3,000 Units in China: FBA, AWD, or Staged Restocking?

“Send everything” can reduce the number of international shipments, but it also starts overseas storage exposure earlier. “Send small batches” reduces early US inventory, but creates more freight events and a longer chain between the reserve and the customer. AWD adds a domestic bulk-storage and replenishment layer, but it is not free storage and it introduces its own inbound, outbound, transport, eligibility, and configuration rules.

On 26 October 2026, an Amazon seller described on Reddit 3,000 completed units of one SKU in China and an expected sell-through period of roughly one year. The seller was comparing FBA, AWD, and China storage while different calculators produced different cost views. The post identifies a real planning problem. It does not provide unit dimensions, cartons, margin, demand variability, Seller Central fee preview, AWD region, capacity, or receiving history, so it cannot support a final allocation.

Turn 3,000 units into a monthly inventory flow

The simple average is 3,000 ÷ 12, or 250 units per month. Use that only as a starting point. A new product may ramp slowly, advertising can create spikes, seasonality can concentrate sales, and stockouts can suppress the very history used to forecast demand.

Create three demand cases: conservative, base, and high. For every month, record beginning sellable FBA units, units in transit to FBA, AWD units, China units, planned replenishment, actual sales, damaged or unavailable units, and ending stock. The model should show where inventory physically sits, not merely a total of 3,000.

Then add time. Measure China warehouse processing, international transit, destination handling, Amazon appointment and receiving, and any internal AWD-to-FBA movement separately. Use the seller's actual historical ranges where available. A fixed generic lead time hides the point at which an order must be released.

Finally, add a service-level decision. The seller may choose a higher buffer for a launch, ranking-sensitive product, or long production replacement cycle. A bulky low-margin product may justify a lower buffer and stricter replenishment cap. The buffer is a business choice informed by variability, not a universal number.

Give FBA, AWD, and China inventory distinct jobs

FBA inventory serves immediate customer orders. Its target should cover expected demand through the next reliable replenishment plus a deliberate safety allowance. Sending a full year of slow-moving units to FBA may expose the later units to monthly storage and ageing-related charges long before they sell.

AWD inventory is bulk inventory inside Amazon's US distribution network. Amazon states that AWD can replenish the FBA network automatically or through seller-requested movements. Its official fee materials separate storage, inbound processing, outbound processing, and transportation. Rates and discounts can depend on region, replenishment mode, programme criteria, and effective date.

China inventory is the reserve that has not yet entered the US warehouse network. It can be inspected, labelled, cartonised, and dispatched in planned batches. It delays US storage exposure and preserves the option to change packaging or labels, but each replenishment must cross the international and Amazon receiving chain.

The three layers can coexist. Define a maximum and minimum for each rather than asking which single location should hold all 3,000 units. A practical policy might maintain a near-term FBA range, use AWD for a domestic buffer if eligible, and release China stock when the combined US supply approaches a trigger. The actual quantities must come from the SKU's data.

Put every option on the same cost perimeter

Begin with international inbound cost from China. Use the final carton count, gross weight, dimensions, destination, service, and quote date. If different batch sizes change the mode or minimum charge, model each batch rather than multiplying one unit rate.

For AWD, include inbound processing, storage, outbound processing, and transportation to FBA. Amazon's official 2026 materials show that these are separate categories. Do not compare AWD's storage line with FBA's complete cost and conclude that AWD is cheaper.

For FBA, include monthly storage, applicable storage-utilisation effects, aged-inventory surcharge exposure, and other fees relevant to the seller's account and product. Amazon states that aged-inventory surcharges apply in addition to monthly storage and are assessed from inventory age within the fulfilment network under current rules.

Add the operational cost of errors: label correction, carton non-compliance, receiving discrepancies, removal, disposal, or delayed availability. These are scenario inputs, not guaranteed charges. Use the seller's records where possible.

Finally, show the economic effect of stockout and overstock separately. Stockout can lose sales and disrupt advertising or rank, while overstock ties up cash and creates storage exposure. Avoid inventing a percentage. Use the product's contribution margin, observed conversion, and planned promotion assumptions.

The model should produce a monthly cash and unit view. A low one-time ocean or air rate does not prove that an all-FBA plan is cheaper after eleven months of inventory ageing. Conversely, avoiding US storage does not prove that twelve small international batches are economical.

