For most organizations, skipping Windows Vista and deploying Windows 7 offers an opportunity to postpone a potentially difficult and costly operating system (OS) migration. However, there are significant timing issues that must be weighed.

• The typical organization will need 12 to 18 months once Windows 7 ships until their critical vendors support the new OS, and they can complete their application testing, image building and piloting.
• Applications that do not run on Windows Vista likely will not run on Windows 7.
• Although Microsoft will support Windows XP through 14 April 2014, support for third-party applications on Windows XP will become a problem by 2012.
• Skipping a release (such as Windows Vista) means that organizations will not be able to move to the next release (Windows 7) leisurely, via PC attrition.

• When planning to eliminate Windows XP, don't expect to deploy Windows 7 for mainstream users before the first half of 2011.
• Test applications during 2009 and 2010, and create remediation plans to ensure you will not have to delay your migrations while you are fixing your applications.
• Plan to eliminate Windows XP from your environment by year-end 2012.
• Prepare budgets to deploy Windows 7 on new and existing PCs during 2011 and 2012, so you can eliminate Windows XP before it becomes problematic.

We continue to get inquiries from clients trying to decide whether to deploy Windows Vista or skip it and wait for Windows 7. Although Microsoft remains circumspect about its target date for the release of the new OS, we expect Windows 7 will likely be released to manufacturing in September 2009 or earlier so it can be preinstalled on new PCs available in store shelves by November. As the release of Windows 7 nears and the beta version garners good reviews for performance and stability, more organizations are beginning to question current Windows Vista deployment plans. It is getting more difficult for Gartner to recommend Windows Vista adoption to organizations that have not yet done any preparation. Current economic conditions are exacerbating the issue as organizations look for projects to trim to lower IT budgets.

For the most part, users want to run applications, not OSs. As long as the client OS is secure, stable and manageable, provides efficient access to the resources and runs their applications, users don't care that much what the OS is. Microsoft has continued to add new features (for example, BranchCache and BitLocker) that will show value and be compelling for customers. However, as Windows has matured, it has become increasingly difficult for Microsoft to innovate or introduce sufficient new features so that organizations can truly see a return on investment for the money spent on the migration to a new OS. Yet, OS standardization matters greatly to the cost and complexity of management/security. Organizations must eliminate older OSs and move to newer ones to ensure support of the OS and the independent software vendor (ISV) applications that need to be run. Therefore, deciding which version of Windows to run will have less to do with the value provided by the new OS, and more to do with reducing risks and minimizing the cost to test and deploy the new OS.

Support for any Windows version is finite, and organizations will have to migrate off it eventually. Organizations can do that steadily or quickly, but they must understand that a decision to stay on one OS longer will always mean moving to a subsequent OS faster. Our advice for most organizations is to deploy Windows Vista via hardware attrition in a strategy we call "managed diversity," where the OS is tied to the PC hardware and does not get upgraded during the life of the PC. This recommendation was never based on any significant value that Windows Vista has over Windows XP. Instead, it was based on the fact that organizations that don't bring in a new OS via hardware attrition will be forced to "forklift" (that is, move all the PCs at once) to the OS that follows. For most organizations, forklift migrations are too expensive and are often delayed to the point where the organization is stuck with huge numbers of PCs running the old OS too long. This frequently results in problems deploying new applications because they do not support the old OS. This situation is playing out with organizations running Windows XP, and that are trying to decide whether to implement Windows Vista or wait for Windows 7.

Gartner classifies Windows releases as "plumbing" releases, which make major changes to the Windows architecture, and "polishing" releases, which build on the plumbing release, improving usability and functionality. Windows Vista was a plumbing release; Windows 7 will be a polishing release, just as Windows XP was a polishing release of Windows 2000. The name Windows "7" seems to be more of a marketing statement to distance the new release from the poorly received Windows Vista than a marker of major architectural changes; it would be better characterized as v.6.1.

Many organizations will see refinements in Windows 7 as a reason to skip Windows Vista. However, the many improvements in Windows 7 have not persuaded us to change our basic advice on Windows Vista deployment. The features target personalized experiences, improved access to information and services, and improvements in device handling, but will not significantly improve compatibility for applications that require Windows XP or reduce application testing over Windows Vista.

This highlights Microsoft's problem with Windows Vista: The OS is good, but its reputation in the industry is bad. The reality of Windows 7 will be that it won't be that much different than Windows Vista, but the perceived improvement may be significant because of Windows Vista's tarnished reputation. Early reviews of the Windows 7 beta bear this out. Windows 7 may be easier to deploy because the political problems of deploying a disliked OS will be eliminated; but many organizations probably will not be able to justify a migration on cost or value alone.

Another hurdle to achieving corporate deployment of Windows Vista on new PCs is the weak economy. Economic concerns have resulted in many organizations reducing or delaying PC purchases. The concept of managed diversity depends on the continual refresh of PCs, which has typically been three years for notebooks and four years for desktops. In this economy, PCs are not getting replaced at the rate of 25% to 33% per year, and many organizations are investigating formally increasing the life of the desktop PC to five years and notebooks to four, for at least some users.

One of the principles of managed diversity is that the organization generally needs to support two versions of the OS in production at a time. At most, the period where three versions of the OS may need to be in-house would be less than a year when the last PCs running the oldest OS are on their way out, and the first PCs running the newest OS are coming in. Moving to a five-year refresh rate would almost guarantee the need to support three OS versions in some years (see Figure 1). This is outside what we consider to be manageable.

Based on our research, it would not be surprising to find that one-half of large organizations will skip Windows Vista and move directly from Windows XP to Windows 7 (see "Sneak Peek: U.S. Symposium Survey Shows Vista Adoption Slow, Office 2007 Moderate"). Of course, this doesn't mean that skipping a version is the right thing to do. Organizations still need to carefully consider what the right course of action will be for their situations.

The primary driver and the main objective for which organizations need to plan is to have Windows XP out of their environments before Microsoft no longer supports it with security fixes (14 April 2014), and before ISVs no longer ship desirable new products that require something newer, which likely will be common by 2012 (see Figure 2). SP1 refers to Service Pack 1. Gartner calls the initial release "SP0."

Organizations that have not been preparing to bring in Windows Vista likely will decide to skip it — a decision at this point that may make sense. Organizations contemplating skipping Windows Vista must plan to take several actions to reduce the risk of this decision (see Table 1).

Skipping Windows Vista puts significant pressure on the organization to execute a subsequent migration to Windows 7 in a relatively short time frame. Therefore, organizations should make a decision quickly, establish a project plan and start preparatory work during 2009. Most companies planning on skipping a version should consider beginning application compatibility testing using the Windows 7 Release Candidate to get an early start on Windows 7 testing. From this standpoint, Windows Vista (or Windows 7 beta and release candidate versions) serves as a good proxy for Windows 7. Applications and devices that do not work with Windows Vista likely will not work with Windows 7, while the majority of those applications that do work with Windows Vista shouldn't experience issues with the new OS.

Even with an advance start on testing, it is still unlikely that organizations could begin production deployment of Windows 7 before early 2011 (see "Windows 7 Won't Need SP1, But Will Still Need 12 to 18 Months Before Deployment Begins"). Thus, an organization contemplating skipping a Windows version must realistically examine whether it has the staffing, budget and skill set to move from Windows XP to Windows 7 in what amounts to 24 months or less.