Getting users excited, creating concrete communication plans, and creating an application map are just three key factors in a successfully testing Windows 7 in your company.
If a Windows 7 pilot test is in your future, you’re not alone. With Windows XP support running out in 2012 and Windows 7’s list of attractive new features, Forrester Research predicts Windows 7 adoption will be aggressive. Its survey of 655 PC decision makers in North America and European enterprises found that two-thirds of companies expect to migrate to Windows 7 eventually (citation: “Windows 7 Commercial Adoption Outlook,” Forrester Research, Inc., November 2009.).
But before that migration can happen, of course, comes pilot tests. Their success depends on many of the same project management best practices that any other project depends on. Communication, buy-in, monitoring, and other project management key factors are crucial. To that end, we put together seven steps to help guide you in conducting a successful windows pilot test.
Start with planning and requirements. Before the pilot test begins, the company should have a clear business case to test Windows 7, with buy-in established from the CEO down to the IT department, says Glen Gramling, senior vice president of sales and marketing for Infogain, an enterprise IT solutions company specializing in integration and custom application development. There should also be buy-in of the roles and responsibilities of each player, as well as an agreed-upon standard of project management methodology and policies. That set of standards is what works best for your company, says Gramling, but of particular importance are the following:
- Get stakeholder support. This step is absolutely necessary. Without stakeholder support, you can surely bet on failure. And if you think support is a given, think again. “Two of the most common culprits for failed projects are lack of stakeholder support or lack of strong project management,” says Gramling.
- Assign a project leader. It is essential that one person be assigned leadership to guide the pilot test from start to finish and to be the go-to person for problems, concerns, etc. Without a strong project leader, and the closed feedback loop that person provides, important information can be lost.
- Create a communications plan. Communication must be thorough and frequent. Gramling recommends weekly meetings for the project management team, with documented ways of sharing problems and issues that happen in between those meeting. Equally important is agreed-upon ways to share information into the ranks of the company.
Get users excited. Windows 7 offers new features, such as its virtualization capabilities, that can make it easier to get buy-in. Elvis Cernjul, vice president of IT for Spiegel Brands, is excited about Direct Access (which enables users to access corporate networks without a VPN) and BranchCache (which promises to decrease the time users spend waiting to download files across the network). Cernjul says that when you demonstrate such features to people and how those features make their jobs easier, Windows 7 can be a fairly easy sell.
Cernjul is currently pilot testing Windows Server 2008, and will embark on Windows 7 pilot testing soon; he hopes to roll out the new OS to the entire company of 600-plus desktops. He expects to approach getting buy-in from end users by pointing out specific ways that Windows 7 features will help people personally rather than talking about how it will help the company’s bottom line or some other top-down approach. Appealing to people personally makes getting buy-in far easier, he says, and thus ensures greater success for everyone.
Assess company’s applications and dependencies. The main hurdle to deploying a new OS are application compatibility issues, says Gramling. Most companies’ applications and technical details aren’t well documented, so he recommends starting with an applications inventory, in order to create an applications map that also documents dependencies (how applications interrelate and affect one another).
Raj Nath, CEO of NathCorp, an application migration consultancy, says that at this point in the process he’ll typically find suites of applications that are based on third-party middleware. He looks for issues such as the middleware vendor not having a new version, or the new version having so many changes that it will take a long time to fix all applications to work with the new middleware. Fortunately, Nath points out, Windows 7 offers technologies to address compatibility issues through offerings such as XPMode and Med-V. These Windows 7 features allow an application to run transparently within a Windows XP virtual environment, but to the end user it appears he simply has Windows 7 on the desktop.
Decide on pilot testers and create a training plan. Nath recommends choosing the first testers from the department with the fewest application compatibility issues. This helps ensure that positive results are gained.
That said, there may be other factors in play. For example, it is highly unlikely that you would want to start with users who perform a mission-critical function, even if they have the fewest compatibility issues. Cernjul, on the other hand, will choose one (or at most two) users from each line of business in order to test the broadest set of applications simultaneously.
It is also important to put a plan in place to educate users on Windows 7’s features. Nath recommends pointing users to Microsoft’s materials to help users get started with Windows 7, which can help ease the training burden on IT.
Embark on the proof of concept phase. In this phase, IT sets up a small number of computers to match the company’s standard configuration, and then tests the deployment plan against the computer. At this stage, it’s crucial to fully test application compatibility issues and to resolve any issues that arise before beginning with pilot testers, says Nath.
Run pilot test. The pilot test itself will enable Cernjul to determine whether upgrading to Windows 7 is worth the effort and cost at this time. The migration is the logical choice at some point, he says, since XP support runs out in 2014. The question is whether now is the right time.
Cernjul’s primary concern is compatibility issues surrounding some of Speigel’s homegrown applications written on .NET. During the company’s Office 2007 pilot test, the system did not work well with the company’s Citrix and SAS systems, for example; those problems needed remediation before Office 2007 could be rolled out throughout the company.
Similarly, a Windows Vista pilot brought up a number of compatibility issues, such as with the company’s BlackBerrys, one reason the company stuck with Windows XP. This time around, Cernjul has little doubt the company will eventually move to Windows 7; but the pilot test will determine if it’s a good idea to wait. Still, even if the pilot tests go well, Cernjul will wait until a Windows 7 Service Pack 1 before he rolls out large-scale deployment.
How long a pilot test takes can depend on a company and its business schedule. Nath recommends that a pilot test be scheduled to take about a month, which is long enough for most businesses to go through an entire cycle of business tasks. This timeframe may be longer for some businesses, such as a publisher that puts out a quarterly magazine. And that timeframe may be shorter for some jobs, for example, an administrative assistant. The important thing is that the pilot test timeframe allows for a full representation of tasks to be tested.
Moreover, Nath recommends that a proof of concept phase for the next set of pilot testers be held during the same time. This allows a speedier rollout, and helps users stay engaged.
At this stage—as with every other stage during a project—it’s important to “review, review, review, and then review some more,” says Gramling. You should monitor progress and communication among team members; also make sure feedback loops are closed so there’s no delay to corrective action when a problem or application dependency issues arises. Gramling’s company generally does deep reviews weekly, and has set in place a well-understood method of communicating should problems arise between those meetings.
For Cernjul, intimate communication is another bonus of the choosing just one person from each group. Doing so allows him to establish an intimate path of communication that enables users to be open and know they’ll be heard. And that communication is priceless.
Write it up. At the end of a pilot test, Gramling writes up a white paper that he leaves with clients, explaining what has taken place, as well as application dependencies and other technical information. Such a task may seem like overkill when doing your pilot testing in-house. Not so, he says. “People move on or change jobs, and such a document can save you time and effort in the future,” he says. “Plus, it’s a great way for stakeholders to tout success.”
I wonder which of these steps is the hardest for most IT managers and CIOs. To some degree, I’m sure that the “hard stuff” is a function of what one is good at. For example, some people are great at generating user excitement and getting buy-in; others excel at designing functional tests so there are no surprises (avoiding, “What do you MEAN that critical app won’t work on Win7?!”).
This question is less for you, Diann, than it’s for readers: What have you found successful in getting buy-in from stakeholders and users? Is it just a feature dog-and-pony show (“Look at how Win7 makes it easier for you to do this!”) or is there a more emotional component?
[...] Expert Voice can give you plenty of pragmatic advice about how to run a Windows 7 pilot and tools to help you change OSs, but hey, who wants to be practical? Instead, see what you can [...]
[...] as an IT organization need to be prepared. You need to practice.” In fact, Gold suggested that the IT department upgrade itself to Windows 7 first and use themselves to find the stumbling [...]