Fictional lab material prepared for the MS-4004 Copilot exercises. Not a real investigation.
To: Prairie Region Operations From: Network Engineering, Western Date: 2026-04-24 Subject: Calgary depot — Tailwind Traders on-time investigation, findings Classification: Internal
What we were asked
Establish why the Calgary distribution centre missed the 95% on-time threshold on Tailwind Traders volume in all twelve months of FY2025, and say what would move it.
Calgary averaged 90.2% on-time across FY2025 against that 95% requirement, ranging from 86.7% in July to 92.1% in October — the weakest site in the network in every month of the year. The month-by-month record is in CanadaPost_Branch_Performance.xlsx.
What we found
Induction timing is the largest single factor we can measure.
The Calgary sortation cut-off for next-day processing is 16:30. Across the sample period, 38% of Tailwind drops arrived after it. A parcel inducted after the cut-off is worked on the following processing day, which consumes the whole of its delivery margin before it has moved.
When we hold induction time constant across sites, Calgary's on-time performance is within 1.4 points of Richmond and Moncton. The gap is not the depot's processing rate.
What we ruled out
| Hypothesis | Result |
|---|---|
| Volume exceeds Calgary's capacity | Ruled out. Mississauga runs nearly twice Calgary's parcels at the network's best on-time rate. Calgary is at 61% of its rated line capacity. |
| Depot processing is slower than other sites | Ruled out. Sort-to-dispatch time is within half a point of the network mean. |
| Address quality in the Calgary catchment | Not supported. Address-related exceptions are 0.4 points above the network mean, which is not enough to account for a gap of this size. |
What we could not establish
- We cannot say induction timing is the only cause. It is the largest factor we could isolate. Roughly a third of the gap is unexplained by anything we measured.
- We did not examine winter road conditions systematically. Calgary's worst months are not consistently the coldest ones, but we have not tested this properly.
- We have no data on why the drops are late. Whether this is Tailwind's dispatch schedule, their carrier to the depot, or something else is outside what we can see. Someone has to ask them.
What we recommend
- Move the working cut-off conversation to the customer. The fix is theirs to make; we can only make the consequence visible.
- Record induction time on every Tailwind drop, weekly, from now on. Without it, nobody can tell whether anything improved.
- Do not commit to 95% by a date. On this evidence the number moves when the drops move, and we do not control the drops.
Contact: Network Engineering Western. Findings are based on FY2025 data and the first four weeks of recovery monitoring.