T. Brian Jones is co-founder and CTO of Inventory Hero. He leads the engineering behind its Amazon data pipeline, demand forecasting, and the AI platform that lets sellers talk to their live inventory, sales, and supplier data in plain language.
Can I choose which Amazon fulfillment center my inventory goes to?
No. Amazon assigns the destination fulfillment centers when you create the shipment, based on the product, current network stock, and demand. We are not aware of any FBA placement program, current or retired, that let a seller name a specific warehouse. Your ship-from address influences the assignment, since Amazon tends to recommend destinations reachable from where the freight starts, but it is an influence, not a choice.
How do I see my real inbound placement fee before I book freight?
Build a draft shipment in Send to Amazon for the actual SKU and quantity you plan to send, then advance to the Confirm Shipping step. Amazon displays a per-unit placement fee estimate for each available option alongside the estimated carrier cost. Read the numbers and leave the workflow without confirming. The draft is not a commitment, and this turns the placement fee from something you discover afterward into an input you can put into a freight decision.
What is the difference between minimal, partial, and Amazon-optimized shipment splits?
They are the three inbound placement options in Send to Amazon, defined by how many destinations you agree to ship to. Minimal splits is a single location and carries the highest per-unit placement fee. Partial splits is two or three locations at a reduced fee. Amazon-optimized splits is four or more locations and carries the lowest fee, commonly zero. Because the optimized option is usually free, the real question is not which fee to pay but whether your own cost of shipping to four destinations exceeds the fee Amazon quotes for shipping to one.
Does paying the inbound placement fee get my units sellable faster?
It gets them sellable sooner but distributed later. Minimal splits means one delivery and one receiving queue, so units become sellable when that single shipment checks in, but Amazon then trans-ships them across the network and the fast delivery promise in distant regions lags. Amazon-optimized splits puts inventory in several regions on arrival, so national coverage starts earlier, but you are exposed to the slowest of four or more independent check-ins. Plan reorder timing on the last destination to check in.
Should my placement strategy change before Q4?
Yes, and in a specific direction: from roughly early October, favor the option with the fewest independent receiving events you have to chase, even if it costs the placement fee. Receiving variance is the binding constraint at peak, and the slowest of four check-ins has a much longer tail than one. The bigger move is to send earlier so no option's slippage matters, subject to your FBA capacity limits.
An Amazon inventory placement strategy is your standing rule for how many destinations you inbound to, where you ship from, and how far ahead you send, given where your demand actually sits. The uncomfortable first fact: you do not get to pick fulfillment centers. Amazon assigns them.1 What you control is the split option, your ship-from address, your timing, and whether you route through AWD. This article is about that routing decision. For the mechanics of the fee itself and how to reduce it, read our FBA inbound placement fee guide; this one picks up where that leaves off.
Because the cheapest inbound is frequently the slowest one, and slow inbound in Q3 turns into a stockout in Q4. Treating placement purely as a fee line optimizes the smallest number in the decision.
Inbound Placement Strategy
A seller's standing policy for how inbound FBA shipments are routed: how many destinations to accept, where freight originates, how far ahead of demand to send, and when to use AWD instead. It sits above the per-shipment fee decision and is driven by transit-time risk and calendar risk, not by fee minimization alone.
Amazon's own 2026 rate change made the trade sharper. Amazon announced on October 15, 2025 that FBA fees would rise by an average of $0.08 per unit sold, effective January 15, 2026.2 Seller-tool summaries of the detailed 2026 schedule report that minimal-split placement fees for standard-size items rose about $0.05 per unit on average and that large standard items in the 3 to 20 lb range were broken into five weight bands, so heavier units inside a tier now cost disproportionately more.3 We could not confirm those two specifics on an Amazon-published page, and the third-party rate cards we checked disagree with each other on the actual per-unit ranges. Treat any rate card you find online as a rumor. The next section is how to replace it with your own number in about five minutes.
