Carrier Performance: How to Measure Carriers and Act on It

Carrier performance is how reliably a shipping carrier delivers your parcels on the date you promised, with the scans, condition, and cost you agreed to. For a merchant, it is best measured on a weekly scorecard of your own shipments, graded against the checkout delivery date, not the carrier's commit time.

I'm Jevon Le Roux, and I run Keeyu, which builds proactive e-commerce operations software. My view is simple: carriers will always drop the ball, and brands can't control carriers. You can measure them on your own data, hold them to the terms you signed, and make sure the shopper hears about a problem from you first. Every order is a promise. Keeyu keeps it.

On-time delivery metrics: what carrier performance really measures

On-time delivery rate is the share of shipments that arrive on or before a committed date. The catch is whose date. A carrier measures itself against its own commit time. You should measure it against the date you showed the shopper at checkout.

The carrier owns the movement, but you own the promise and the visibility. If your product page said "arrives Thursday" and the parcel lands Friday, the shopper does not care that the carrier hit its ground commitment. You broke a promise, and the carrier was the instrument.

Why carrier on-time numbers are not comparable to yours

Published on-time figures are useful context, but they are graded against the carrier's own clock. ShipMatrix's report on the 2025 peak season covers December 1 to 31, 2025, across all parcel services combined. It put on-time performance at 95.3 percent for FedEx, 97.2 percent for UPS, and 94.1 percent for USPS, measured against each carrier's actual commit times.

The same report notes that UPS added an extra day to more Ground lanes than FedEx during peak. So a parcel that would be late against a FedEx commitment could count as on time for UPS.

When a carrier extends its commit window, its score goes up and your promise does not move. Track both numbers, but run your decisions on the one tied to your checkout promise. For the full method of setting and hitting that date, see my guide to improving on-time delivery rate.

Three dates to capture on every shipment

• Promised date: what the shopper saw at checkout or in the confirmation email.
• Carrier commit date: what the carrier's service level says, from the label or rate response.
• Actual delivery date: the delivered scan.

With those three you can tell whether a late order was a carrier miss, a promise you set too tight, or a dispatch delay on your side.

Carrier scorecards and analytics: the metrics that matter

A carrier scorecard is a fixed set of metrics, calculated the same way every week, that shows how each carrier is doing for your volume on your lanes. It turns "the carrier has been bad lately" into a number you can take into a meeting.

I would measure orders in breach of the promise you actually sold and parcels with no scan after a set threshold. I would also count affected customers who never opened a ticket. That last one matters because silent shoppers are the ones who leave. The scoreboard should sit alongside traditional support metrics, not under them. Issue rate and fulfillment rate tell you more about carrier health than first response time ever will.

The scorecard

Segment before you compare

A single blended number for each carrier hides most of the useful signal. I'd run separate metrics by dispatch type, because a pallet delivery and a small parcel have nothing in common on transit time or damage risk. Split by service level, by origin warehouse, and by destination zone. Then compare carriers only within the same segment.

Set thresholds, not just averages

Averages tell you what happened. Thresholds tell you what to do. Pick a no-scan threshold, such as 48 hours with no movement on a ground shipment. Treat every breach as a live issue, not a line in next month's report. The carrier will not tell you a parcel has stalled. You have to see it.

Worked example

Here is an illustrative month for a brand shipping 10,000 orders across two carriers. The numbers are invented to show the math, not drawn from any real carrier.

• Carrier A: 6,000 shipments, 5,520 on time to promise (92.0 percent), 240 no-scan gaps over 48 hours (4.0 percent), 60 lost or damaged (1.0 percent), spend $54,000.
• Carrier B: 4,000 shipments, 3,840 on time to promise (96.0 percent), 80 no-scan gaps (2.0 percent), 20 lost or damaged (0.5 percent), spend $40,000.

