Under 1%. Industry reporting commonly puts 3% to 5% as ordinary for direct-to-consumer fulfillment, so a sustained sub-1% reported error rate is the mark of a well-run operation. But the reported rate understates reality — unreported errors can multiply it several times over — and order complexity makes two providers' numbers hard to compare directly.
What e-commerce did to the fulfillment industry
Traditional fulfillment was a pallet-in, pallet-out business. Staff were forklift drivers, the software tracked cases, and an order was a truckload. Direct-to-consumer e-commerce broke that model completely: it requires a different SKU organization, staff picking and packing individual units all day, software that integrates with a dozen sales channels, and a tolerance for volume that swings 5x in a week.
That transition has been difficult for a lot of traditional warehouses, and many chose to stay out of DTC entirely and stick to B2B. It is worth knowing which kind of company you are talking to, because the error profile of the two is not the same.
Why not a 0% error rate?
Ideally, from the client's side, it is zero. In practice it cannot be. B2C fulfillment involves far more physical touches, far more SKUs to pick from, and far more orders per day than B2B. Add human nature, mislabeling by manufacturers, and conditional instructions (if the packing slip says X, do Y), and zero stops being a target and becomes a fiction.
There is surprisingly little published literature on the subject, but the reporting that exists puts 3% to 5% as common across the industry. A genuinely good operation runs at 1% or less. That is the standard we hold ourselves to, using the same care on client orders that we used on our own e-commerce business.
What the error rate really means
Three things distort the number, and all three matter when you are comparing providers.
1. It only counts the errors that were reported
The rate is what reaches customer service. It excludes carrier damage and manufacturer defects — correctly, since those are not the 3PL's doing — but it also excludes every mistake nobody mentioned.
Assume errors are roughly evenly split between those that favor the customer and those that do not. The ones in the customer's favor mostly go unreported. That alone puts the real rate at around double the reported one. Then factor in price point and product type: on a low-value item, plenty of people simply let a wrong pick slide. Depending on the category, the true rate can be three or four times the reported figure, or more.
This is not an argument against the metric. It is an argument for knowing what you are looking at — and for asking a provider whether they also track internally caught errors, which is a much better signal of process health than the customer-reported number alone.
2. Small samples lie
One error in your first ten orders is a 10% error rate. The next ninety orders ship clean and it is 1%. Early numbers on a new account are noise, and there is a real learning curve at launch: a new brand's quirks, its SKU variants, its packaging rules all take a cycle or two to settle. Ideally there is enough prep time to work through that before go-live. Client timelines do not always allow it.
3. Complexity is not comparable between accounts
Consider ten orders. Nine have one item each. The tenth has 91 items. One item ships wrong. Is that a 10% error rate (one order of ten was wrong) or a 1% error rate (one item of a hundred)?
It is usually calculated per order, because an order being wrong is what forced the client to take action. But it demonstrates the point: an account with a high average units-per-order, a large SKU count, or many variants within a SKU carries far more chance of error than a single-SKU, single-item account. One client's error rate is not equivalent to another's, and a provider quoting a company-wide average is blending very different kinds of work.
Who should pay for the errors?
Any mistake a 3PL makes should be covered by the 3PL. The real question is what "covered" includes.
Work the arithmetic. Say a product costs $100, shipping is $10, and the 3PL charges $2 for order processing and pick and pack — of which they might net $0.50. If the vendor has to replace the product outright, they would need to ship 200 packages to break even on that single error. At a 1% error rate, one error every hundred packages, the fulfillment company loses money permanently.
In practice, errors rarely mean a lost product; they mean return shipping and re-shipping costs. But even that can run negative for the vendor on a large, express or international shipment.
So what is the right split? There is no single answer. It comes down to:
- The vendor clearly communicating expectations up front, so both sides are on the same page
- That understanding being written into the agreement, not left to goodwill
- A vendor being willing to cover more when they price in their own internal "insurance" against an acceptable error rate
- A relationship good enough to work through the cases that fall outside what either side would call normal
- The client budgeting for the reality that the rate will not be zero
What to ask a prospective 3PL
- What is your reported order error rate, and over what period?
- Is it calculated per order or per unit?
- Do you track errors caught internally before shipping?
- What is the rate on accounts that look like mine — similar SKU count and units per order?
- What do you cover when an error happens, and what is written into the agreement?
- What did you change the last time an account's rate went the wrong way?
The last question is the most revealing one. Any operation will have a bad month. What matters is whether anyone noticed and what they did about it.
AccuracyOperationsChoosing a 3PL