universally upgraded to Windows XP, most skipped Windows Vista, and most are now in the midst of their Windows 7 upgrade cycle. Thus the conventional wisdom is that most Enterprises will skip Windows 8. I think the picture has become far more complicated.
Windows 8 has two strikes against it for Enterprises. First is the life-cycle problem I describe above, and the second is that with a dramatically new user interface the employee re-training requirements are the highest since the introduction of the Chicago shell in Windows 95. On the other hand Microsoft continues to show that it understands and is avoiding key mistakes it made with Windows Vista such as requiring major hardware upgrades and breaking application compatibility. Windows 8 will run on all hardware that Windows 7 runs on, and likely run better on it than Windows 7 does. And application compatibility also remains extremely high, meaning existing Windows 7 applications won’t need work to run on Windows 8. Windows 8 also likely won’t have dependencies that force infrastructure upgrades (e.g., you can likely deploy Windows 8 clients without upgrading your Active Directory version). Microsoft has avoided the third strike.
Are their good enough reasons to rush out and upgrade all the PCs in your shipping department, call centers, factory floors, retail outlets, or even Information Worker desktops to Windows 8 after you’ve just deployed Windows 7? No. But I am now going to argue that is unimportant, because I believe most business are going to deploy hybrid Windows 7/Windows 8 environments.
Enterprises are as impacted by the shift to tablet computing as much as Consumers, and in many ways perhaps more so. The problem they face today is that as their Information Workers, or business units, want to incorporate tablets the real choices are iPads or a dizzying array of highly fragmented (from OS version and UI standpoint at the very least) Android devices. Application compatibility with their existing apps? None. Training applicability with their corporate PC setup? None. Management compatibility with existing practices? None. Ability to enforce information security policies? Limited. Ability to write new applications that work on their desktops and notebooks as well as their tablets? Limited (to websites basically). Expertise within their development organization to write applications for IOS and Android devices? Limited. And the list could go on. The bottom line is that, from an IT perspective, a Windows 8 tablet is far easier to incorporate into their existing an environment than any other option. So while they may wait for “Windows 9” before taking the Enterprise through a full upgrade cycle, Enterprises will begin targeted deployment of Windows 8 surprisingly quickly.
If Microsoft is really lucky (and has done a great job) the way this will start is by the CEO (and other CxO level executives) deciding they want a Windows 8 tablet and essentially forcing IT to let them put it on the corporate network. This is what happened with iPhones and iPads being allowed access to corporate email. It will then spread to other executives, and hopefully IT will find supporting them fairly easy. Other employees will have purchased Windows 8 tablets for personal use and be pressuring IT to let them use the same technology at work. IT will be forced to develop policies for allowing and supporting Windows 8 tablets within the environment, which then opens up Windows 8 deployment more generally. In other words, this time it will be Microsoft riding the “consumer-driven” IT trend.
More likely though is the path of a business unit having a requirement to deploy tablets (e.g., traveling sales or service reps) and IT responding defensively with a preference for Windows 8 tablets. And once IT has the capability in place to support Windows 8, and the policies to support bringing it into the Enterprise, business units will start to drive non-tablet deployments. The pace at which this happens is perhaps the most debatable point in this entire blog posting because it depends on how quickly Enterprise-oriented ISVs write Windows 8 Metro applications. And that is a chicken and egg problem. Enterprise ISVs aren’t going to switch focus to Metro until they have confidence that their customers are going to start deploying Windows 8.
I want to be clear that I’m not claiming fast deployment of Windows 8 in the Enterprise, I’m only claiming Enterprises won’t skip it completely (as many expect). Adoption will be near non-existent in 2012 and tepid in 2013. But by the time Windows 9 ships most Enterprises will already have deployed Windows 8 to a surprisingly broad part of their employee population. And then they can do the full cycle rollout of Windows 9 to replace both Windows 7 and Windows 8.
In the end it doesn’t matter for Microsoft if Enterprises move completely to Windows 8 or retain Windows 7 as their desktop/notebook operating system. What matters is that as companies deploy tablets they deploy Windows 8 tablets over competing offerings. A hybrid Windows 7/Windows 8 environment is total victory for Microsoft.
