# Fulfilment operations decision report

## Decision

Review Harbour express first, then Eastgate. Harbour has 32 late orders among 55 mature orders (58.18%), and its express cell is 18/18. Across all hubs, 90 of 197 mature, non-cancelled orders are late: 45.69%.

| Hub | Mature | Late | Late rate | Open overdue |
| --- | --- | --- | --- | --- |
| Eastgate | 55 | 31 | 56.36% | 3 |
| Southbay | 32 | 7 | 21.88% | 1 |
| Harbour | 55 | 32 | 58.18% | 3 |
| Westhill | 55 | 20 | 36.36% | 3 |

## Denominator and completeness

The 216 orders comprise 12 cancelled, 7 not yet due, 107 on-time deliveries, 80 late deliveries and 10 still-undelivered overdue orders. These cohorts sum to 216. Mature denominator = 107 + 80 + 10 = 197; late numerator = 80 + 10 = 90. Using 90/216 understates lateness. The seven not-mature orders are not completed successes.

## Actions

First inspect the ten overdue_open orders, sorted by due_at then order_id, and open each event timeline. Confirm whether delivery is still outstanding or whether the delivery feed is incomplete. Follow the hub operations process after confirmation; this workbench does not dispatch vehicles or send customer notifications.

For Harbour express, compare collection-to-delivery intervals with scheduled departures and handover scans. The route notes suggest queueing or address correction as leads. They do not prove the cause of a delay or establish employee responsibility. Costs and refunds are absent, so no loss estimate is supported.

## Review limits

The cutoff is 2026-09-15 00:00 UTC. Future deliveries are excluded. Delivered lateness and open overdue minutes remain separate, because an unfinished order does not have a completed delivery duration. Southbay has a different service mix and a small express sample; use task four before comparing hubs. In the next review, preserve this historical snapshot and introduce a new observation date rather than rewriting the earlier decision with later evidence.