I am attending Realization's Project Flow 2010 conference in Chicago. One of the opening workshops yesterday was on "metrics and measures" for project management.
How are projects measured today? I would expect everyone is familiar with the overall metrics: things like number of projects completed; speed at which those projects are completed (cycle time); budget; and scope. These all tell us how things went. These metrics can give an overall sense of whether things are getting better or worse, but how do you manage and measure during the middle of your project; your portfolio of projects?
What are the internal project measures used today? Here are some:
- Adherence to schedule, or "will you finish your work on time?" Strangely, this leads to all sorts of interesting effects, none of which result in finishing the overall project on time.
- The efficiency / utilization or "are you busy?" metric. This metric could be used to help expand capacity, but it is rarely the only factor. It is more often used to either justify adding new projects or laying off people. The struggle with either of these options is that they have knock-on effects to whole system: more work piles up somewhere else. Piles of work essentially guarantees delays.
- Active projects. Strangely, the number of active projects doesn't necessarily correlate to the number of projects that are completed. But this metric sticks around in many organizations. The effect of adding more projects is the same as above: more work piles up and priorities are confused. And projects are delayed. See Little's Law for more.
I have left out a lot of the logic that connect the metrics to the ill effects. What I really wanted to recount are the metrics that make sense to help drive projects faster through the entire system.
So, what could some better metrics be that will help drive the right behaviors with the goal of FINISHING projects more and faster?
- Is the system currently meeting its targets? In other words, are projects delivering on time (and on budget, on scope)? Is the due date performance getting better / worse?
- Does the current prediction for hitting targets look positive? How is project cycle time doing?
- Project WIP. Control the number of active projects. The project management office and management in general need to resist the clarion call to add projects for every new good idea. Once the organization decides the right level of project WIP, the PMO has to ensure that the number of projects does not overwhelm the capacity of the system. The key here is that controlling the project WIP will ease the conflicts that cause people to multitask and delay projects.
- Task WIP at the constraint. If the constraint gets too much work, you may see more and more delays on projects that use the constraint.
- Task priority compliance. Are people working on their highest priority tasks first. Don't just work on the easy stuff or the stuff that is interesting: work on what is important to the success of the overall project. This requires the WIP control mentioned above. Without project WIP control, the priority of individual tasks goes haywire because there is simply too many things happening for the individual task owners. (I've written many times about the bad effects of multitasking.)
- Are tasks started with Full Kit? In other words, be sure to check the next week or two of upcoming work to be sure everything is in place for those tasks: people, equipment, supplies, documentation, etc. Tasks that start without full kit will take longer than necessary and often require rework.
- Issue resolution (stuck tasks). Work that can't move forward can be a killer on projects, depending on where that task is in the project. The metric here asks, "Are issues being raised immediately" (by task owners)? And are these issues being reviewed by line (and upper) management one a timely basis. The priority here should be the same priority information used in task priority compliance.
One interesting thing was said at the outset of the discussion: not everyone in a project environment is responsible for the same things, so they should not all be measured the same way. This is a clue to the kinds of measures you might want to use.