5.12 Plan B Strategies

< Day Day Up >

We will review the construction of Plan Bs in a minute. Before we do, let us take a peek at the traditional strategies used to prevent or react to Murphy's Law as presented in Exhibit 6. You are probably familiar with most of them.

Exhibit 6: Plan B Strategies

Plan B Strategy	Description
Pilot (proof of concept)	Instead of rolling out everything to 5000 Metro users or to all 200 sites, start out with a small, manageable sample. Set realistic goals and test parameters that allow you to gauge whether or not the implementation strategy works as intended. Engage beneficiaries from the planning phase forward.
Phased rollout	This is a natural extension of the pilot concept. Users are "converted" to the new network or application in measured segments or modules, not "flash cut." This gives you the time to measure success and react.
Day One versus Day Two	It makes sense not to rebuild the whole world at once. If you are remodeling your home, would you demolish bathrooms and the kitchen at the same time and try to live there? Identify any deliverables that can be pushed off until Day Two.
Backing out	In one major project, we rolled out laptops to hundreds of users, but left their old desktop PCs in place for a week while we tested new machine functionality. The thought was if we ran into issues, the old PC could be reconnected so the production user could carry on while the new laptop was reconfigured. Back-out strategies can be much grander in scope or scale, such as turning the old routers back on or reverting to the legacy payroll system.
Belt and suspenders	In the IP Telephone effort featured throughout the book, we wired the new site to accommodate both IP Telephone and traditional PBX voice technology in the event that the newer IP Telephone technology did not pan out. Although expensive, the cost was rationalized by the awful prospect of a voice outage impacting thousands of users.
Alternates	This strategy can be applied to facilities, computing platforms, network connectivity, resource, or vendors depending on the risk at hand. I once contracted a service bureau to process time-sensitive data to my specifications to meet a potential shortfall in the in-house development of an application to process millions of records as the front end to a complex mainframe accounting process. This was neither trivial nor cheap, but it was smart.

Three things should be clear from this discussion:

In the complex project world, the cost of Plan Bs is quite dear.
All require detailed preparation.
Long lead times may be associated with them.

This brings us to the next important Plan B consideration.

< Day Day Up >