> How do you reconcile the two?

This sounds like a riddle / koan.

Business analysts pulled these metrics from a data warehouse and the semantics don't line up.

I've worked on growth and data science teams, and there ought to be a book "How to lie with data science" as a successor to "How to lie with statistics".

There's been situations where I've seen the same metric queried in two different ways with contradictory results. The slight of hand here is that the query is query, and unless the metric is defined as the query, the metric is not the query.

I remember Nassim Taleb proposing that books are fractals and can't be summarized / compressed.