Two factors drive IT infrastructure migrations. First, IT management, senior management, and technology users have to be uncomfortable with the existing configuration. Second, compelling new technology must offer features and benefits that IT management and senior management can use to justify a major investment in time and money. In other words, large-scale migrations have both push and pull aspects. Unhappy or unsatisfied users push companies to migrate to new technology. And a new technology's potential benefits pull companies to invest in it.

An exclusive analysis of Survey.com data underscores the push component of the Win2K migration. Almost all organizations that migrate to Win2K have large Windows NT server and workstation installations. Figure 1 shows the top five features that IT managers who either have no plans to evaluate Win2K or haven't yet begun the evaluation process dislike the most about NT deployed on servers. Figure 2 shows the top five features IT managers who have completed or are in the process of completing Win2K deployments dislike the most about NT running on servers.

The differences between the two groups are stark. IT managers who still show no inclination to move to Win2K server implementations are upset about NT reliability, security, and total cost. They've invested significant resources getting their NT installations working correctly. As a result, they apparently are loathe to move too quickly to a new platform.

In contrast, IT managers who have moved to Win2K were more concerned about NT's management limitations. Their focus on remote management, Hierarchical Storage Management (HSM), and systems management features indicates that their IT infrastructures are complex and sophisticated. Control is a central issue.

Although more than 50 percent of the respondents who haven't yet begun Win2K evaluations indicated that reliability/stability, cost, and security were the worst NT features, no feature was mentioned by more than 33 percent of respondents who have already moved forward with Win2K deployments. In those settings, NT's shortcomings were much more context specific and surmountable.

Similar disparities in NT assessments exist among those who haven't started Win2K evaluations and those who have begun to deploy the OS on desktops and laptops. Figure 3 shows the five most mentioned NT shortcomings on the desktop among IT managers who haven't started the evaluation process. Figure 4 lists the five most mentioned NT shortcomings on the desktop for those who have begun Win2K deployments.

Once again, IT managers who aren't moving quickly toward Win2K are more focused on cost, performance, and reliability/stability issues, while those who are deploying Win2K are more concerned about efficient management. And although cost is an issue, only about one-fifth of those who have deployed Win2K listed it as a "worst" NT feature, compared to about half of those who haven't yet evaluated Win2K.

Figure 5 and Figure 6 compare the same groups for laptop deployments. Again, those IT managers who aren't evaluating Win2K are focused on cost and performance issues. Those who are deploying Win2K care more about management issues.

The message from these numbers is clear. Win2K deployments are still in the domain of the early adopters who are based in sophisticated IT shops. The push factors—the reasons that drive them to look at Win2K's advantages—cluster primarily around management issues, where NT limitations are most acute. For IT managers who haven't yet stepped up to the Win2K plate, the push factors are cost, performance, and reliability. They must be convinced that Win2K will deliver significant improvements in these areas in which they are most unhappy with their NT installations.