Measuring activity or measuring progress
The number of tasks created, hours logged in or messages exchanged tells you how busy people are, not where you stand. As early as 1992, Kaplan and Norton argued that measurement systems strongly steer people's behavior and that purely financial indicators can mislead an organization trying to improve (The Balanced Scorecard). The same logic applies to a project: what you put on the dashboard ends up steering what the team does.
A good indicator meets three conditions: the team can influence it, it moves fast enough to trigger an action, and you know in advance at what threshold you step in.
Five basic indicators
1. Completion rate
Tasks completed divided by tasks planned over the period. It is up to you to set the threshold that triggers a discussion (for example, two cycles in a row below the level you consider normal): underestimated scope or overestimated capacity?
2. Cycle time
The time between starting a task and finishing it. If it lengthens while the workload stays stable, something is lingering somewhere. It is tied to work in progress by Little's law (Little, 2011): the more tasks open at once, the longer each takes to get out, at constant throughput.
3. Blockers
The number of blocked tasks and the age of each blocker. Beyond a few days, a blocker calls for an explicit decision rather than silent waiting.
4. Workload per person
How tasks and effort are distributed. The aim is not arithmetic equality, but avoiding the breaking point nobody sees coming.
5. Velocity, if you work in sprints
The work completed per sprint, in effort points. It is used to compare a team with itself over time, never to compare two teams. A useful clarification: the Scrum Guide does not mention it; it is a widespread practice, not a rule of the framework.
Three traps
When the measure becomes a target. If you ask to "improve velocity", estimates inflate and the indicator no longer measures anything. Measure to learn, not to grade.
The dashboard nobody reads. An indicator nobody has looked at for two weeks can be removed. Every figure you keep should have a ritual attached to it.
The average that hides everything. A 90% completion rate can hide a critical milestone at 0%. Look at the distribution, not only the average.
Sources
- Kaplan, R. S., & Norton, D. P. (1992). The Balanced Scorecard: Measures that drive performance. Harvard Business Review, January-February 1992. hbr.org
- Little, J. D. C. (2011). OR Forum: Little's Law as Viewed on Its 50th Anniversary. Operations Research, 59(3), 536-549. doi:10.1287/opre.1110.0940
- Schwaber, K., & Sutherland, J. (2020). The 2020 Scrum Guide. scrumguides.org