Endpoints
Metrics glossary
Every metric has akey that you pass as the {metric} path segment to Retrieve Metric. List Metrics returns the same catalog at runtime, with each metric’s unit and the properties you can filter or break down by. An unknown key returns a 404.
Units
Filtering and breaking down
Each property works two ways unless noted otherwise:- Filter — pass
property=valueto narrow the series to a single value, for example?payment_method=card. - Break down — pass
breakdown_by=propertyto split each point into one entry per value. For example,?breakdown_by=currencyreturns an entry forusd, an entry foreur, and so on. Onebreakdown_byat a time.
400.
Query parameters
These apply to every metric on Retrieve Metric, independent of its properties:Revenue
Payments
Fees
Every fee metric accepts acurrency filter.
Disputes & refunds
Authorizations
Members & users
Partners
Advertising & traffic
Snapshot metrics
Snapshot metrics are captured once per day and preserve what was known at that
time. They are day-only and UTC-only:
interval must be day and time_zone
is not accepted. You still pass from and to to choose the range of daily
points. Use each metric’s windows catalog field to determine which
snapshot_window values it accepts.Cohorted dispute snapshots
Cohorted metrics attribute each dispute to the date of its original paid receipt. The attribution window in the metric name answers “how long after purchase can a dispute count?” Thesnapshot_window answers a separate question: “how many consecutive days of sales are pooled into this point?” Use 7d for a more reactive series or 28d for a smoother denominator.
Each captured point uses the latest sales cohort whose complete attribution window has elapsed. For a snapshot captured on date D, its cohort ends on D - attribution days - 1 day; the extra day ensures the final attribution day is complete. To reproduce a cohort-end-date chart, plot each point on that inferred cohort end date. Comparing 7d, 14d, and 28d attribution values at the same raw API timestamp compares different mature cohorts; align them by cohort end date before expecting 7d ≤ 14d ≤ 28d.
