3PL onboarding, from SKU file to live order.

3PL onboarding, step by step

An onboarding is a transfer of knowledge that was never written down. Your team knows that the blue variant ships in the bigger carton, that one supplier always mixes sizes in a carton, and that the shop sends gift lines at zero value. A warehouse knows none of this until someone records it. Most of what follows is about getting that knowledge written down before it costs you an order.

The sequence and who owns it

How long this takes depends on your data and systems, so ask what any timeline assumes. The order matters more than the calendar.

SKU master data: the file the rest depends on

The SKU master, or item master, is the record the warehouse trusts over anyone's memory. Pickers confirm against it, packers choose cartons from it, and a carrier can add charges for what it failed to predict. Start with active SKUs only.

Fields worth having

• Identity: SKU code as used by the shop and by your ERP, plus a description a person can read without opening the box.
• Barcodes: one for each handling level you actually use: unit, inner pack, outer carton.
• Physical data: weight and dimensions per unit and per carton, carton quantity, and how cartons build onto a pallet.
• Tracking rules: lot, expiry date or serial number, and whether capture is required on receipt, on dispatch or on both.
• Handling: fragile, top load limit, hanging, this way up, and any restricted product characteristics such as lithium batteries.
• Packing: approved carton or mailer, inserts, and rules like no shared parcel with liquids.
• Set logic: the component list for any bundle, and whether it is pre-assembled or built at pick.
• Value: the figure used for loss and damage discussions, so the first claim is not also the first conversation about value.

How it fails in practice

• One barcode on two SKUs. This can happen when a supplier reuses a code across colours. The scanner confirms, the customer receives the wrong colour, and everyone looks at the picker.
• Pack and unit confused. The shop sells single units, the supplier counts packs of six, and stock looks wrong by a factor nobody notices until a count.
• Dimensions copied from a catalogue. The carton selection is wrong, and the carrier's measurement decides your freight bill instead of your own data.
• Two codes for one product. The shop and the ERP disagree, so an integration creates phantom items or rejects orders.

Ask the provider to measure and scan a sample on the first inbound and compare it with your file. A file that passes without a single question deserves a second look.

ASN: telling the warehouse what is coming

An ASN, advance shipping notice, announces a delivery and its contents before it arrives. Without one, a truck is a surprise: the goods can be unloaded but not reconciled against anything.

• Sender, carrier, expected date, and how the delivery must be booked at the dock.
• Number of pallets and cartons, pallet type and height, and whether unloading equipment is needed.
• Each SKU with quantity, plus lot and expiry where tracked.
• Your purchase order or container reference, so a difference can be traced to a document.
• A flag for mixed-SKU cartons and pallets, which take longer and are often the source of count errors.

Agree in writing how receiving works. Is the count checked against the ASN, or done blind and then compared? Who hears about a difference, and at which point in the process? Are damaged cartons photographed and held, or received as damaged stock? What happens to goods that arrive with no notice at all? And brief your suppliers, because they are the ones who will label the cartons. Receiving rules only work if your factory has seen them.

Stock cutover: the moment of greatest risk

Cutover is the point at which sales stop drawing on old stock and start drawing on new stock. If you are moving from your own storeroom or another provider, this is where overselling and lost stock can happen. Three approaches are common.

A cutover checklist

1. Fix a stock freeze: the exact moment after which the old location ships nothing more.
2. Obtain a final count at origin, and get it in a form that can be compared by SKU.
3. Decide what happens to open orders, backorders and pre-orders at the freeze.
4. Reconcile three figures: origin count, what the transport documents say, and what was received. Assign every difference to someone.
5. Name the stock master. Decide which system publishes availability to the shop, and turn the old feed off before turning the new one on.
6. Change the returns address in your labels, your terms and your customer emails, and decide who collects parcels still travelling to the old address.
7. Check how your previous provider releases stock and what notice the contract requires. An unsettled invoice at that stage can delay the release of goods.

Two stock feeds writing to one shop can produce very confusing errors. Availability appears to correct itself, then flips back overnight. Make sure only one can write.

Integration testing: use your awkward orders

Connecting a sales platform to a warehouse is usually the easy part. Whether orders behave properly is another matter. Test with data exported from your own shop, because invented orders tend to be too tidy.

• Single-line order, multi-line order, and an order for several units of a single SKU.
• Bundle SKU: does it explode into components, and does stock fall on the right items?
• Free gift at zero value, and an order with a discount spread across lines.
• Long address, a company name, a parcel locker or pick-up point, and special characters in German and Polish names.
• Cancellation before release to the floor, and a cancellation after release. Also an address change.
• Order with partial stock: does it wait, split or fail?
• One order sent twice: does the system reject the duplicate?
• Multi-parcel dispatch: do all tracking numbers reach the shop and the customer?
• Stock feed: full snapshot or changes only? What happens if the connection drops overnight?
• Return registered in the warehouse: which status reaches your systems, and who acts on it?

One question separates a tested integration from a demonstration: when a message fails at the weekend, who sees it? Agree who monitors failures, how they are told, and what the manual fallback is. Print and scan real carrier labels at the same time, because a label that looks correct on screen can still fail on a scanner.

Two paperwork items that get forgotten

Personal data. A fulfilment provider receives customers' names and addresses in order to act on your instructions. Under Article 28 of the GDPR, where processing is carried out on behalf of a controller, it must be governed by a contract or other binding legal act, and the controller must use only processors providing sufficient guarantees. Settle your data processing terms during onboarding, not after the first order.

Packaging evidence. The EU packaging regulation (PPWR), applicable since 12 August 2026, addresses fulfilment service providers directly in Article 45, and the German Central Agency Packaging Register explains that a provider who is not the packaging manufacturer has to obtain registration and compliance information before providing services for that market. How this applies to your provider and your shipping routes is a question for an adviser. Keep your registration details to hand either way. See what to verify for Germany.

Acceptance: decide what passing means before you start

Without written criteria, go-live happens because a date arrived. Write the test before the pilot, keep it short, and name the person who signs it on each side.

1. Master data: a sample of SKUs has been scanned, weighed and measured, and differences are corrected.
2. Inbound: the first delivery is reconciled to the ASN, and differences are closed with a named decision.
3. Orders: the whole test script has passed, and the failed cases have been fixed and re-run.
4. Pack-out: you have opened parcels yourself and compared them to your packing standard.
5. Stock: a cycle count on chosen SKUs matches the system, and the shop displays the same figure.
6. Returns: a test return has travelled to the warehouse and followed your procedure, including the report you receive. See how returns are routed.
7. Billing: a sample period has been priced against the offer, so the first real invoice is no surprise. See what 3PLs charge for.
8. Contacts: you know who handles exceptions and who may approve unplanned work.

Then raise volume in steps instead of redirecting everything at once, and review exceptions together in the early period. A repeated exception should lead to a changed rule.

What to ask a provider before signing

• What do you need from us in the master data file, and what do you check yourself?
• What happens when an inbound delivery does not match its notice?
• Can we see your test approach, and do we test with our own order data?
• Who owns a count difference discovered after goods have been received?
• Which of our products need a procedure you do not have yet?
• What does leaving look like: how is stock released and how is data handed back?

At ShipRevo the sequence is scope, quote, set-up, first inbound, then test orders and test returns before volume starts. Details are in how it works. If your system has no ready connection, we can build a custom integration: see technology and integrations. Timing depends on your data and systems, so we agree it once we have seen them.

Updated: 2026-11-01