This is the step most sellers skip, and skipping it is why the placement fee usually shows up as a surprise after the freight is already booked. In Send to Amazon, create a shipment for the actual SKU and the actual quantity you intend to send, work through packing details, and advance to the Confirm Shipping step. Amazon displays a per-unit placement fee estimate for each available placement option there, alongside the estimated carrier cost.4 Write the numbers down. Then leave the workflow. A draft shipment is not a commitment, nothing is charged, and you can delete it or come back to it later.
That single habit changes the shape of the decision. The break-even further down this page is an inequality with two sides:
extra freight and labor per unit < minimal-split placement fee per unit
The right-hand side is Amazon's to publish and it varies by your SKU, dims, weight band, and current network state. Nobody on the internet can give it to you. But you can read it yourself, for free, before you ask a carrier for a single quote. Do that first, then go get lane quotes knowing what number you are trying to beat.
The three options, by the count of destinations you accept:5
Option (Send to Amazon label)
Destinations you ship to
Placement fee
What you are actually buying
Minimal shipment splits
1
Highest per unit
Amazon doing the network distribution for you
Partial shipment splits
2 or 3
Reduced
Fewer lanes to manage, part of the fee back
Amazon-optimized shipment splits
4 or more
Lowest, commonly $0
Nothing; you do the distribution yourself
Read that table carefully, because it is not a three-way price comparison. The optimized option is commonly free. So the honest framing is: Amazon's default costs you nothing in fees, and the placement fee is the price of a convenience upgrade you buy from Amazon when shipping to four destinations is expensive or impractical for you. Sellers pay it for concrete reasons: one LTL lane out of a prep center instead of four, a shipment too small to divide into four sane pallets, a carrier relationship on one route, or a calendar too tight to chase four receiving queues.
The middle tier is the one sellers forget, and it is modeled alongside the other two below. If four destinations is operationally painful but one is expensive, two or three is often the actual answer.
Run it per shipment, in dollars per unit, with both the placement fee and the freight delta as variables you supply. Take a container-sized send: 2,400 units of a 1.2 lb large-standard SKU, 8 pallets, moving out of a Los Angeles-area prep center.
The only thing we can compute for you is the arithmetic that turns a total freight-and-labor premium into a per-unit break-even. So here it is as a sensitivity table rather than a fabricated rate card. Find the row closest to your own quote spread:
Your total premium for splitting (extra freight + BOL + appointments + pallet labor)
Break-even minimal-split fee per unit, 2,400 units (approx)
Same premium at 6,000 units (approx)
$240
$0.10
$0.04
$480
$0.20
$0.08
$720
$0.30
$0.12
$1,200
$0.50
$0.20
$2,400
$1.00
$0.40
To use it: get your one-lane quote and your four-lane quote, subtract, add roughly $25 per extra destination for BOL, appointment, and pallet labor, and find the row. If the minimal-split fee Send to Amazon quoted you is above that row's break-even, split. If it is below, pay the fee and ship one lane.
We deliberately do not publish "the" freight numbers. LTL pricing swings several-fold across lanes, density, accessorials, fuel, and season, far more than the placement rate cards vary, so a worked example built on invented freight would be a worse number than the one we refused to print.6 Size the pallet count first in the FBA pallet calculator, or the freight container calculator if you are planning the container behind it.
Three things fall out of the table that matter more than geography:
Shipment size dominates. The same premium spread over 6,000 units instead of 2,400 is 40% of the per-unit cost. Big sends push almost everyone toward splitting. Small sends push almost everyone toward paying the fee.
Partial splits usually sit closer to optimized than to minimal on freight. Going from one destination to two roughly doubles your lane count but often less than doubles your cost, because the second lane still moves 4 pallets and holds decent density. The jump from two to four is where per-lane density collapses. Price partial explicitly, do not assume it is the midpoint.
You cannot always divide freely. A 600-unit replenishment at 120 units per pallet is 5 pallets. Split across four destinations and you are shipping 2, 1, 1, and 1, which means three lanes moving a single pallet each. Carriers price a partial or lone pallet against a minimum charge, so your per-lane cost does not fall in proportion to the freight you removed from it. Below roughly one full pallet per destination, the split premium stops behaving linearly and the fee usually wins.
