This may sound obvious, but it is not obvious, otherwise we'd have beautiful, credible, executable, assessable, and risk adjusted schedule for all the work we do.
There are ten best practices described in this document. All of them obvious in their titles, and much less obvious in their execution. While it may seem trite to say it, these ten best practices are about all there is to building good schedules. But for some reason we rarely see actually good schedules.
These 10 are (if you don't want to download and read the full 211 pages:
- Capture all activities - this means for all the terminal nodes in the Work Breakdown Structure (WBS) there should be work activities that produce the products and services for that deliverable. These work activities should be contained in Work Packages, which are derived from the Accomplishment Criteria that support the Significant Accomplishments of the Program Event in the Integrated Master Plan (IMP). But that another topic
- Sequence all the activities - time (and order) prevents everything from happening at once - Einstein. But the order of performance requires careful consideration, not just arranging the work to fit the availability of resources, funding, or whims of management.
- Assign resources to all activities - this means a resource loaded schedule. A schedule, where the resources have been carefully considered and the work effort carefully considered as well.
- Establish durations of all activities - start with the reference class approach.What can you learn from somewhere else on how long it will take to do this work. Use the 1, 2, 5 estimating technique. Use some technique, do not just make up the numbers - have a model of how these numbers came about. There is lots of materials for estimating, read them.
- Verifying the schedule is traceable horizontally and vertically - it should start with the vertical traceability. As an IMP "jock" the vertical traceability defined the architecture of the program. Then the horizontal traceability come in.
- Confirm the Critical Path is valid - any good scheduler knows what the critical path is, she doesn't need to have a software tool tell her. If you're flying to low earth orbit or beyond, the propulsion system had better be on the critical path. If not your schedule is not credible.
- Ensure reasonable float - this number needs to be developed through a Monte Carlo Simulation IAW (in accordance with) DI-MGMT-81650.
- Conduct a schedule risk analysis - is the inverse of the float analysis. Both are needed.
- Update the schedule using actual progress and logic - answer the question how long are you willing to wait before you find out you're late? You'd better be updating - statusing - the schedule at half or less, because when you do discover you're late, you'll not have enough time to fix it.
- Maintain a baseline - no baseline, no credibility to the forecasts of future performance. You can't manage a project without variances. Rubber banding the baseline is not program management, it's cooking the books.
Each chapter goes into details for each practice. Each chapter has a case study to show how to apply the principles of the practice.
Download this, read it, start to apply it. Even if you're schedules don't improve, now you'll know why?