POWER BI IMPLEMENTATION STARTER — NOT A PUBLISHED PBIX REPORT
Import output/monthly.csv, output/cohorts.csv and output/countries.csv into Power BI Desktop.
These are separate aggregate tables: do not join them on country or customer.
Monthly grain = one row per calendar month. Cohorts grain = one cohort + relative month.
Suggested visuals: monthly value line; order trend; AOV KPI; country contribution; cohort matrix.

DAX measures (Monthly table):
Positive Invoice Value = SUM(Monthly[positive_value])
Orders = SUM(Monthly[orders])
Average Order Value = DIVIDE([Positive Invoice Value], [Orders])

Cohort matrix cell (rows: cohort; columns: age):
Retention % = DIVIDE(SUM(Cohorts[active_customers]), SUM(Cohorts[cohort_size]))
Do not interpret grand totals across different ages as a retention rate. Disable matrix totals.
Format retention as Percentage; all monetary values as GBP.
Use the definitions and limitations in project-methodology.txt in the report information page.