For a pallet-shipping seller, this is the strongest answer on the page, and it is the only lever here that sidesteps the placement trade rather than pricing it.
Send bulk into Amazon Warehousing and Distribution and you inbound to AWD, not to FBA. Replenishment from AWD into FBA is then Amazon's move across its own network, priced as a published transportation into FBA rate of $1.40 per cubic foot base, which covers FBA inbound placement.7 There is no split option to choose on that leg, no four-lane quote to chase, and no per-shipment inequality to solve.
That is not free, and it is not always cheaper. You are now paying AWD storage on top ($0.48 per cubic foot per month in the East, Southeast, and South Central regions, $0.57 in the West, before smart-storage or Amazon-managed discounts) plus $1.40 per box inbound and outbound processing.7 The comparison to run is total landed-into-FBA cost per unit through AWD versus your own split-or-pay number from the table above, which is the AWD versus FBA decision in a different frame.
AWD tends to win when you are importing in container quantities, your reserve stock would otherwise sit in FBA against a capacity limit, and you would be re-solving the split decision on every replenishment anyway. It tends to lose when your whole inventory comfortably fits in FBA and turns quickly.
This is the part of the decision with no fee attached, and it is worth more than the fee.
The split option changes when units become sellable, which is the only date your reorder math cares about. Two effects pull in opposite directions:
Minimal splits is one delivery and one receiving queue. Units are sellable when that shipment checks in. Amazon then trans-ships internally, so fast delivery promises in distant regions lag, but your sellable date is governed by a single event you can see.
Amazon-optimized splits puts inventory in several regions on arrival, so national coverage starts earlier, but your inventory is not fully deployed until the slowest of four or more independent check-ins clears, each in its own queue.
The operator rule: plan the reorder date on the last destination to check in, not the first. A shipment is not "in" when the first FC scans it. Here is how to actually run that.
Where to observe the lag. Every FBA shipment carries a per-destination status in Seller Central under Shipping Queue, and the shipment detail view timestamps when each destination moved to Receiving and to Closed. For each split shipment you have already sent, record two dates per destination: carrier delivery date and the date units first appeared as sellable. The gap is your check-in lag for that destination.
What to do with the numbers. Build a small history, ten to fifteen shipments is enough to be useful:
For each shipment, take the maximum check-in lag across its destinations, not the average. A four-destination shipment's real lag is its slowest leg.
Take the median of those maximums as your planning lag, and the 80th or 90th percentile as your buffer case. Do not use the mean; check-in lag is skewed by a long right tail, exactly like lead-time variability on the supplier side.
Fold the planning lag into total lead time as its own component, alongside production, ocean or air transit, and prep. It is not part of transit time and it is not part of prep. It is its own stage, and it is the one that grows in Q4.
Carry the gap between your median and your 90th percentile as buffer stock expressed in days, then convert to units with your days of supply math.
Why slowest-of-N gets worse with N, not better. If each destination independently clears receiving within your planning lag 85% of the time, one destination hits that lag 85% of the time and four destinations all hit it about 52% of the time. The number of destinations is a direct multiplier on the chance that at least one leg runs long. That is not an argument against splitting. It is an argument for sizing the buffer to the option you actually chose: a four-destination send needs a wider buffer than a one-destination send of the same units, and if you plan both on the same lag you will be late on the split ones and not know why.
Keep this visible alongside your other stock locations with multi-location inventory tracking, and track it the way you would any other lead-time component. Receiving queues are also the reason FBA capacity limits and placement interact: capped space and a long queue together mean the units you sent are neither sellable nor removable.
A little, and less than you would hope, so keep the analysis proportional.
You can pull the Amazon-Fulfilled Shipments report in Seller Central (Reports > Fulfillment > Sales > Amazon Fulfilled Shipments), which carries ship-city, ship-state, and ship-postal-code alongside sku and quantity-shipped, capped at about a month of data per request.8 Roll shipped units up by state and you can see roughly where your customers are.
