A useful measure can change its meaning

A university may track the number of visits to its advice office as an indicator of demand. An increase seems straightforward: more visits must mean more need. Yet the relationship can change when the service changes how people gain access.

Suppose the office introduces an easy online booking system and extends its opening hours. People who previously found it difficult to attend may now make appointments. The rise in visits could reflect improved access rather than a sudden increase in underlying problems. Conversely, fewer visits after a complicated booking rule would not necessarily show that students need less support.

The measure still provides useful information about workload. What changes is the strength of the inference from workload to need. To interpret the pattern, staff may compare waiting times, unsuccessful booking attempts, and reports from students who do not attend.

This example illustrates why an indicator should be reviewed when the system around it changes. A number can remain accurately counted while becoming a less direct answer to the original question. Good evaluation preserves the distinction between the observation itself and the explanation attached to it, especially when a new procedure influences who appears in the records.