Name the question before the number
Start with the question the team is trying to answer. “Are open tasks becoming harder to clear?” is a question. “Add a chart of tasks” is a request for an object. The question helps determine which records matter, which period to examine and which comparisons are meaningful.
Identify the person who will act on the information. If nobody can describe a possible action, the metric may be interesting background rather than a decision tool. That can still be useful, but it should not be presented as a signal that automatically requires a response.
Make the calculation reproducible
Define the unit, the population, the period and any exclusions. Another person should be able to reproduce the calculation from the same records. Write down how missing values, reopened work and duplicate entries are treated.
For an illustrative example, two teams both report “completion rate.” One divides completed tasks by tasks created that week. The other divides completed tasks by all tasks available that week, including older work. The labels look identical, but the numbers answer different questions. A shared definition is needed before comparing them.
Use a metric definition sheet
- Question: what decision or investigation does this support?
- Owner: who interprets it and who maintains the definition?
- Calculation: what exactly is counted, divided or compared?
- Population and period: which records and dates are included?
- Limits: what is missing, uncertain or deliberately excluded?
- Response: what would a meaningful change prompt us to check?
The sheet can be brief. Keep it close to the report so that its meaning is available when somebody reads the number.
Add context before adding more charts
A change in a metric may reflect a different mix of work, a change in collection rules or an incomplete reporting period. Write the context needed to interpret it. If a process changed halfway through the period, mark that change rather than treating the series as fully comparable.
Use a small number of supporting views when they resolve an actual ambiguity. For example, a total open-work count may need an age breakdown to reveal whether older tasks remain unresolved. More charts do not compensate for an unclear definition.
Treat thresholds as prompts for investigation
If the team uses a threshold, explain why it is meaningful and how it was chosen. An arbitrary red or green status can make a dashboard appear more decisive than the underlying evidence. A threshold should help focus attention while leaving room to inspect the cause.
Write the checks that follow a change: confirm the source is complete, review changes in the workload, inspect a sample and speak to the people doing the work. The metric is a starting point for a decision, rather than a substitute for understanding the situation.
Review the metric when the work changes
When the process, source or intended decision changes, revisit the definition. Record substantive changes so that historical comparisons remain understandable. Retire a metric that no longer informs a useful question instead of keeping it because it has always appeared in the report.
If the source itself is uncertain, begin with a focused data audit. If the numbers are being used to choose a direction, put the evidence and its limits in a short decision memo. Reporting is easier to use when the path from record to decision is visible.
Published by Naltyx. Examples are illustrative; any Naltyx product mentioned is our own. How we write.
Let’s talk