Data Health & Reconciliation

• Source: Housecall Pro API (read-only), pulled live and cached 30 min.
• Bucketing: a job counts in a month by its scheduled date.
• Canceled jobs are excluded from sales (matches HCP's revenue reports).
• Tech credit: revenue & job count go to the job's assigned technician.

Verdict Only 8 of 6,726 jobs all year are missing an assigned tech — HCP is being used correctly for tech assignment. Those 8 jobs ($219) are the only revenue that can't be credited to anyone.

Our YTD $712,240 / 288 jobs vs HCP $712,837 / 294 jobs. Dollars match to 99.9%. The remaining ~6 jobs / $597 (about $100 each) are small jobs HCP's report counts but we don't — almost certainly because the owner's report buckets by completed/invoiced date while we bucket by scheduled date, so it picks up a few jobs scheduled in late 2025 or left unscheduled.

To match exactly switch our bucketing to the same date type the owner's report uses (Completed date). The data is clean — this is a settings alignment, not a Housecall Pro problem.

Our live pull counts every non-canceled job credited to the assigned tech, bucketed by scheduled date. HCP method applies the same rule the owner's revenue report uses for cancellations (canceled work removed). The difference is the canceled jobs each tech had on the books.

Heads up Canceled-removal is only part of the story. On the Miguel spot check, his printed YTD report is $712,837 — about $88K below even his canceled-excluded $801,104 — because the owner's report also buckets by a different date (Completed/Invoiced). To make every tech tie out to the printed report exactly, confirm that date type and I'll add a third "printed report" column.

The old totals counted canceled jobs; the corrected totals exclude them (matching Housecall Pro). Below is every field tech and what changed. Same jobs, same attribution — the only difference is canceled work removed.