The diagnostic question
There is one question that cuts through this quickly, and I recommend asking it in your next review meeting:
“What decision does this number change?”
If the honest answer is “none — we track it,” then the number is not a metric. It is decoration. It may be interesting; it is not operationally useful.
The point of this question is not to be difficult. It is to force clarity about the link between measurement and action, which is exactly the link that quietly breaks in growing organisations.
Three failure patterns
Reporting as reassurance. Numbers presented to demonstrate that things are under control rather than to test whether they are. The tell is that the report always looks broadly the same, and unfavourable figures come pre-packaged with an explanation.
Metric inflation. A dashboard with forty indicators is a dashboard with no priorities. Nobody can hold forty things in mind, so people default to whichever number is currently causing them discomfort. Five or six metrics that genuinely drive the business will outperform forty that merely describe it.
Analysis as delay. “Let’s get more data” is sometimes rigour and sometimes avoidance. The distinction is whether anyone has specified what the additional data would change. If more data would not alter the decision, the request is procrastination wearing a lab coat.
What changes it
Assign an owner to every metric. Not a team — a person. Someone who explains movement and is accountable for the response. Unowned numbers get discussed; owned numbers get acted on.
Pair every metric with a threshold and a response. “Resolution time above X triggers a staffing review.” Deciding the response in advance, when nobody is under pressure, produces better decisions than deciding in the moment.
Separate the exploratory from the operational. Analysts need freedom to explore. Operators need a short, stable set of numbers they check consistently. Mixing the two gives you an operational view too noisy to act on and an analytical practice too constrained to discover anything.
Review the metrics themselves, periodically. Businesses change. A number that was central two years ago may now be measuring a problem you already solved. Retiring metrics is as important as adding them, and considerably less popular.
The cultural condition
None of the above works if bad numbers are dangerous to present.
If the person reporting a decline expects to be attacked, you will receive carefully framed reports and lose the early warning that makes the whole exercise worthwhile. Leaders set this directly, in how they respond in the first thirty seconds after unwelcome news.
The response that builds a healthy data culture is dull: “Thank you. What do we think is causing it, and what do you need?”
That is not softness. It is the only way to keep getting accurate information — and accurate information, arriving early, is worth more than any dashboard.
Dr. Mohamed Mousa writes about financial services, technology, and leadership.