This question is not about technology but more about methodology.

If the business is trying to monitor unique customer visits (by time, by loc etc), rather then customer visits.

If we have John Doe visiting in January and he visits again in February, it is one Unique Customer visit in January and one Unique Customer visit in February.

When we look at this aggregated to Calendar Quarter, isn't the software aggregating it to two (2) Unique Customer visits? But the business views this as one Unique Customer visit in the Quarter.

This is the same as Bank Accounts Balances. They are non additive from month to month, but we can scorecard it and show report on year level.

I have to remind myself, that when we scorecard, we are really not reporting the actuals, we are reporting on actuals compared to targets. In this case, even if the aggregation is incorrect, the comparesement to target (that is also aggragated) will be correct on the Quarter or Year level

Am I explaining this correctly or just misleading myself to an answer that I like?

Thank you,

You're explaining it right. These special types of calculations are usually for counts, and the problem you're trying to avoid is commonly called "double counting."

In WebFOCUS there's a feature called Distinct Count that allows you avoid double counting; using a hidden field that is silently compared when doing the sum (say an equipment code or a patient or customer code) and the counting only spreads across time breaks or other Dimensional breaks when the values in this field don't match.

There's a feature in development for PMF that supports the no-double-count requirement, called Distinct Count Measures. It doesn't use the Distinct Count feature (it can't for these measures, because for relational efficiency PMF performs all filtering on the DB side, and this requires a hidden filter).

The next GA release - 5.2 in the Fall - will have this new capability.

thanks