Before you act on it, three things have to be true, and often only one is:
Enough units. A regional share computed on a few hundred units is noise. Below roughly 1,000 units in the window, and at least a couple of hundred in the region you think is hot, do not act on the difference at all.
The right baseline. US population share (South 38%, West 24%, Midwest 21%, Northeast 17%)9 is a crude proxy, and it biases toward false positives, because Amazon spend per capita is not uniform: Prime penetration, income, urban density, and category mix all skew it. If you sell more than one SKU, the better baseline is your own catalog's national mix, which already absorbs those effects. Compare a SKU against your other SKUs, not against the Census.
No seasonal or climate confound. Patio furniture, snow gear, and hurricane supplies are regionally non-uniform by design and by month. A skew that is really a season is not a routing signal.
And then the honest caveat: your observed geography is partly an artifact of your own placement. Conversion follows the delivery promise, and the delivery promise follows where Amazon happened to put your stock, so a region where you were repeatedly out of position under-reports demand it actually had. Exclude stockout days using the Inventory Ledger before computing shares, the same way you would for restock planning.
The one lever it connects to. If a genuine skew survives all of that, the thing you change is not the split option, it is the ship-from address, because that influences which destinations Amazon recommends.1 Concretely: a seller importing through Long Beach whose demand runs 35% Northeast against a 20% catalog baseline should price a second prep relationship on the East Coast, or a transload that puts freight on rail eastbound before it is inbounded, so that Amazon's recommended destinations start closer to the customers. That is a supplier and 3PL decision with real switching cost, which is precisely why it should be triggered by a durable skew and re-checked once a quarter, not read off a monthly report.
From roughly early October, availability beats fee arbitrage, and the tie-break flips toward fewer destinations.
That is the opposite of the year-round default below, so here is the defense. At peak the binding constraint is receiving variance, not trans-ship time. Minimal splits gives you one check-in: when it clears, the units are sellable, and Amazon's internal trans-ship degrades your delivery promise in far regions but does not make the units unsellable. Optimized splits gives you four independent draws from a queue distribution whose right tail is getting fatter by the week, and your inventory is not deployed until the slowest one clears. Trading a slightly worse delivery promise for one receiving event you can actually chase is the right trade in November. One delivery you can chase beats four you cannot.
Two things people get backwards about peak:
Storage is an argument against sending too much early, not for sending early. The standard-size monthly rate steps from $0.78 to $2.40 per cubic foot for October through December, then reverts in January.10 But on a 0.2 cubic foot unit that is about 32 cents per unit per month of difference. It is a reason not to park six months of cover in FBA in October. It is not the reason you cannot send early. The real constraints on sending early are your FBA capacity limits and, on the other side, the low-inventory-level fee threshold that penalizes running too thin. Plan against those two, not against the storage rate.
The lost-week comparison is not as lopsided as it looks. A SKU doing 40 units a day at $12 contribution loses about $3,360 in a stocked-out week. The placement fee on a 2,400-unit send at $0.40 per unit is $960. But the $3,360 is not certain and the $960 is, so comparing them directly is not sound. Risk-weight it with your own numbers: if paying the fee cuts your chance of a late week from, say, 15% to 5%, you bought $336 of expected value for $960 and it was a bad trade on those inputs. Change the inputs to 30% and 10% and it is a good one.
The conclusion that survives the arithmetic is not "always pay the fee at peak." It is buy buffer weeks, not fee cents. A week of extra cover, sent in September, dominates every version of this comparison, because it makes the slippage question stop mattering. Spend your October attention on the send date and the buffer, and let the split option be the tie-break it actually is.
There is precedent for Amazon leaning on this decision with pricing. In October 2024, Amazon told sellers it was seeing tight capacity at some West Coast facilities and temporarily cut the minimal-split inbound placement fee range by about 5 cents per unit for inventory destined for the Eastern US.11 We found no equivalent Amazon announcement for 2026 as of August 2026, so do not plan a 2026 shipment around an East Coast discount. Do plan around the pattern: when Amazon wants volume steered somewhere, it shows up as a changed number in your Send to Amazon quote, which is one more reason to read that screen every shipment instead of memorizing a rate.
