Rollout
Having got the system ready and the bugs fixed, it now becomes necessary to deploy it. How will this be done? If there are thousands of users in multiple geographic locations, there are choices to be made:
-
Big-bang rollout to every user on a single day
-
Staged rollout, site by site
The former is difficult to orchestrate and leads to huge problems in the event of a failure at a site. It also requires that all of the training and infrastructure work be completed in time. On the other hand, it does mean that there are no synchronization issues with previous systems.
The staged approach is easier to manage and allows some flexibility in dealing with hold-ups in the other aspects of delivery. However, it introduces a whole new can of worms in terms of keeping the new system and the old synchronized. This may require multiple data transfers in both directions, as well as temporary modifications to escalation and reporting processes. It may also mean extra work for the legacy systems with which the systems must integrate.
Finally, you must prepare for the worst. Have a contingency plan to cope with total failure. No matter how much you test a system in a lab environment, and no matter how well you pilot it, you can never be 100 percent sure that it will work until it does!