By Bernardo Campelo — Forbes Business Council E-Commerce Leader, Amazon SPN Certified provider, Amazon SP-API authorized partner, and Founder of PrepVia.
I talk to sellers every month who know their 3PL is costing them money and stay anyway. Not because they are indecisive. Because the math of leaving looks worse than the math of staying, and nobody has ever shown them a sequence that does not involve a window where their inventory is in neither building.
That fear is rational. A badly run migration produces exactly what they are afraid of: units in transit during a restock, a stockout on a hero ASIN, and a recovery that takes longer than the problem they were solving.
But the fear is only rational for the hard cutover, which is the version almost everyone imagines and nobody should ever run. Done properly, switching providers does not have a gap at all, because you never move the inventory you are currently selling from.
The 60-second version
Never do a hard cutover. Do not empty one warehouse into another. Point new purchase orders at the new provider and let the old one drain down on the inventory it already has.
The overlap is the whole method. For four to eight weeks you are live in both facilities. You pay a little in duplicated overhead. You pay nothing in stockouts, which is the only cost that actually threatens the business.
Read the exit clause before you sign anything — including with your current provider. Notice period, storage rates after termination, and per-unit transfer-out fees decide what leaving costs, and they are written before you ever want to leave.
The 2026 timing: Black Friday inventory has to reach Amazon by October 21 or 28 depending on your split option. Working backward through prep, freight, and production, a migration started after August is a migration running live through peak. Start it now or start it in January.
Why Sellers Stay Too Long
The decision to switch is almost never triggered by a single failure. It is triggered by accumulation: the turnaround that used to be two days is now six, the inventory report needs an email, the Q4 queue is unpredictable, and the invoice has line items nobody can explain.
None of those individually justify the disruption of moving. Together they cost more than the move would, but they cost it slowly, and slow costs do not force decisions. The stockout risk of switching is immediate and vivid. So sellers stay, and the accumulation continues.
The way out of that trap is to notice that the comparison is wrong. You are not comparing "stay" against "hard cutover." You are comparing "stay" against a phased migration, whose actual worst case is paying two providers a little for a couple of months.
The Overlap Method
The principle is simple: new inventory goes to the new provider, old inventory sells down where it is. You are never dependent on units that are currently on a truck.
- Sign the new provider and integrate it first, while shipping nothing. Connect the APIs, create the accounts, confirm the address, and get the rate card in writing. Integration problems should surface while you still have zero dependency on the outcome.
- Send one small test shipment. One SKU, a few hundred units, all the way through: receiving, prep, shipment creation, freight, check-in at Amazon. You are buying information about how the new provider actually behaves, not saving money on this shipment.
- Route the next purchase order to the new provider. This is the actual switch, and it involves moving nothing. Your supplier ships to a different address. That is the entire operational change.
- Let the old provider drain. Keep inbounding from its existing stock as normal. It is still your working inventory. Nothing is stranded, nothing is in limbo.
- Move only what is left, and only if it is worth moving. When the old facility is down to a tail, compare the transfer cost per unit against the value of the units. Slow movers and low-margin SKUs are frequently cheaper to liquidate or inbound to the marketplace than to truck across the country.
- Close the account in writing and get a final reconciliation. Written confirmation of zero units remaining, and a final invoice you have checked line by line against the rate card.
Steps three and four are the whole trick. The migration happens at the purchase-order level, not at the pallet level, so there is no moment when your sellable inventory is in a truck.
The cost of the overlap is real but bounded: you carry storage or minimums at two facilities for four to eight weeks. Compare that against one week of stockout on your best ASIN and the arithmetic answers itself.
Read the Exit Clause Before You Need It
The leverage in this relationship is entirely in the contract, and the contract is signed when everyone is optimistic.
Six terms determine what leaving actually costs. Check them in your current agreement today, and negotiate them in any new one before you sign.
| Term | What to look for | Why it matters on exit |
|---|---|---|
| Notice period | 30 days is normal; 90 is a lot | Sets the earliest date you can be fully out |
| Storage rate after termination | Whether it changes once notice is given | Some agreements reprice storage during wind-down |
| Transfer-out fee | Per unit, per pallet, or per hour of labor | Often the single largest exit line item |
| Outstanding balance conditions | Whether release of goods is tied to payment | A billing dispute should not become an inventory dispute |
| Data and record export | Whether you can export inventory and shipment history | Your receiving records support marketplace reimbursement claims |
| Loading appointment obligations | Who schedules and who pays for the outbound truck | Determines whether you can actually execute on your timeline |
The Q4 Timing Problem
Migrations have a season, and in the second half of the year the calendar decides for you.
Black Friday inventory must reach Amazon by October 21 for FBA shipments using minimal shipment splits, or October 28 using Amazon-optimized splits. Work backward through check-in buffer, freight, prep, and receiving and the goods have to be standing at your prep facility in early October. Work back further through ocean freight and production and the purchase order is placed around now — the full calendar is in the Q4 2026 inventory deadlines.