Carrier A looks cheaper per label: $9.00 against $10.00. Cost per on-time delivery tells a different story. Carrier A is $54,000 / 5,520 = $9.78. Carrier B is $40,000 / 3,840 = $10.42. The gap shrinks from $1.00 to $0.64.

That is before you price in the 40 extra lost or damaged parcels, the tickets from 160 extra stalled shipments, and the shoppers who quietly never come back. Each lost parcel means a claim and a reship decision, which my guide to lost in transit claims, reships and refunds walks through.

Now split by lane. Say Carrier A's misses cluster on two western destination zones, where its 90th percentile transit is a full day longer than Carrier B's. That is not a reason to fire Carrier A. It is a reason to move those two zones and keep the rest.

SLA accountability: using the scorecard in disputes and renegotiation

Service level agreement accountability means holding a carrier to the delivery, scan, and liability terms in your contract. It takes evidence specific enough that the carrier cannot wave it away.

You can staff agents to answer where is my order forever while the carrier black hole stays uninstrumented. That is backwards. The data you need to hold a carrier accountable is the same data you need to stop the tickets. Instrument the promise, and the dispute file builds itself.

Money-back guarantees and late-delivery refunds

A money-back or service guarantee is a carrier promise to refund shipping charges when a parcel misses its commit time. Terms vary by carrier, service, contract, and ship date. Read your carrier agreement and the carrier's current published terms before you build a refund process. Never assume last year's terms still apply.

Where a guarantee or contractual SLA does apply, the claim usually hinges on the carrier's commit time, not your promised date. That is why the scorecard tracks both.

Building a dispute file

• Shipment ID, service level, and carrier commit date for every late parcel.
• Full scan history, with gaps flagged against your no-scan threshold.
• First-scan time, so the carrier cannot blame a late handover if the label was scanned at pickup.
• Photos and packing records for damage claims.
• Monthly totals by lane, so the conversation is about patterns, not one bad parcel.

Renegotiation and volume allocation

Your scorecard is a negotiating document. If one carrier is consistently stronger on a lane, that is leverage on rates for the others. Allocation decisions should follow cost per on-time delivery and lane-level transit, not the headline label rate. Move volume a zone at a time, measure for a few weeks, and keep the change only if the numbers hold.

Every order is a promise. Keeyu keeps it.

Comparative carrier data: benchmarks across carriers and lanes

Comparative carrier data is any measure that lets you rank carriers against each other, either from public reports or from your own shipments split by carrier. Both have limits, and the limits matter when you are deciding who gets volume.

Public numbers are a starting point, never the decision. As the ShipMatrix peak report shows, carriers change commit times during peak, which moves their scores without changing the parcel's real arrival. Your own scorecard, measured against your promised date on your lanes, is the comparison that counts.

Public sources worth reading

• The ShipMatrix 2025 peak report also gives the 2024 peak results. FedEx was at 91.8 percent, UPS at 96.5 percent, and USPS at 90.4 percent, so you can see the year-over-year swing.
• USPS posts quarterly service performance results by mail class, plus a dashboard that shows performance by area. Note that its Package Services class means Bound Printed Matter, Media Mail, and Library Mail, not most e-commerce parcels.

Running your own head-to-head

The cleanest comparison is a controlled split. Route a share of shipments on the same lanes and service level to two carriers for a set period. Then compare the scorecard rows side by side. Peak season is the best time to learn and the worst time to switch, so test before October, not during it. For software that pulls this data together across carriers, see our ranking of 3PL and carrier visibility tools.

Customer experience impact: what late and lost parcels cost you

Customer experience impact is the effect a carrier failure has on what the shopper does next: contacting support, leaving a review, or buying again.

When a carrier drops the ball, the question is who tells the shopper first. Every time a shopper has to go find out something, you've already added friction. Brands can't control carriers. They can control communication.

