The intelligence layer
Most “commission tools” just diff two spreadsheets. CommissionSight is built around a delta engine that scores every member against prior periods — month over month, quarter over quarter, and year over year — and pinpoints exactly who continued, whose commission moved, who dropped, and who reappeared after a gap. The result is an auditable book, surfaced identically in the app grid and over the API.
What you get
- A stable identity per member, so the same person matches across every file — and reappearances are caught even after a long gap.
- A status for every member against your chosen baseline:
| Status | Meaning | Flags it can carry |
|---|---|---|
| 🟢 green | Continuous and unchanged | NEW (first appearance) |
| 🟡 yellow | Present, but something changed | COMMISSION_CHANGED, DATA_CHANGED, REAPPEARED, REAPPEARED_WITH_DELTA |
| 🔴 red | Was present, now absent | DROPPED |
CHARGEBACK is a cross-cutting flag layered on top of any status (any policy clawed back this
period). See Commission owed & chargebacks.
- Per-field deltas (
prevValue → currValue) so a yellow is never a black box — you see exactly what moved and by how much. - Churn you can trust — every scored period is guarded by data-quality signals: a partial-statement flag (a half-loaded file can’t masquerade as mass attrition) and identity-change detection (a carrier re-issuing its member ids — plan-year changes, system migrations — is called out as an id change, never reported as your whole book dropping). See job warnings.
- Any baseline — prior month (default), prior quarter (QoQ), prior year (YoY), or any two periods
you pick via
/comparisons. - Explainable statuses — every flag reconstructs to a plain-English reason from the API fields, so
an auditor can trace any member straight to the statement it left on (
GET /v1/members/:id/last-seen).
The same scored dataset powers the color-coded UI grid and these endpoints, so the app and your own systems always agree.