Metric sprawl is the accumulation of metrics past the point where anyone can act on them. It is the most common failure mode in mature measurement programmes, and it develops slowly enough that no single step looks unreasonable.
Every new tool ships with metrics, and switching one on is free.
Every incident produces a request for a new indicator, and refusing looks defensive.
Every audit finding gets a metric attached as evidence of response.
Every new executive wants a view shaped their way, and the existing views stay in place alongside it.
Nothing is ever retired, because retiring a metric requires proving it is not needed, which is harder than leaving it.
Attention is finite. Thirty metrics in a pack means each gets a minute, which means none gets examined.
Definitional inconsistency becomes inevitable. At forty metrics you can hold the definitions in a document. At three hundred you cannot, and duplicates with slightly different scopes start appearing in different reports.
Production cost rises until the reporting cycle consumes the analysts who should be interpreting the results.
And the credibility hit is the worst part. When three versions of patch compliance appear in one meeting, the audience stops trusting the whole set rather than adjudicating between them.
Ask two questions of every metric. Who acted on this in the last twelve months, and what goal or obligation does it serve.
Metrics that fail both go. Metrics that fail the first but serve a real obligation move to a context layer that is retained but not reported.
Expect the executive set to come down to somewhere between six and ten. Expect resistance to be strongest for the metrics with the least evidence of use, because those are the ones defended on principle rather than on record.
Then install the thing that prevents recurrence: a rule that no metric is added without a named consumer, and a scheduled review that retires dormant ones by default. See metric governance.
Metric Maestro records which metrics are actually consumed and when each was last computed and viewed, which turns the annual cull from an argument into a query.
Because definitions are held centrally rather than embedded in individual reports, duplicates with divergent scopes are visible before they reach a meeting rather than after.
From the blog