Two carriers quote you nearly the same rate per parcel. On the monthly invoice they look interchangeable. But one delivers on time 96% of the time and damages almost nothing; the other runs 84% on-time, generates a steady trickle of "where is my order?" tickets, and quietly loses you repeat customers every week. On paper they cost the same. In reality, one is far more expensive than the other — you just can't see it on the invoice.
That gap is invisible until you measure it. Most merchants pick carriers on headline price and gut feel, then never look back. Yet the cost of a bad carrier doesn't show up as a line item — it shows up as refunds, re-ships, support hours, one-star reviews, and customers who simply don't come back. Industry surveys are blunt about the stakes: roughly 70% of shoppers say on-time delivery is essential to trusting a brand, about 75% return after a good delivery experience, and 60% say they won't shop again after a late one. Your carrier is delivering that experience on your behalf, under your name. If you're not scoring their performance, you're outsourcing your reputation to a black box.
This guide is about opening that box. It covers the carrier performance metrics that actually predict cost and customer loss, how to turn them into a simple per-carrier scorecard, and — the part most merchants skip — how to act on the numbers: rebalance volume, renegotiate with evidence, and know exactly when a carrier has to go.
Why You Have to Measure Each Carrier Separately
If you ship with a single carrier, "performance" and "the carrier" are the same thing — there's nothing to compare. The moment you run a multi-carrier setup, that changes completely. Now every order is a choice, and each carrier is good at different things: one owns fast urban delivery, another is cheaper on heavy parcels, a third reaches rural regions the others can't. An average across all of them tells you nothing useful. The whole point is to see them side by side, on the same lanes, over the same period.
This is also why measurement is worth the effort: with more than one carrier connected, you can act on what you find. If one carrier's on-time rate on a given route collapses, you can shift that volume to a better performer tomorrow — not renegotiate a single-carrier contract you're locked into. Measurement without options is just anxiety. Measurement with a multi-carrier setup is leverage.
Fix it: Commit to scoring every carrier on the same metrics, broken down by carrier and by route or region — not one blended number. A carrier that's excellent in Istanbul and terrible in the east is two different stories hiding inside one average.
The Metrics That Actually Matter
You don't need forty KPIs. You need a handful that predict the two things a bad carrier costs you: lost customers (late, damaged, failed deliveries) and lost margin (the true cost per delivered order). This is different from your store-wide shipping KPIs — those measure your operation; these measure each carrier's contribution to it.
On-Time Delivery Rate
The single most important number. It's the share of shipments that arrive on or before the date you promised the customer.
On-time rate = (shipments delivered by the promised date ÷ total shipments) × 100
The subtlety: measure against the date you promised the customer, not the carrier's internal service level. A carrier can hit its own SLA and still make you look late if your estimated delivery date was more optimistic than their real transit time. Benchmark reports put the average across mainstream e-commerce near 88%, with top-quartile merchants above 95–97%. A widely cited rule of thumb: any carrier running below 90% on-time deserves a serious conversation or a reduced share of your volume.
Fix it: Track on-time rate per carrier per route, monthly. Where one carrier is consistently below 90% on a lane another carrier serves well, that's your first volume-shift candidate.
Transit Time and Its Variability
On-time rate tells you whether a carrier hit the promise; transit time tells you how fast and — more importantly — how predictably. A carrier that always takes three days is easier to build a business on than one that averages two days but swings between one and five. That variability is what wrecks your delivery-date promises and drives support tickets.
A useful early-warning sub-metric is first-scan time — the gap between when you create the label and when the carrier logs its first physical scan. A creeping first-scan time usually means pickup delays or handoff problems on the carrier's side, and it shows up here before it shows up as a late delivery.
Fix it: Watch the spread, not just the average transit time. Rising variability or first-scan lag is an early signal a carrier is straining — catch it before it turns into a wave of late deliveries.
First-Attempt Delivery Rate
The percentage of parcels delivered successfully on the first attempt. Every failed first attempt means a redelivery, a possible return to sender, a frustrated customer, and extra cost — and carriers vary a lot here depending on their local density and process. This metric pairs directly with your own work on reducing failed deliveries: some failures are address-quality problems you own, but a persistent gap between carriers on the same routes points at the carrier.
