Is a 98% on-time rate good?
The number alone tells you nothing. Here is the framework a neutral operator uses to judge whether your 3PL's metrics are actually good.
A 98% on-time rate can be excellent or a warning sign, because the number by itself tells you almost nothing. What matters is how on-time is defined, what it quietly excludes, whether it fits your actual order profile, which way it is trending, and whether it matches what your customers actually experience. A good 3PL metric is one that is defined honestly, measured against your specific orders, and consistent with how the relationship actually feels. When the number looks fine but the relationship does not, that gap is the real signal, not the metric.
Both statements can be true at the same time. That is the whole reason a single number can never be trusted on its own.
The number is not the answer. The definition is.
Ask what a good on-time rate is and you will get confident answers and round numbers. Almost none of them survive the follow-up question: on time measured at what?
In auditing 3PL contracts, a consistent pattern shows up. The headline percentages sit close together, most of them in the high 90s, but the definitions behind them barely overlap. One contract stops the clock when the shipping label is created. Another when the carrier scans the box. Another when the package reaches the customer's door. And some rates only hold inside a normal day's volume, so the orders most likely to be late are the ones least likely to be counted.
| Provider | Reported on-time | What on-time actually means |
|---|---|---|
| Provider A | 98% | Measured when the shipping label is created. The order can sit on the dock for days and still count as on time. |
| Provider B | 98% | Measured when the carrier picks up. Closer to reality, but says nothing about when the customer gets the box. |
| Provider C | 97% | Measured at delivery to the customer. The strictest definition on this list, and the lowest number. |
| Provider D | 99% | Holds only within a normal daily volume cap. Orders above the cap, the ones most likely to run late, are excluded. |
Four nearly identical on-time numbers, four completely different definitions of what on-time even means. The 97% here may be the best operator on the list.
Five questions that tell you if your number is good
No borrowed benchmark required. Your own contract and your own customers hold every answer.
How is it defined?
On-time can be measured at label creation, carrier pickup, or delivery to the customer. Two providers can both report 99% and mean opposite things. Ask what event stops the clock.
What does it exclude?
Rates get flattering when the denominator shrinks. Volume caps, held or backordered lines, weather and carrier exclusions, and non-operational days can lift a number without service improving.
Does it fit your order profile?
A benchmark borrowed from a different operation is noise. Peak versus steady, DTC versus B2B, simple picks versus kitting all change what good looks like for you.
Which way is it trending?
Direction beats level. A steady 96% can be healthier than a 99% that has slipped two points in two quarters. Ask for the trend line, not the snapshot.
Does it match lived experience?
The most important check. If the dashboard says 98% but customers say packages are not arriving, the gap between the number and the experience is the real metric. A good number that nobody believes is not a good number, it is an unexamined one.
The metrics that actually matter, in context
Same rule for every one of them: meaningful only against your profile, your definitions, and your trend.
Order accuracy & misships
The share of orders picked, packed, and shipped correctly. Line-level and order-level accuracy are different tests. Agree on which one you are measuring, then judge it by trend.
Order cycle time
How long an order takes from received to shipped. Averages hide the tail. Ask about your slowest days and your peaks, not just the mean.
Inventory accuracy
How closely the system's counts match the shelf. Drift here quietly creates oversells, stockouts, and billing disputes that surface later as service failures.
Receiving turnaround
How fast inbound stock becomes sellable. A slow dock makes every downstream metric look worse while the scorecard blames fulfillment.
Returns processing
How quickly returns are inspected and back on the shelf. In DTC this is working capital sitting in a bin. Measure it in days, against your own volume.
How to read all of them
Every metric on this list gets the same five questions: how it is defined, what it excludes, whether it fits your profile, which way it is trending, and whether it matches lived experience. No target number survives without that context.
The scorecard is not the relationship
A QBR that is a data dump of green metrics can sit right next to a partner who has quietly decided to leave. The honest benchmark for a partnership is not a color on a dashboard. It is a neutral read of whether the relationship is actually healthy: whether commitments are believed, whether problems get raised early, and whether both sides still want this to work.
And if the gap between the number and the experience turns out to be real, do the math before you act on it. Price what leaving would actually cost before you start an RFP, because a switch has a way of costing far more than the quote that tempts you into it.
The honest benchmark is not a number on a dashboard, it is whether the relationship still works for both sides.
Get an honest read on your partnership.
A Revenue-at-Risk Review is a confidential, neutral read on what is actually working and what is at risk, for both sides. The first conversation is free and confidential.
Prefer email? info@logisticsresolve.com
3PL metrics, answered
Is a 98% on-time rate good for a 3PL?
A 98% on-time rate can be excellent or a warning sign, because the number alone tells you almost nothing. It is good only if the definition is honest, the exclusions are small, it holds through your peaks, the trend is stable or improving, and it matches what your customers actually experience. If the dashboard says 98% and customers are complaining, trust the gap, not the number.
What is a good on-time delivery rate for a 3PL?
There is no single good on-time rate, because providers define on-time differently and order profiles vary. Contracts commonly set on-time SLAs in the high 90s, but a rate measured at label creation and a rate measured at delivery are different metrics wearing the same name. Judge your rate by its definition, its exclusions, its fit to your order profile, its trend, and whether it matches lived experience.
What is a good order accuracy or misship rate for a 3PL?
There is no universal target for order accuracy or misships, and any page that gives you one is guessing. Judge accuracy the way you judge on-time: whether it is measured at the line level or the order level, what the denominator excludes, how it holds up during your peaks, and which way it is trending. A stable rate that matches customer experience is worth more than an impressive one that does not.
Why do my 3PL's metrics look good when service feels bad?
Usually because the metric and your customers are measuring different things. A rate can be counted at label creation while customers wait on delivery, and exclusions like volume caps, backorders, and carrier delays can shrink the denominator until the number looks great. The gap between a green dashboard and a bad experience is itself the signal, and the first thing to check is the definition behind each number.
What KPIs should I track for my 3PL?
Track on-time performance measured at an event you actually care about, order accuracy or misship rate, order cycle time, inventory accuracy, receiving turnaround, and returns processing time. For each one, agree on the definition in writing, measure it against your own order profile, and watch the trend over time. A handful of honestly defined metrics beats a page of flattering ones.
How do I know if my 3PL is underperforming?
Look at the gap between the numbers and the experience, not just the numbers. If the scorecard is green but customers are complaining, escalations are rising, or QBRs feel defensive, the relationship is drifting even if the metrics are not. A Revenue-at-Risk Review is a confidential, neutral read that tells both sides whether the numbers reflect reality and what is actually at stake.