Carrier performance metrics are per-carrier measures of delivery outcome, speed, reliability and cost, computed from the events your own systems hold: the promise shown at checkout, every tracking scan, the carrier invoice and the claims log. The distinction from the carrier's report is the clock. A carrier's service standard starts when the carrier scans the parcel in, and "delivered" includes delivery to a pickup point the customer did not choose; your customer's clock started at checkout, when a date was shown. The gap between the two clocks is where handover delays, missed cut-offs and pickup-point substitutions hide. Owning the scans means you can measure the promise you made, not the standard the carrier sells.
Most brands know what they pay each carrier and have a feeling about which one is good. Few have a number. This page is the scorecard: eight metrics, how to compute each from data you already have, and what to do with the result.
Short answer: run eight metrics per carrier, per service and per destination country, computed from your own tracking scans: on-time delivery rate against the promise you showed, handover time from label to first carrier scan, first-attempt delivery rate, exception rate, damage and loss rate, return-leg transit time, cost per parcel including surcharges, and claim resolution time. Compute them monthly, compare carriers on the same lanes, feed the result into routing rules, and bring it to rate negotiations. The carrier's own report measures against the carrier's own definitions; yours measures against what you promised the customer.
Free for DTC brands: Pango runs the shipping and returns audit on your own order and carrier data as part of a demo: fifteen numbers across checkout, carriers, warehouse handover, tracking and returns, the three that matter most for your operation, and the same orders shown running on one system. Book the free audit.
Why your own data, not the carrier's report
A carrier's performance report is built on the carrier's clock: the service standard starts when the carrier scans the parcel in, and "delivered" includes delivery to a pickup point the customer did not choose. Your customer's clock started at checkout, when a date was shown. The gap between the two clocks is where handover delays, missed cut-offs and pickup-point substitutions hide. Owning the scans means you can measure the promise you made, not the standard the carrier sells.
It also means one definition across carriers. Two carriers will never define "attempted" or "exception" the same way; normalising their statuses onto one set is the precondition for comparing them. The status vocabulary problem is visible in any two of the carrier guides on this site, for example Royal Mail and DHL.
The eight metrics
| Metric | Definition | Computed from |
|---|---|---|
| On-time delivery rate | Delivered on or before the date shown to the customer, divided by delivered | Checkout promise date and the delivered scan |
| Handover time | Hours from label creation to the carrier's first physical scan | Label timestamp and first scan; long gaps look like "label created, not moving" to the customer |
| First-attempt delivery rate | Delivered on the first attempt, divided by delivered (home delivery and pickup point measured separately) | Attempted and delivered scans |
| Exception rate | Parcels with any exception scan (delay, address, damage, customs, hold) divided by shipped | Normalised exception statuses |
| Damage and loss rate | Parcels reported damaged or never delivered after the carrier's search, divided by shipped | Claims and tracking outcomes |
| Return-leg transit time | Days from the customer's return drop-off scan to warehouse receipt | Return label scans and dock receipt |
| Cost per parcel | Invoiced amount including fuel, remote-area, oversize and re-delivery surcharges, divided by parcels | Carrier invoices matched to shipments |
| Claim resolution time | Days from claim filed to paid or rejected | Claims log |
Two of these deserve a note. Handover time is the metric the customer feels first and the carrier's report never shows, because the carrier's clock has not started. Cost per parcel is only honest with surcharges included; a cheap rate card with heavy remote-area and re-delivery fees is not cheap.
Cut the scorecard by lane
A carrier is not good or bad; it is good on some lanes and bad on others. Cut every metric by service (home, pickup point, express), by destination country and, where address data allows, by destination type (urban, rural). The useful output is a lane table: for each country and service, the best carrier by on-time rate at what cost per parcel. That table is the input to carrier selection rules such as "route to the carrier with the highest measured on-time rate to this country last month, unless its cost per parcel is more than X above the alternative".
How to use it
Routing. Refresh the lane table monthly and let it drive the performance-based rules. A carrier whose on-time rate to one country slipped last month should lose that lane automatically, not after a quarterly review.
Negotiation. A brand that can show a carrier its own handover time, exception rate and surcharge total per lane is negotiating on facts. One Pango customer uses its analytics as a third-party source of truth in carrier renegotiations for exactly this reason; the analytics module covers warehouse handover time, warehouse and carrier performance, lost and delayed packages and carrier attribution.