And what could go wrong? First, I completely dismiss a Vista-like scenario in which Microsoft ships something that has lots of negatives and no compelling value. The biggest threat to Windows 8 in the Enterprise is that in a Consumer-driven IT world the consumer rejects Windows 8. In that case the pressure for IT to make alternate devices, particularly Apple devices but also Android devices, equal partners to PCs in their computing environment could become irresistible. For those Enterprise IT guys who are frustrated by Microsoft’s consumer focus these last few years, that is the scenario that it is up against. To win in the Enterprise it must have a strong position with consumers. Failure to do that puts Windows in the same camp as MVS (z/OS), VMS, Solaris, etc. Enterprise-oriented operating systems that can continue to be important and produce healthy revenue for a decade (or decades) beyond their prime, but long ago lost their central role in the future of computing. And Microsoft is just not ready for that to be Windows’ fate.
I think tablet demand by corporations and end users has been established by iPad sales. Tablet demand is driving corporations to buy new technology at a faster pace than ordinary upgrade cycles.
The key challenge for Microsoft is driving end user demand for Windows 8 tablets because it delivers better consumer scenarios than iPad. Success here leads to bottoms up deployment of Windows 8 tablets into the enterprise. Because of Apple’s head start with iPad, I give this a low chance of success. It also introduces complexity because end users may also want to bring in Macs and iPads in addition to Windows 8 devices – so promoting a grass roots adoption route for Windows 8 isn’t entirely in Microsoft’s best interest.
The other scenario is to show Windows 8 tablets solve IT problems really well and are a compelling product for corporate IT to standardize on – and let IT make it the friction free choice for departments that need tablets. Today, much line of business tablet software is custom built – so an IT department that standardizes on iPad or Windows 8 will need to create custom software either way. As you point out, corporate developers are already familiar with Microsoft development tools, IT can reuse their device management technologies, and because its Microsoft, the tablets should work well with the Microsoft back office servers already in the corporation. Together, this makes Windows 8 tablets and easy choice for Corporate IT. Once the tablet decision has been made, it becomes easier to convert PCs from Windows 7 to Windows 8 if only to reduce TCO through device homogeneity.
I really hope Microsoft isn’t planning on releasing Windows 8 and then just sitting on it for a full three-year Windows dev cycle with no new user- or developer-visible features. The “Metro-style” environment is basically version 1.0 of a new platform, one where the market and competition is still changing and advancing rapidly – it’s very unlikely that the first version will be competitive through that whole period.
Windows Phone is a good analogy for what they should do, where the core (kernel, filesystem, etc.) seems to be on a longer cycle, but the built-in apps (which for mobile tend to be considered part of the OS), browser, shell, and application platform layer (Silverlight/XNA) are on an annual cycle. They already split out Windows Live Essentials and IE and put them on a shorter cycle, so that’s good, but they need to do the same for the Metro-style parts of the shell and for the new app platform (WinRT) as well. There needs to be a “Mango update” for Windows 8.
Besides it being a new and less mature platform and space, and one that – despite it probably getting some degree of enterprise adoption as you argue – will still be on balance mostly consumer-oriented in the short term, there’s another reason a mid-cycle shell / platform update is more appropriate here than for previous versions of Windows. The Metro-style app platform is much more restrictive in how apps can interact with each other, extend the system, and provide platform features. In the past third-party-provided alternative browsers like Firefox, middleware platforms like Flash and Java, and system-level utilities like Google Desktop sometimes filled holes in the system that Microsoft was slow to fill themselves, and while these may often have been perceived as a threat to Microsoft, they kept people who needed these features from having to leave the Windows platform altogether. This will be more difficult in the Metro-style environment, leaving Microsoft on the hook to fill any holes quickly.
I know this is a two-month old post, but this quote is crucial: “For those Enterprise IT guys who are frustrated by Microsoft’s consumer focus […] To win in the Enterprise it must have a strong position with consumers.”
For any developer, sys admin, etc. who enjoys working with Microsoft products and wants to continue doing that, MS must win in the consumer market. In the not-to-distant future the 20 and 30 somethings will be making the decisions and they will expect the hardware/software quality in the office to be no less than what they use in their personal lives.