Use replenishment triggers instead of calendar habits

A basic trigger is expected demand during replenishment lead time plus safety stock, minus FBA sellable and confirmed inbound units. The seller can then decide whether the replenishment should come from AWD or China.

Amazon explains that AWD auto-replenishment uses forecast demand, weighted lead time, and seller-provided guardrails to determine movement toward FBA. Sellers can use Amazon's recommended level, set minimum or maximum limits, or manage movements manually, subject to current programme conditions. Manual and automated options do not have identical capacity and fee effects.

Treat in-transit stock cautiously. A created shipment is not the same as received and sellable inventory. Track China dispatch, export handover, arrival, Amazon delivery, receiving, transfer, and available status. The trigger should distinguish each stage.

For China replenishment, include enough time for warehouse release, international transit, destination delivery, and Amazon receiving variability. Update the trigger with actual cycle data after each batch. If three batches consistently take longer at Amazon receiving than planned, the model should change; the seller should not keep the original assumption because it was convenient.

Review the plan every month. Replace forecast with actual sales, record the latest unit volume and fee preview, and shift the next release. A rolling model adapts to a faster or slower sell-through without moving the whole reserve prematurely.

Test three strategies and one hybrid

All to FBA minimises the number of upstream replenishment decisions and can support immediate availability, but it front-loads inbound and exposes slow-moving units to FBA storage and ageing. Test it when demand is strong, predictable, and supported by margin and capacity.

All US reserve through AWD places the bulk closer to FBA and can use automatic or manual replenishment. Test SKU eligibility, carton requirements, region, storage tier, processing, transportation, and the exact auto-replenishment configuration. Amazon also states that inventory cannot simply be moved back from FBA into AWD, so an over-allocation to FBA is not reversed through that path.

Staged replenishment from China delays US storage and preserves flexibility. Test multiple freight batch sizes because minimum charges and mode changes may make very small releases inefficient. Include the larger safety stock needed for the longer replenishment chain.

Hybrid inventory assigns each layer a job. An initial FBA quantity supports launch or current sales, AWD holds a shorter US buffer when its full cost and eligibility work, and China holds the remaining reserve. The seller reviews the split monthly. This often creates more control than a single-location decision, but it also requires clean inventory and shipment records.

YANCHAO can support the China-side work: batch inventory records, instructed inspection, labelling, cartonisation, measured shipment data, dispatch planning, and international tracking references. Amazon capacity, receiving, AWD eligibility, rates, and replenishment decisions remain governed by the seller's account and Amazon's current systems.

Build the spreadsheet before moving stock

Use one row per month and separate columns for demand, FBA beginning and ending stock, FBA inbound, AWD stock, China stock, each transfer, and every cost category. Add a source column identifying whether a figure came from Seller Central, an invoice, a warehouse measurement, a freight quote, or a planning assumption.

Mark assumptions clearly. Unit volume, cartons, margin, demand variability, and lead time are high-impact inputs. Run sensitivity tests by changing one at a time. If a small change in unit volume reverses the AWD decision, obtain a reliable packaged measurement before shipping. If a one-week receiving delay creates a stockout, increase the buffer or release earlier.

Do not use a forum calculator result as the final rate. Import current Seller Central fee preview, AWD fee page, inventory reports, and account charges. This article also publishes no live YANCHAO freight price because the required packed dimensions and shipment-specific inputs were not available and the real-time freight tool could not be accessed in this environment.

FAQ: Practical Follow-up Questions

Can the seller simply send 250 units every month?

Not safely from the average alone. The batch must cover replenishment lead time, demand variation, safety stock, carton efficiency, and Amazon receiving. A rolling trigger is more reliable than twelve equal calendar shipments.

Is AWD always cheaper than holding slow inventory in FBA?

No. Compare AWD storage, inbound and outbound processing, transportation, eligibility, and applicable discounts against FBA monthly storage, ageing exposure, and other account-specific fees. Use the same unit volume and month.

Can staged China shipments cost more than the storage they avoid?

Yes. More batches can increase freight and handling. Compare the incremental cost of staged inbound with the US storage and ageing cost avoided, then show stockout risk separately. The answer depends on volume, margin, lead time, and demand.