Customer promise. Measured on-time rate per lane is what an honest delivery promise at checkout is built on. How the promise is computed is in what is order promising.
Support. Exception rate and first-attempt rate predict ticket volume. If one carrier's exception rate on one lane doubles, the WISMO tickets follow within days; the mechanics are in reduce WISMO tickets.
Failed deliveries. First-attempt rate and attempt-to-outcome are their own subject, with the seven causes a brand controls, in how to reduce failed deliveries.
Building it
- Store every tracking event for every shipment, with the carrier's raw status and your normalised status, timestamped.
- Store the promise shown at checkout on the order.
- Match carrier invoices to shipments at line level, including surcharges.
- Log claims with filed and resolved dates.
- Compute the eight metrics monthly, per carrier, service and country.
- Publish the lane table to the people who write routing rules and negotiate rates.
The reason this is rare is not analytical difficulty; it is that the events live in the carrier's system, the promise in the checkout, the invoice in finance and the claim in an inbox. Pango collects all four on the order record because it runs the outbound routing, the label, the tracking and the return on the same system, and connects to more than 100 carriers through prebuilt connectors with every carrier's statuses normalised onto one set. The category is described in carrier management system, and the module at delivery management.
The category term for that layer, and how to buy it, is covered in carrier management software.
What a scorecard looks like
A carrier is not good or bad; it is good on some lanes and bad on others. So the scorecard is not one row per carrier but a lane table: one row per carrier, service (home, pickup point, express) and destination country, with the eight metrics as columns and the month as the period. Where address data allows, add destination type (urban, rural). The useful reading is per lane: for each country and service, the best carrier by on-time rate at what cost per parcel, and which carrier's exception rate moved since last month. That table is the input to carrier selection rules such as "route to the carrier with the highest measured on-time rate to this country last month, unless its cost per parcel is more than X above the alternative".
A carrier-mix dashboard is the same table with two additions: the share of volume each carrier carried per lane, and what the blended cost and on-time rate would have been under a different split. That is the view that answers "should we move the Netherlands pickup-point lane from carrier A to carrier B", and it only works on your own data, because the carriers' reports do not share lanes or definitions.
Which platforms offer carrier scorecards?
By their own descriptions, the post-purchase vendors in Pango's competitor research describe branded tracking, notifications, exception management, returns and AI agents; none of the research's recorded descriptions is of a carrier scorecard or a carrier-mix dashboard, so this page does not rank them on one, and no platform is named here as offering the "most reliable" scorecard. The practical test for any platform is the one in the previous section: can it hold the promise shown at checkout, every scan with a normalised status, the invoice at line level and the claims log on the same order record, per lane and per month? A tracking-only platform holds the scans; a shipping app holds the labels and sometimes the invoice; a returns tool holds the return leg. The scorecard needs all four, which is why it is rare. The category that is supposed to hold them is described in carrier management software, and the connection work that feeds it in what is carrier integration.
For managing carrier contracts specifically, the software question is secondary to the data question: a contract review runs on cost per parcel against the contracted rate, per service and zone, with surcharges itemised, and on the on-time and exception rates per lane. Whatever holds those four numbers is the contract tool.
How Pango uses them for routing
Pango collects all four sources on the order record because it runs the outbound routing, the label, the tracking and the return on the same system, and connects to more than 100 carriers through prebuilt connectors with every carrier's statuses normalised onto one set. Its shipping module does multi-carrier orchestration, routing, rate shopping and shipping rules, so the lane table is not a report that a person reads and then edits rules from; the measured on-time rate, exception rate and cost per parcel per lane are inputs the routing rules can read. Its analytics cover warehouse handover time, warehouse and carrier performance, lost and delayed packages and carrier attribution, which is the set a live customer uses as a third-party source of truth to renegotiate carrier terms. The module is at delivery management.
The bottom line
Eight metrics, computed from your own scans and invoices, cut by lane, refreshed monthly, and used in routing and negotiation. That is the whole discipline. The hard part is getting the events, the promise, the invoice and the claim onto one record; once they are, the scorecard is a query. The shipping and returns audit starts with which of the four you have today, and a demo shows the lane table built from live data.
