Medical Device Link.

Originally Published MX November/December 2002

BUSINESS PLANNING & TECHNOLOGY DEVELOPMENT

Many engineers, when asked to estimate how long a task will take, figure out how much time they think it should take, triple that amount, and then add a bit for good measure. This liberal approach stems partly from knowledge that life in R&D abounds with the unexpected. It also results from engineers being conscientious people who are serious about keeping their commitments. Therefore, when pressed to predict the duration of a project, they make the situation as safe as possible for themselves under the reasonable belief that they will be held accountable for their estimate. In addition, experienced engineers have usually had the frustrating experience of asking for just the amount of time they believed a project would need, only to have some of it stripped away for a new management priority.

Figure 1. A project timeline based on liberal time estimates supplied by participating resource groups (a) will be longer than one based on likely actual task times required that includes a buffer period to absorb delays (b). (click to enlarge)

So margins of safety naturally get built into task time estimates. These inflated estimates are subsequently strung together into a project timeline that establishes a needlessly long expected project duration (see Figure 1). Such a timeline is problematic for several reasons:

• People try to start their task only on its start date.
• They focus on delivering their completed assignment on the last day allotted for it.
• When a task goes faster than planned at one stage, the people to take it through the next project phase likely will not be ready to start early, and so the opportunity for the project to gain time will be lost.
• When any task takes longer than planned, the project is at greater risk of coming in late.
• Projects get rescheduled with almost every update.

The best that can be hoped is that the project is completed "on time," with all the safety margins having been consumed.

The same project, with the safety margins removed from the individual tasks and a generous time buffer built in, could be scheduled in 25% less time. A project organized this way is more likely to be completed in a predictable and timely manner, for several reasons

• Each task time estimate expresses median duration, and is somewhat aggressive.
• Employees are encouraged to work on tasks as soon as received and to pass completed jobs to the next project group immediately.
• They are to focus on speed and quality in performing their tasks.
• When a task goes faster than estimated, buffer is gained.
• When a task takes longer than estimated, buffer is available to be consumed.
• The degree of project expedition is reflected in buffer consumption, which is a clear and early indicator of any risk to on-time completion.

In a scheme organized like this, regular project updates measure the size of buffer against the length of the process chain. In the example illustrated, the starting buffer length is 33% of the 10.5-week chain. During project implementation, if remaining buffer is measured at 25%, then contingency plans, say, for adding human or physical resources or air-freighting materials, are likely to be made. If the buffer has been reduced to 10%, then contingency plans are implemented immediately. A buffer measured at 30% well into the timeline is good news, and a visible indicator of a well-planned project with good likelihood of culminating in a timely and successful market introduction.

Copyright ©2002 MX