Amazon MCF vs. Overseas Warehouse Fulfillment: Where the Cost Lines Cross

Bottom line up front: if you're doing a steady 300+ orders a day with fewer than 50 SKUs, shifting 70% of your non-Amazon orders from MCF to a US overseas warehouse will typically cut per-order fulfillment cost by 15–25%. But with 200+ SKUs and under 100 orders a day, stick with MCF — the warehouse's fixed fees will eat every dollar you saved.

MCF — Amazon Multi-Channel Fulfillment — lets you fulfill orders from other channels out of your FBA inventory. One inventory pool for every channel. Sounds beautiful. But in 2024 I ran three projects migrating clients off MCF, and the conclusion was the same every time: MCF is a convenience tax, and the more volume you push through it, the more that tax stings.

Tear Apart the Cost Structure: MCF vs. Overseas Warehouse

MCF costs three things: a per-unit fulfillment fee (tiered by size and weight — Amazon publishes the rate card), FBA storage fees, and inbound placement and transfer fees. Its signature trait: zero fixed cost. One unit or ten thousand, you pay per piece, and adding SKUs costs nothing extra.

The overseas warehouse runs on different math: inbound receiving fees, storage billed per pallet or per cubic foot, pick-and-pack billed per order plus per unit, last-mile postage — and one immovable monthly account fee or minimum commitment. Its signature trait: high fixed cost, low marginal cost. The more orders you run, the thinner the fixed cost spreads.

Once you see the structure, the dividing line draws itself: low volume with scattered SKUs, MCF wins. High volume with concentrated SKUs, the warehouse wins. There is no universally cheaper option — only the cheaper option for YOUR order profile.

Both sides carry hidden costs that never show up on the quote:

On the MCF side: FBA long-term storage fees (inventory sitting over 180 days), and the invisible stockout cost when inbound congestion hits during peak season. On the warehouse side: returns processing fees (billed per unit — painful for high-return categories), the one-time systems integration cost (API work typically runs in the low five figures in RMB through an outside developer), and switching cost — if the 3PL relationship sours and you want to move, the transfer fees and stockout risk alone will ruin your quarter.

So never compare on the "per-order handling" line alone. Amortize every hidden cost, then compare. In my project work I always build a 12-month total cost view — peak surcharges, long-term storage, all of it — before the recommendation stands.

Splitting Traffic: MCF and Warehouse Order Routing Rules

A cross-discipline note here: MCF vs. overseas warehouse isn't either-or — it's traffic splitting. And the key to the two modes coexisting long-term is writing the split rules into your system instead of letting ops decide by gut feel every day. My standard four steps:

Step 1: grade SKUs by 90-day velocity — A/B/C. A-graders (fast movers, time-sensitive, low return rate) stay on MCF for speed and delivery reliability. B and C grades move to the warehouse for cost. Re-grade monthly; upgrades and downgrades trigger automatic inventory transfers.

Step 2: write the routing rules into the OMS. My template is three rules: rush orders from your own store go MCF expedited; subscription repeat orders go to the warehouse (low time sensitivity, cost priority); multi-SKU bundle orders ship from a single location first — splitting across warehouses needs human approval.

Expand the three rules into a routing table and load it straight into the OMS rule engine:

Step 3: split safety stock by channel. The MCF pool and the warehouse pool each get their own waterline — never share one. Size them by each channel's last-90-day sales, and rebalance weekly as velocity shifts.

Step 4: reconcile weekly. Pull the MCF invoice and the warehouse invoice, compute cost per order, and any drift beyond 10% on either side triggers a review. Too many sellers migrate and then stop watching — three months later the rates moved and nobody noticed.

The Math: Where the MCF Cost Lines Cross

Illustrative table based on a home-goods client's actual 2025 Q2 operating profile (rates per the latest official schedules — treat the figures as a worked example):

Assumptions: standard size, approx. 1.5 lb per unit, 500 orders/day, 32 SKUs, 70% of orders eligible for the warehouse.

Real money saved. But the most important row is the fixed-fee allocation: at 500 orders a day it's $0.40; at 50 orders a day it becomes $4.00 and wipes out the savings. That row IS the crossover math.

Here's the counterintuitive call: the decision variable isn't total volume — it's order concentration. If 80% of your orders come from 20% of SKUs, the warehouse is a slam dunk. If you carry 300 SKUs and half of them move a few units a month, MCF's zero fixed cost is a moat. I've seen a seller at 800 orders a day stay all-in on MCF — 400+ SKUs, and migrating would have bled money on inbounding and systems integration alone.

Three Failure Modes — This Pitfall I've Seen Three Times

Failure mode one: no routing rules, one buyer's two items ship in two parcels. One via MCF, one from the warehouse. Two packages, two postage charges, and customer service gets asked "why did this ship separately?" Three clients, same hole. Some numbers: one of those clients saw 12% of multi-SKU own-store orders split into two parcels in March 2025 — nearly $4,000 a month in extra postage, and support tickets up 30%. After the rules went into the system, the split rate dropped below 1%. The fix is step two's routing table: same-order multi-SKU ships from one location first; cross-warehouse splits need approval.

Failure mode two: lopsided inventory. A client pushed everything to the warehouse for the savings, MCF stocked out, Amazon-channel orders missed their ship windows, and account health went red. The rescue: hard-split safety stock by channel sales ratio, keep at least 14 days of sales on the MCF side, and trigger automatic transfers from the warehouse below the red line — transfers, not a fresh purchase order after you're already out.

Failure mode three: migrating and then never reconciling. Year one at the warehouse is priced to win your business; year two brings a 15% storage hike and peak-season pick surcharges, and per-order cost quietly climbs back to MCF levels. One client went eight months without reconciling — when I ran the numbers, the hybrid setup was costing 6% MORE than all-MCF. In my project delivery standard, reconciliation is a weekly motion, not a quarterly one.

The MCF-vs-warehouse question isn't answered in Amazon's rate card. It's answered in your own order data. One action item for this week: pull 90 days of orders, rank SKUs by concentration, and see what share of orders the top 20% of SKUs drive. The answer will jump out at you.

By Lin Xiao, WMS Implementation Consultant