JB Analytica reference architecture — a published example. The business and every figure on it are invented.

Customers

The same realised revenue again, at the grain a business actually asks about: per customer, all time. lifetime_realised_net_revenue_eur is a column on the dim_customers mart, summed there from the same order lines the performance page totals by month — so a customer's lifetime value and the year's revenue cannot disagree, because neither is worked out twice. dbt/tests/assert_marts_reconcile.sql fails the build if they ever do.

These people do not exist. The names, the cities and the email addresses are generated to look Belgian and nothing more; there is no customer behind any row on this page.

The table and the chart above both filter on has_ordered, which is a mart column rather than a condition this page invents: 79 of the 600 customers have never placed an order. That is not a gap in the data — it is what a customer table looks like, and Architecture explains why the generator was left alone to produce it.