Audience by purchase frequency
Clients who bought N times in a window, with an optional value floor per purchase — the audience the QON builder can't express, because its RFM is lifetime (no window) and has no notion of a purchase day. Answers "who bought 3+ times over $X this semester", "high-value repeat buyers", "spent more than $Y since January", in a single call instead of paging orders and aggregating client-side.
An occurrence is an order (groupBy=order) or a purchase day
(groupBy=purchase_day), where the tickets of the same day are summed —
retail splits one purchase across several tickets, so the day, not the order, is
the unit the campaign means. On Cloe's 2026 this is 14% of the audience.
valueMode sets the floor: all = every occurrence ≥ valueMin (recurring
high value); any = at least one; sum = the window total. valueMin is in
cents, inclusive. minOccurrences is the minimum count of orders/days.
Per row: occurrences (count), value (window total), minValue/maxValue
(smallest/largest occurrence) and lastDate — the diagnostics that let you show
why each client qualified. Values are in cents; the window and the day
are cut from order_date in the org timezone, like the sales report. Rows
are paginated (e/p); total is the full audience size.
This is the order side only — client id + metrics. Name, email, phone,
last store and store group live in register; so do the "exclude these
experiment lists" and "must have contact" filters. Enrich and filter with the
register operations, and close the loop with
register.audience-list.person.create (which takes the client ids) — no CSV in
the middle.
Try-it values populated from environment ACME (Demo)
Authentication
Token
Send your user token in the x-guper-authorization header.