Missed Lead Pipeline Monitor
Data/automation health for the Stage A → Stage B missed-lead pipeline: Stage A reviews call transcripts and flags candidate leads; Stage B (Cloud Run Job, twice daily) attempts to create each one in Cubby. This tracks pipeline health and outcomes — not lead conversion (see the Leads Monitoring dashboard for that).
Time period:
Pipeline Health
Leads Created Over Time
Stage A flagged vs. Stage B created, per day
Every day in the selected period is plotted, even with zero activity, so the timeline never skips a day. Stage A = candidates flagged from call review (whether or not Stage B ever created them); Stage B = leads actually created in Cubby. Filter by facility below, or switch on the status breakdown to see what Stage B did with each day's prospects -- anything it hasn't gotten to yet shows as "pending".
Lead Outcome / Usefulness
Current Cubby status of Stage-B-created leads
Live join to cubby_replica.leads_flagged by cubby_lead_uid — current status only, not a trend over time. Newest 20 leads per page.
—
Conversion rate (14d+ old)
—
Eligible leads (14d+ old)
—
Resolved a current Cubby status
| Caller | Facility | Lead ID | Lead UID | Lead created | Age (days) | Current Cubby status |
|---|
Known limitation. This is current status only, not a trend — a leads_status_history snapshot table
(populated by a small daily job) would be needed to compare, e.g., week-1 vs. week-3 conversion. Deferred as
v2 of this dashboard. Also note: pending backfill rows just haven't had
cubby_lead_uid
filled in yet (it lands progressively, within ~8h) and are expected; one still pending after a day usually means Cubby's
replica row for that lead has no facility or created date, so Stage A's match can't find it — only rows showing unresolved
(a cubby_lead_uid is set but no matching Cubby lead was found) are worth a data-pipeline check.
Drill-Down / Worklist
All flagged candidates
Every candidate Stage A flagged in the selected period, with Stage B's outcome. Click a row to see the call that triggered it. Click a column header to sort. 20 rows per page.
| Identified | Caller | Facility | Tier | Bucket | Signal | Status | Attempts | Reason |
|---|
Outcome Breakdown
Stage B status, last 30 days
Every candidate Stage A flagged in the selected period, by what Stage B did with it.
Why leads were skipped or failed
last_error grouped by known pattern (skipped / failed / existing-tenant rows, selected period).