Project Tracking in Routine
The effectiveness of software project controls will be revealed to you routinely as a regular part of tracking project activities. One of the primary reasons for systematic project tracking is to gain the earliest possible insight into developing and potential problems and opportunities.
Three standard ways for you to monitor the effectiveness of your project control are to evaluate:
– Work package completion percent
– Milestone success rate
– Trend lines
Work package completion percent offers the quickest feedback, but gradually developing problems will be hard to see. Milestone success rate is evaluated less frequently, but developing problems are more apparent. Trendlines are both shortterm and long-term, but require collecting a series of data points before the trend can be considered revealing. Although they are of relatively less value early in a project, trendlines become very convincing indicators of project control effectiveness during the middle and later phases of a project.
Work Package Completion Percent: Work package completion percent is the ratio of completed packages to those still incomplete, including those not yet started. Since a typical work package for an individual or very small team (two or three people) is usually four weeks or less, on a project of a dozen people or so, you may have one or more packages being completed each week. This allows very detailed insight into project control effectiveness as a function of the changing status of work packages, which is reported during weekly status meetings.
Discrete status changes for work packages include:
- Work package available
- Assigned
- In progress
- In unit test
- In integration test
- In system test
- On hold
- Completed
- Products added to baseline
As I mentioned above, any single work package should never be tracked in terms of being partially complete. Instead, use binary status tracking, whereby a single work package is shown as either 0 percent complete or 100 percent complete. With binary status tracking, there are no intermediate stages of percent complete.
Work package completion percent is the number of work packages that are 100 percent complete compared to the total number of work packages. If you planned to have 25 percent of the work packages completed by the fourth month of a project, then comparing the actual number of completed packages at the fourth month to the planned amount provides excellent feedback on your overall project control effectiveness.
Milestone Success Rate: Milestones characterize major targets of progress within your project plan. Generally, a greater number of milestones is better than relatively few. Milestones allow you to verify the progress, or lack thereof, on the project. They represent your commitment to complete certain activities or parts of products by certain dates.
Failure to achieve milestones on their originally planned dates is indicative of ineffective project controls. The most obvious indicator of not achieving a milestone is, of course, needing to slip the date further into the future. If you adjust the date long before the date actually arrives, this typically creates less of a problem. It indicates that you are monitoring the project closely and adjusting plans to accommodate the latest developments. However, if the first clear indication that a date has to slip is the arrival of that date concurrent with an inability to pass the milestone acceptance criteria, this clearly indicates inadequate project tracking and control.
In the former case, you are comparing actual progress to plans, anticipating future events and capability, and adjusting dates as a function of project performance and predictions based on accumulated project data. In the latter case, you thought you were done with a major activity or a part of the system, but you weren’t. In either event, adjusting the milestone date requires you to:
- analyze the scope and complexity of defects within the product
- estimate the number and types of latent defects
- determine the amount of rework that will be necessary to address known and potential problems
- determine the resources available for performing the rework
- adjust the milestone sufficiently far into the future to allow rework and internal retesting to be completed
Another way to miss a milestone date is not to have the anticipated features that were originally planned for that milestone. This is characterized by reducing or relaxing the acceptance criteria for a particular milestone. Although in your plans, it will look like you achieved the milestone (that is, the date didn’t move), you should still consider this to be a signal of possibly ineffective project controls.
Trendlines: Trendlines give you a graphical depiction of relative values of data over time. Usually, you will plot several trendlines on the same chart, or profile, to facilitate instant comparisons between related values. For example, you can use a project profile to plot trendlines for:
- total hours worked
- total size of completed product
- productivity (total size of completed product/total hours worked)
- total defects detected
- defect density (total defects detected/total size of completed product)
These five trendlines are all plotted on the y-axis, and time is plotted on the x-axis. Three of the values (total hours worked, total size of completed product, and total defects detected) are simple measurements that either increase from period to period or remain the same. The other two are metrics calculated from the measurements.
Plotting multiple trendlines on the same profile allows comparison of related metrics and a projectwide interpretation of trends as a function of related project activities.
Tweet |
|