Fully Managed, Semi-Managed, Self-Fulfillment: Which Mode Makes You the Importer?

Marketplace structure notes from the FulfillDock by SBT desk, 12 October 2026. Platform behaviour described here comes from the press and analyst reports cited inline; no figure in this article is ours, and no marketplace has reviewed or approved it. The mode comparison is our own operational reading.

The useful way to sort marketplace fulfillment modes is not by who runs the storefront. It is by three questions that decide your costs, your risk and your paperwork: who is the importer of record, who owns the inventory while it sits in the destination country, and who has to answer a compliance officer when the goods are already onshore. Fully managed selling answers all three away from you. Semi-managed answering moves two of them back to you. Self-fulfillment moves all three.

Three modes, drawn by what you still control

The middle column is the one sellers choose when the other two stop working, and it is also the column most often entered without noticing what moved. You keep the listing you already had; you lose the assumption that someone else's import paperwork covers your goods.

What happened in May 2025, briefly

The structural push came from duty treatment rather than from platform strategy. A presidential order signed 30 July 2025 suspended the duty-free de minimis allowance for every country of origin, effective 29 August 2025; China and Hong Kong consignments had already lost that treatment at the start of May 2025. Reuters reported on 12 May 2025 that on 2 May Temu said its United States sales were then being handled by locally based sellers.

We mention the sequence for one reason: it explains why a mode that reads as a commercial choice is really a customs consequence. Once a low-value parcel can no longer cross without duty, the goods move as an import, and an import has an importer. We covered the entry mechanics of that change in an earlier note on US de minimis rules, and the Union's per-line-item charge on low-value parcels in another, so this article stays on what it means for the seller's operating model.

What local stock asks of a seller who has never imported

1. A real entry, with a named party. Bulk stock arriving into a destination-country warehouse is declared by someone with the standing to declare it. If you have never filed an entry, you now need a broker, a classification decision per product family, and an answer to who signs as importer of record. On the US side that answer also decides who signs a certificate of compliance; on the EU side it decides whether you have the responsible economic operator your labelling requires.
2. Inventory that is yours in a country you do not staff. The capital moves forward: you buy production for a forecast rather than for an order. Two numbers become yours by default, the sell-through rate per warehouse location and the exit route for a SKU that does not clear. Dead stock in a bonded or third-party facility is a monthly fee and a disposal decision, not a shelf you can ignore.
3. A service clock you have to staff for. Platforms set standards for local sellers on dispatch and delivery performance, and they enforce them with listing consequences. The operational reality behind a promise is a cut-off time, a pick queue and a carrier pickup that has to arrive. Sellers who treat the local warehouse as storage rather than as an operation tend to discover this in their first slow week.

Semi-managed: what the platform keeps, what it hands over

The clearest public description of the model we have found is an analyst note from ChinaTalk, which describes merchants shipping product into the consumer's country or region, largely by sea, storing it in local warehouses and fulfilling from there, while the platform retains pricing and front-end control. That split is the whole story: the platform keeps demand, you take the supply-side work, and the handover point is the moment the goods become an import you are responsible for.

Two consequences follow that are easy to miss from a dashboard. First, quality problems become yours in a market where returning them is not trivial; the reverse route needs to be planned before the forward one, because a rejected pallet sitting in a rented bay has no self-destruct button. Second, your supplier's documentation now travels further: the same test report that satisfied a parcel-route buyer may be the thing a customs or marketplace auditor asks for by batch, tier and date.

What each mode does to your cost shape, without naming a single rate

We are not going to tell you which mode is cheaper, because it is not a comparison anyone can win in the abstract. What we can describe is where the money sits. A direct-parcel model concentrates cost in per-shipment clearance and duty handling, spreads inventory risk across many small lots, and leaves the delivery promise distant. A local-stock model moves that cost into bulk entry, storage handling and inventory carrying, buys a short delivery promise, and adds a forecasting obligation you cannot delegate to a warehouse.

Every seller who switches mode believes they are changing logistics. Most are also changing their balance sheet and their compliance surface at the same time, which is why the switch is better made for a bounded set of fast-moving SKUs than for a catalogue.

Where our own network stops

Being straight about the boundaries, since this page is read by people choosing a structure. Our overseas stations store, pick and dispatch, and they support delivery and replenishment including into Amazon's inbound network. They are not intake points for customer returns and they do not take goods back for inspection. A marketplace model that assumes easy local reverse logistics needs a separate answer for that leg, and it should be settled before the first container of stock is booked rather than after the first defective batch is reported.

Four questions before you choose

1. Which mode names you, or an entity you control, as importer of record on the destination side? If you do not know, that is the answer for now.
2. Can you produce, per SKU family, the compliance documents an importer would be asked for?
3. How many weeks of stock are you prepared to carry per location, and what happens to the remainder?
4. Who handles a return, in the destination market, in writing?

Question one is the hinge. Answer it and the other three are planning; dodge it and they arrive anyway, with a deadline attached.