Compare parcel delivery performance with the same events and shipment mix

Write the comparison rule before looking at the average

A route can look faster because its report starts later, includes different destinations or omits parcels still in transit. Before a carrier or consolidator review, define the question: warehouse dispatch reliability, time before postal acceptance, or postal delivery performance. Give the review a date, eligible shipment population and target tied to a named event.

MIT CTL’s public abstract of a 2024 exploratory European omnichannel retail study discusses the breadth of fulfilment costs, including inbound work and storage. It is background for considering the complete service scope; it does not establish US carrier rankings. The cohort method here is our practical review proposal, not a result from that study.

Separate label creation, physical handover and postal acceptance

USPS distinguishes a label created while awaiting the item from USPS possession. It also describes acceptance pending as receipt of a container or pallet before individual processing starts. Keep the original event text, timestamp and source. A batch receipt, a warehouse dispatch statement and an individual parcel scan should remain separate observations.

For China–US movements, retain the international reference and the link to the final parcel. If the question is about the full customer journey, begin from the agreed order or dispatch event and include the relevant international stages. If it is about USPS domestic delivery, define the postal start event and disclose parcels without that evidence. Neither clock should be presented as the other.

Build comparable groups before combining results

Group parcels by the contracted product, origin or gateway, destination group, dispatch period and the handling characteristics relevant to the review. Keep important differences in packed size, goods acceptance or delivery requirements visible. Record sample counts instead of allowing one large group to hide a small group with unresolved cases.

USPS’s delayed-mail help page explains that delivery standards differ by mail class and lists non-guaranteed standards and destination-specific conditions. Use the applicable purchased service and day-counting rule when agreeing a target. A reported expected date is not a universal guarantee or a China-to-door commitment.

Use a worksheet that preserves unknown results

The downloadable parcel cohort worksheet contains a header and one blank row. Create one row for each parcel included in the agreed review; parcel_reference can be an internal reference. Keep cohort_reference, route_reference, product, destination_group, booked_at, start_event, start_at, target_at, calendar_reference and review_as_of with the record.

Add delivered_at only from the applicable recorded event. Use outcome for the result at the review point, latest_event for the exact status, evidence_source for the supporting record and exception_owner for the next action. Keep personal addresses and customer contact details in the approved private system rather than circulating them in a comparison sheet.

Report pending and unknown parcels beside timely deliveries

Use mutually exclusive outcomes for the chosen target: delivered by target, delivered after target, past target and still open, not yet due, and unknown because required timing evidence is missing. Record how returned, cancelled or intercepted parcels are handled before starting; if excluded, show the reason and quantity separately.

An illustrative review has 100 eligible parcels: 80 delivered by target, 10 delivered after target, five past target and still open, three not yet due and two unknown. The delivered-only result is 80 of 90, or 88.9%. The all-eligible by-target share is 80 of 100, or 80%. These are different descriptions of this invented sample; a complete assessment must also explain the ten parcels outside the delivered group.

Do not label an immature cohort’s current share its final performance. Compare the counts at a stated review date, or wait until the agreed observation window closes while continuing to report unresolved evidence. The example sets no carrier benchmark or contractual service target.

Keep elapsed time, cost and exception work connected

If reporting a delivery duration, state its events and calendar and identify which parcels had both timestamps. Show the completed sample alongside the open queue. Investigate where a gap occurred: order release, international transfer, postal acceptance, delivery attempt or return handling. A missing scan alone does not identify the party responsible for delay.

Review invoiced charges, rework, replacement and return costs on the same cohort and agreed charging basis. Keep a quotation separate from the actual invoice. A route change should consider customer information and exception ownership alongside observed timing; a network announcement alone supplies none of these shipment results.

Keep reporting lag separate from parcel transit

If reviewing information delays, record the event occurrence time, source recording time and partner receipt time as separate observations. State which pair of times the interval uses. An update received late can describe an earlier movement; it does not itself identify a longer physical journey. Retain unresolved time-zone or status mappings for review.

Agree the next test and preserve the original result

Record the proposed change, owner, eligible trial group, target definition and review date. Keep the original record when a later scan arrives, with an updated observation rather than an erased earlier result. Bring exact carrier events and supporting handover records to the responsible provider for unresolved cases.

For Jeton Express, prepare the route, order profile, parcel references and agreed service scope for review. This worksheet does not fetch tracking events, calculate a contractual entitlement or certify carrier performance. Use the linked USPS guidance for actual status-specific actions and the existing route-review guide for a proposed network change.