Tickets are the visible part

A stalled parcel creates a where is my order ticket a few days later. Late dispatch and carrier black holes destroy customer lifetime value while ticket dashboards still look fine. Silent churn runs ahead of visible tickets. The shopper who never writes in is the one you lose without noticing, which is why the scorecard counts affected customers who never opened a ticket.

What to do with a breach

• Detect it at the threshold, not when the shopper complains. A missed first scan or a 48-hour gap is the signal.
• Tell the shopper before they ask, with a revised date and what you are doing about it.
• Fix it with the carrier in parallel: open the trace, file the claim, or reship.
• Log the outcome against the carrier and lane so it shows up on next week's scorecard.

Promise with confidence

When you can see whether each order is tracking against the promise, you can advertise faster delivery commitments with more confidence. You will know early when one is slipping. That is what proactive e-commerce operations means to me. You stop waiting for the complaint and start working the order the moment the data says it is at risk.

Keeyu gets shoppers what they want, on time, as promised. For the wider operating model, start with our guide to proactive e-commerce operations. To see how Keeyu puts it into practice, book a Keeyu demo.

Every order is a promise. Keeyu keeps it.

Carrier performance FAQ

What is carrier performance in e-commerce?

Carrier performance is how reliably a shipping carrier delivers your parcels on the promised date, with complete scan data, in good condition, and at the agreed cost. Most merchants track it with a weekly scorecard covering on-time rate, first-scan time, scan gaps, exceptions, loss and damage, claims, and transit time by lane.

How do you calculate carrier on-time delivery rate?

Divide shipments delivered on or before the promised date by total shipments delivered in the same period. Use the date the shopper saw at checkout, not only the carrier's commit date, because carriers can extend commit windows during peak. Our guide to improving on-time delivery rate covers setting that promise.

What should a carrier scorecard include?

Start with on-time rate against your promised date, first-scan time, no-scan gaps, and exception rate. Add lost and damaged rate, claim approval rate, transit time by lane, and cost per on-time delivery. Segment each metric by service level, origin warehouse, and destination zone before comparing carriers.

Are carrier money-back guarantees still available?

It depends on the carrier, the service, your negotiated contract, and the date you ship. Terms change, so a guarantee you used last season may not cover you now. Read your carrier agreement and the carrier's current published terms before building a refund process around them. Track commit dates so you can claim when one applies.

How do I use carrier data to allocate shipping volume?

Compare carriers within the same lane and service level, then rank them on cost per on-time delivery rather than label rate. Shift volume one zone at a time, measure for a few weeks, and keep the change only if the scorecard holds. Our ranking of 3PL and carrier visibility tools covers software that helps.

Why do carrier delays increase support tickets?

When a parcel stalls and nobody tells the shopper, they go looking for answers, and that search becomes a where is my order ticket. Many affected shoppers never write in at all and simply do not buy again. Telling them first, with a revised date, cuts the tickets. Our example of how to inform a customer about a delivery delay has wording.

What is a no-scan gap and why does it matter?

A no-scan gap is a stretch of time with no carrier tracking event on a shipment in transit. Long gaps often mean a parcel is stuck, misrouted, or lost. Setting a threshold, such as 48 hours on ground service, lets you act before the shopper notices, and it gives you evidence for claims.

• 1. ShipMatrix, "Parcel Carriers delivered Merry Christmas to consumers in 2025 Peak", https://shipmatrix.com/wp-content/uploads/2026/01/SMx-on-parcel-carriers-OTP-in-peak-of-2025_Jan-12_2026.pdf. On-time performance for December 1 to 31, 2025 (FedEx 95.3, UPS 97.2, USPS 94.1 percent) and peak 2024 (91.8, 96.5, 90.4 percent), all parcel services combined, against actual carrier commit times; UPS adding an extra day to more Ground lanes than FedEx.
• 2. United States Postal Service, "Service Performance Results", https://about.usps.com/what/performance/service-performance/. Quarterly service performance results by mail class and an area dashboard; Package Services covers Bound Printed Matter, Media Mail and Library Mail.