Amazon MCF or a 3PL: The Fulfillment Decision for Multi-Channel Brands

A brand doing eight figures on Amazon turns on its own Shopify store, and for about six weeks the fulfillment question looks solved. Amazon already holds the inventory. Multi-Channel Fulfillment will ship it anywhere. Nobody has to go negotiate with a warehouse.

Then the first full month of off-Amazon orders clears, somebody reconciles the invoices against the order count, and it stops looking solved.

We've had this conversation with enough brands to know it almost never turns on the per-order rate, which is where everyone starts. It turns on three things nobody priced: whose inventory pool the units left, what the box looks like when it lands on a doorstep, and who owns the return. Fulfillment is an operations decision wearing a logistics costume, which is why it sits inside our Amazon operations management practice rather than next to it. What follows is the framework underneath that.

What does Amazon MCF actually do?

It uses your existing FBA inventory and Amazon's fulfillment network to ship orders that didn't come from Amazon. You pass the order in through the API or a connector app, Amazon picks it from the same pooled units your Amazon orders ship from, and it goes out in either Amazon-branded or unbranded packaging depending on the option you choose.

The appeal is real. One inventory pool, one set of operational habits, no second warehouse relationship to manage, and delivery speed that comes from a network you could not build or buy at that scale. Buy with Prime pushes the same idea further by putting a Prime delivery promise on your own product pages.

What MCF is not is a branded fulfillment experience, and it is not a pool of inventory you control independently of Amazon. Both of those matter more than most brands expect going in.

The seven dimensions the decision actually turns on

Price per order is one row in this table and it's rarely the deciding one.

Dimension	Amazon MCF	A 3PL
Inventory pool	Shared with FBA. Every off-Amazon order draws down the same units your Amazon rank depends on.	Separate. Your Amazon stock stays untouched by DTC demand.
Packaging	Amazon-branded or plain. No inserts, no custom box, no unboxing moment.	Your box, your inserts, your kitting and bundles.
Setup effort	Low. A connector and a careful settings pass.	Weeks. Contract, onboarding, integration, and a stock transfer.
Delivery speed	Amazon's network, which is the strongest argument in MCF's favor.	Depends on the provider's footprint and how many nodes you pay to stock.
Returns	Routed through Amazon's process, which a customer who bought on your site may not expect.	Whatever you specify, including inspection rules and restock conditions.
Channel breadth	Good for marketplace overflow where packaging doesn't matter.	Better when you sell in several places with different packing and labeling rules.
Fee control	Amazon sets the rate and changes it. You find out when it changes.	Contracted, negotiable, and reviewable at renewal.

That last row is the one brands underweight. A fulfillment cost you don't control is a margin assumption you don't control, and if off-Amazon is becoming a real share of your revenue, you've handed the economics of your own channel to a platform you also compete with.

Where MCF clearly wins

Three situations, and they're common.

Amazon is the business and DTC is a rounding error. If your own site is a few percent of revenue and mostly exists so the brand has a home, a second warehouse relationship is overhead you don't need. Turn MCF on, accept the packaging, move on to a problem that matters more.

You're testing whether a channel exists. Standing up a 3PL to validate demand is backwards. MCF lets you find out whether anyone buys from your site at all before you sign anything with a term length.

Marketplace overflow. Orders from marketplaces where the customer never sees your brand as a sender anyway. Packaging is not doing any work there, so paying for it is waste.

Where a 3PL clearly wins

Being fair about this is the whole point of the framework. A 3PL is the right answer more often than Amazon-first brands expect.

The brand experience is the strategy. If your DTC positioning rests on how the product arrives, the insert, the tissue, the card, the sequencing of a subscription box, MCF removes the thing you were selling. We've watched brands spend heavily on DTC acquisition and then ship the order in a plain box with no insert, and the repeat rate tells you what the customer concluded.

You need kitting, bundles or anything custom. Multi-unit bundles, gift sets, subscription assortments, serialized inserts. This is ordinary 3PL work and it is not what MCF is for.

Off-Amazon volume is no longer small. There's a crossover point where contracted rates and a negotiated relationship beat a platform rate you can't influence. Where it sits depends on your weight, dimensions and destination mix, which is why the next section is a formula and not a number.

You want your Amazon inventory protected from DTC demand. Which brings up the thing that actually bites.

The inventory question nobody asks until it bites

MCF draws from the same pool as FBA. So a good week on your own site consumes the units your Amazon rank depends on, and a promotion that works too well off Amazon can put your best ASIN out of stock on the channel paying the bills. Amazon's restock limits sharpen this: the units are capped, and off-Amazon demand is now competing for the cap.

Most brands discover this after the stockout rather than before it. The planning fix isn't complicated, but it has to exist: forecast the two demand streams separately, hold a buffer sized against your Amazon velocity rather than your blended velocity, and treat an off-Amazon promotion as an inbound planning event.

We've worked this problem from the other side. On a home and kitchen account that kept hitting a 15,000 unit restock limit, the fix was smaller and more frequent inbound shipments, daily rather than weekly monitoring of sales and inventory data, and prioritizing the best-selling ASINs in every restock decision. That account earned its way to a restock limit above 176,000 units and held a 97% in-stock rate. It's written up as an inventory optimization case study if you want the detail.

The lesson transfers directly: whoever is shipping your off-Amazon orders, the constraint is the forecast, not the warehouse.

How to run the math on your own account

Ignore blended averages, including ours. Pull your own last ninety days of off-Amazon orders and compute, per order:

1. Fulfillment cost per order, at your real weight and dimension mix rather than a representative SKU. Quote the 3PL on that same file.
2. Inbound and storage, including the second stock position a 3PL requires and the working capital sitting in it.
3. Return cost per order, at your actual return rate, not the category's.
4. Packaging, as an explicit line. If branded packaging is strategy, price it as strategy rather than treating it as free.
5. The stockout cost, which is the one everybody omits. Take your Amazon contribution margin per unit, multiply by the units a rank recovery costs you after a stockout, and assign a share of it to the pooled-inventory risk.

That fifth line is usually what flips the answer. A 3PL can look more expensive per order and still be cheaper once you price the risk of your Amazon rank being hostage to a DTC sale.

The hybrid most multi-channel brands land on

FBA for Amazon. MCF for marketplace overflow where packaging is irrelevant. A 3PL for branded DTC. It looks like complexity on a slide and in practice it's three clean rules that each match a channel's actual job.

The honest trade-off: you're now running two fulfillment relationships and reconciling two invoice formats, and somebody has to own that weekly. Brands that split fulfillment without assigning an owner end up with inventory in the wrong node and nobody noticing for a month. If no one on your team has operations as their first job rather than their third, the hybrid is the wrong shape for you and a single provider is better even at a worse rate.

What we'd tell you to do first

Before you quote anybody: separate your off-Amazon demand forecast from your Amazon one. You cannot evaluate either option on a blended number, and almost every brand we talk to is working from a blended number. Once those two forecasts exist the decision usually makes itself, and the quotes you collect afterward are confirming a choice rather than driving one.

If you're mostly Amazon with a small DTC tail, turn MCF on and go spend your attention somewhere it earns more. If DTC is where the brand is actually being built, the plain box costs you more than any freight line will ever show. Most brands past $5M end up running both, and the ones who do it well decided that deliberately instead of discovering it in an invoice.

Run the numbers on your own file before you sign anything. If you'd rather have an operator run them with you, book a call with our team, or read how we structure Amazon operations management.