Fix it: Compare first-attempt rates across carriers on identical routes. If one is markedly worse where the others succeed, the difference is the carrier — not your addresses.
Exception Rate
The share of shipments that hit a delay, hold, misroute, or other disruption in transit. Exceptions are the leading indicator that sits upstream of late deliveries and angry customers — a rising exception rate today is next week's on-time collapse. Tracking it by carrier tells you which partner introduces the most friction into your fulfillment.
Fix it: Set a threshold (say, exceptions on more than 2–3% of a carrier's shipments) that triggers a review. Exceptions are also where proactive tracking notifications earn their keep — they turn a silent delay into a managed one.
Damage and Claims Rate
The percentage of shipments that arrive damaged or become a claim. Damage is expensive twice: once for the refund or replacement, and again for the customer you likely lose. Rates vary meaningfully by carrier and handling, and a carrier with a stubbornly high damage rate is telling you something about how they treat your parcels. Read this metric next to your lost and damaged parcel process — the claims data you're already filing is this metric, if you attribute it by carrier.
Fix it: Attribute every damage claim to its carrier and route. A single-digit damage rate is normal; a carrier that's consistently double another on the same product mix and packaging is a packaging problem you've ruled out — or a carrier problem you haven't acted on.
True Cost per Delivered Order
Not the headline rate — the all-in cost per successfully delivered order, including surcharges, volumetric/desi weight adjustments, redelivery fees, and the cost of failures. This is the metric that exposes the "cheap" carrier that isn't. A carrier with a low base rate but a high failure and redelivery rate can easily cost more per delivered order than a pricier, more reliable one.
True cost per delivered order = (base rates + surcharges + redelivery + claims/refunds) ÷ successfully delivered orders
Fix it: Rank carriers by cost per delivered order, not per label. This single reframing often flips which carrier looks cheapest — and it's the number to bring to a rate negotiation.
Building Your Carrier Scorecard
A scorecard is just these metrics, per carrier, side by side, reviewed on a schedule. It turns scattered gut feelings into one page you can act on. A minimal version:
| Metric | Carrier A | Carrier B | Carrier C | Target |
|---|---|---|---|---|
| On-time rate | 96% | 84% | 91% | ≥ 95% |
| Avg transit (days) | 1.9 | 2.4 | 2.1 | — |
| First-attempt rate | 94% | 88% | 92% | ≥ 92% |
| Exception rate | 1.8% | 4.1% | 2.6% | ≤ 2% |
| Damage/claims rate | 0.4% | 1.2% | 0.6% | < 1% |
| Cost / delivered order | ₺58 | ₺61 | ₺57 | lower is better |
A few rules that keep a scorecard honest:
- Break it down by route or region. A blended national number hides the fact that Carrier B might be your best option in one city and your worst in another. Route-level data is what lets you rebalance rather than replace.
- Weight by volume. A carrier with a 99% on-time rate on ten parcels a month isn't proven. Read the metrics alongside how much you actually ship with each.
- Compare like with like. Only compare carriers on lanes and parcel types they both serve. Judging a rural-focused carrier on urban next-day is unfair and misleading.
- Review monthly, trend quarterly. A monthly scorecard catches problems fast; a quarterly view of the trend tells you whether a bad month was a blip or a decline. Monthly-with-quarterly-trend is the cadence most operators settle on.
Fix it: If you're pulling this together from carrier portals and spreadsheets by hand, you'll do it once and abandon it. The data has to land in one place automatically for the habit to survive — which is the difference between a scorecard you actually keep and a one-off audit.
Turning the Scorecard Into Action
A scorecard nobody acts on is a report. The value is entirely in what you do next, and there's a clear escalation ladder.
1. Route by performance, automatically. The lightest-touch action: send each order to the carrier that scores best on its route, not one carrier for everything. Time-sensitive orders to your most on-time carrier on that lane; cost-sensitive orders to your cheapest reliable one. This is the everyday payoff of measuring — every order goes to the partner the data says is best for it.
2. Rebalance volume. When a carrier slips on specific lanes, shift that share to a better performer without touching the rest of the relationship. You keep the carrier where they're strong and stop feeding them where they're weak. This is only possible because you measured by route.
