Have you considered instead of evaluating a median, evaluating a % of
orders over a certain threshold?For example you could export your
revenue at transaction id level for a given month via data warehouse,
use excel to derive a median, and then use that as your
benchmark.(Visit: Revenue > 200 & Orders Exists [visits]) / (Visit:
Orders Exists [visits]).Although not exactly what you are hoping to find
it will give you similar directional value.