That PO is the one you want landing at the new provider. Which means the integration, the test shipment, and the decision all have to happen before it ships.
| Start the migration in… | What happens |
|---|---|
| Late July to August | Test shipment clears, Q4 POs route to the new provider, the old one drains through the fall. This is the window. |
| September | Tight. Possible for domestic suppliers with short lead times; risky for anything on the water. |
| October to December | Do not. You would be running a first-time integration through the highest-volume weeks of the year. |
| January | The other good window. Volume is low, mistakes are cheap, and you have three quarters to settle in. |
If you are reading this in late July and you already know your provider will not hold up in peak, the honest answer is that this month and next are the decision window. Not because a vendor wants urgency, but because the arithmetic of ocean freight and an October 21 deadline does not move.
What to Verify Before You Commit
- Run the test shipment end to end, including check-in. Not to the new warehouse — all the way to received at Amazon. The failure modes live in the last mile of that chain.
- Confirm the guaranteed peak turnaround in writing. Ask specifically what the commitment is in October, and what the remedy is if it is missed. See why most prep center SLAs are misleading.
- Verify live API integrations by name. Amazon SP-API at minimum; Walmart and TikTok Shop if you sell there. Ask what is live today, not what is planned.
- Get the complete rate card, including exit terms. Receiving, per-unit prep, storage after the free window, returns, minimums, and transfer-out. Negotiate the exit terms while you are the one being courted.
- Check real-time inventory visibility yourself. Log in during the test shipment and confirm you can see units move without asking a human.
- Do not move slow movers. Compare per-unit transfer cost against unit value. Liquidating or inbounding a slow tail usually beats trucking it.
- Export your historical records before closing the old account. Receiving logs, shipment history, and discrepancy reports. You will need them for claims after the relationship ends.
Frequently Asked Questions
How do I switch 3PLs without a stockout?
Do not transfer inventory. Route new purchase orders to the new provider and let the old provider sell down the stock it already holds. You run in both facilities for four to eight weeks, which costs some duplicated overhead but creates no window where your sellable inventory is in transit. Only move whatever tail remains at the end, and only if the transfer cost is less than the value of those units.
How long does it take to migrate to a new 3PL?
Plan for four to eight weeks from signing to fully operational, though the phases matter more than the total. Integration and a test shipment take one to two weeks. The real switch happens the moment you send a purchase order to the new address. Draining the old facility takes as long as your existing inventory takes to sell, which is usually the longest phase and the one you do not have to rush.
When is the worst time to change fulfillment providers?
October through December. Running a first-time integration through peak means learning a new provider's failure modes during the weeks that generate most of your annual revenue. The two good windows are mid-summer, so the switch is complete before Q4 inbound deadlines, and January, when volume is low and mistakes are inexpensive.
Can a 3PL refuse to release my inventory?
What a provider may do is governed by the agreement you signed, and many agreements tie release of goods to a settled account balance. That is why the exit terms matter more than almost any other clause: notice period, transfer-out fees, storage rates during wind-down, and whether an outstanding balance affects release. Read those terms before signing, and keep the account current through a migration so that a billing dispute never becomes an inventory dispute.
What does it cost to move to a new 3PL?
Three buckets. Transfer-out fees at the old provider, charged per unit, per pallet, or per labor hour. Freight from the old facility to the new one, if you move anything. And the overlap cost of paying both providers during the drain-down period. The overlap is usually the largest of the three and the one most worth paying, because it is what eliminates the stockout risk.
Should I tell my current 3PL I am leaving before I am ready?
Give the contractual notice, but time it around your operational plan rather than your emotional one. Notice periods run from the date you give notice, and some agreements change storage pricing once notice is served. The practical sequence is to integrate the new provider, run a test shipment, confirm it works, then give notice with a wind-down schedule you have already planned.
Final Take
Most sellers who stay with a provider they have outgrown are not making a decision. They are avoiding one, because the version of the switch they have in their head is a hard cutover, and a hard cutover genuinely is dangerous.
The phased version is not. It is a purchase order sent to a different address, followed by a few weeks of patience while the old inventory clears. The worst realistic outcome is two invoices for a couple of months. The worst outcome of staying is a peak season spent watching a queue you cannot influence.
The only genuinely bad option is doing this in October. If you already know the answer, the calendar says decide in the next few weeks.
Talk to PrepVia about a phased migration →
24–36h prep · 32h Amazon check-in or free · No minimums · Net-30 · Amazon SPN Certified
Related Reading
- Amazon Q4 2026 Inventory Deadlines — The calendar that sets your migration window
- 3PL vs Prep Center vs Fulfillment Center — Make sure you are switching to the right category
- Why Your Prep Center SLA Is Misleading — What to demand in writing from the new provider
- Amazon Q4 Prep Capacity Crisis — Why providers that work in May fail in October