3. Renegotiate with evidence. Walking into a rate negotiation with "you feel expensive" gets you nowhere. Walking in with "your on-time rate on this lane is 84% against a competitor's 95%, and here's the exception data" changes the conversation — you're negotiating from evidence, and a carrier that wants to keep your volume has a reason to fix the problem or adjust the rate.
4. Replace. The last resort, reserved for carriers that fail the criteria below. Because you already run multiple carriers, replacement is a volume shift, not a cliff — you ramp the failing carrier down and a proven one up.
Acting on carrier data pays off measurably: benchmark analyses suggest that renegotiating or replacing chronically underperforming carriers (those below ~90% on-time) typically yields a 3–5 percentage-point improvement in on-time rate within 6–12 weeks — which flows straight through to fewer support tickets, fewer refunds, and more repeat orders.
When to Actually Switch a Carrier
Switching has real friction, so don't do it on one bad month. Do it when the scorecard shows a pattern you've already tried to fix:
- Chronic on-time failure — consistently below ~90% on lanes that matter, across multiple months, not a one-off.
- Unpredictable transit times — high variability that keeps breaking your delivery-date promises even when the average looks fine.
- A persistent exception or damage rate well above your other carriers on the same product mix and routes.
- Rising hidden costs — accessorial and surcharge creep that pushes true cost per delivered order past more reliable alternatives.
- Unresponsiveness — a carrier that won't engage with the data or renegotiate when you bring evidence. That, on its own, is a reason to move volume.
Fix it: Give a failing carrier one documented chance — share the scorecard, agree on a target, set a review date. If the next review doesn't show improvement, execute the volume shift you'd already planned. Deciding the exit criteria before you're emotional about a bad peak season keeps the decision rational.
Common Mistakes That Hide the Truth
- Judging on price alone. The headline rate is the most visible number and the least complete. Cost per delivered order is the honest one.
- Blending everything into one average. A single national on-time number is where good and bad carrier performance cancel out and you learn nothing. Segment by route.
- Measuring against the carrier's SLA, not your promise. Your customer was told a date by you. That's the bar. A carrier hitting its own SLA while missing your promise is still a problem you own.
- Reviewing once a year. Carrier performance drifts — networks get congested, service areas change, peak season strains everyone. An annual look catches the decline months too late. Monthly is the habit that works.
- Collecting data you never act on. The most common failure of all. A scorecard's only job is to trigger a decision.
Your Carrier Performance Checklist
Run your carrier program against this list:
Measure the right things:
- On-time rate is tracked per carrier, per route, against your promised date
- Transit time and its variability (plus first-scan time) are monitored
- First-attempt delivery, exception, and damage/claims rates are attributed by carrier
- Cost is measured per delivered order — surcharges and failures included — not per label
Make it a habit:
- Every carrier is scored on the same metrics, side by side
- Metrics are broken down by route/region, not blended into one national number
- The scorecard is reviewed monthly with a quarterly trend view
- The data lands in one place automatically, not hand-assembled from portals
Act on it:
- Orders route to the best-scoring carrier for their route
- Volume is rebalanced off weak lanes toward stronger performers
- Negotiations use performance evidence, not gut feel
- Clear, pre-agreed criteria decide when a carrier is replaced
Stop Guessing Which Carrier Is Good
The carrier you think is your best one and the carrier the data says is your best one are often not the same — and the difference is measured in refunds, support tickets, and customers who quietly don't come back. You can't manage what you can't see, and for most merchants carrier performance is the single largest unmeasured cost in their shipping operation.
If you do only three things, do these: measure on-time, damage, and true cost per delivered order for each carrier separately; break every number down by route so you can rebalance instead of just replacing; and review it monthly so you catch a decline while it's still a blip. That alone moves you from picking carriers on price and hope to managing them on evidence.
Doing this by hand — logging into each carrier's portal, stitching exports into a spreadsheet, attributing every claim and surcharge — is exactly why most merchants never keep it up. When every carrier's on-time rate, transit time, first-attempt success, damage rate, and true cost land in one dashboard, broken down by carrier and route and updated automatically, the scorecard stops being a project and becomes something you glance at — and the next time a carrier starts slipping, you see it in a week instead of a quarter, and you already have a better one connected to shift the volume to.