Pick a default per SKU family rather than re-litigating every send.
Default to Amazon-optimized splits for large sends, because the per-unit split premium falls fast as shipment size rises and the fee on the other side is usually zero.
Drop to partial or minimal splits when the shipment is small enough that destinations fall below about a pallet each, the freight lane is awkward, or the calendar is tight.
Route through AWD when you are importing in container quantities and would otherwise re-solve this every replenishment.
Always build the draft shipment first and read the per-option quote at Confirm Shipping before you ask a carrier for anything.
Size the buffer to the option you chose, using the slowest-of-N method above, and re-pull your regional shares once a quarter rather than once a month.
From October, stop optimizing the fee, favor fewer receiving events, and send earlier against your capacity limit.
Amazon assigns the destination fulfillment centers for an inbound shipment; sellers select a placement option (how many destinations), not named warehouses. Amazon Seller Central help, "FBA inbound placement service fee," accessed August 2026. https://sellercentral.amazon.com/help/hub/reference/GD3CJHDPKMSHRAJU . Amazon's recommended destinations vary with the ship-from address entered in Send to Amazon, which is why changing where freight originates changes the set you are offered. ↩↩2
Amazon Selling Partners, "Update to U.S. Referral and Fulfillment by Amazon fees for 2026," published October 15, 2025. https://sellingpartners.aboutamazon.com/update-to-u-s-referral-and-fulfillment-by-amazon-fees-for-2026 . States "Unless otherwise noted, all changes will be effective January 15, 2026" and that "FBA fees will increase by an average of $0.08 per unit sold, or less than 0.5% of an average item's selling price." The page does not itself detail the inbound placement fee change. ↩
Reported, not Amazon-confirmed. Multiple seller-tool summaries of Amazon's detailed 2026 US fee schedule (Nova Analytics, "Amazon Inbound Placement Fees Jump in 2026"; SKU Compass, "Amazon Inbound Placement Fee 2026," May 10, 2026; AMZ Prep, "Amazon Inbound Placement FBA Fees Explained in 2026," all accessed August 2026) state that minimal-split placement fees for standard-size items rose about $0.05 per unit on average effective January 15, 2026, and that large standard 3 to 20 lb items were split into five weight bands. We could not verify either specific on an Amazon-published page, and the per-unit rate ranges published by those sources conflict with one another, so no rate card is reproduced in this article. Read your live per-unit estimate in Send to Amazon. ↩
Amazon Seller Central, Send to Amazon workflow: the Confirm Shipping step displays estimated total cost including the FBA inbound placement service fee and estimated carrier cost, with a per-unit estimate shown for each available placement option. A draft shipment can be created and abandoned before this step without charge. Observed in the Send to Amazon workflow and corroborated by Amazon Seller Forums discussions of the placement options, accessed August 2026. ↩
The option names "Minimal shipment splits," "Partial shipment splits," and "Amazon-optimized shipment splits" are the labels shown on the Send to Amazon placement-option screen in Seller Central, not headings on Amazon's public pricing page. Corroborated by Amazon Seller Forums threads referring to the options by those exact names, accessed August 2026, for example https://sellercentral.amazon.com/seller-forums/discussions/t/8623db79-5dfc-434b-bd2b-aafcd2aba4bb . Destination counts: minimal is a single location, partial is two or three, Amazon-optimized is four or more. ↩
No freight rate card is published in this article by design. LTL and drayage pricing varies several-fold by lane, density, accessorials, fuel, and season, which is a wider spread than the third-party placement rate cards we declined to reproduce. Supplying an invented freight number would produce a less reliable break-even than asking you for your own quotes. The sensitivity table is pure arithmetic on a premium you provide. ↩
Amazon Warehousing and Distribution pricing, effective January 15, 2026: base monthly storage $0.48 per cubic foot (East, Southeast, and South Central regions) and $0.57 per cubic foot (West region), lower with smart-storage or Amazon-managed discounts; transportation into FBA $1.40 per cubic foot base ($1.26 with Amazon-managed), which covers FBA inbound placement; inbound and outbound processing $1.40 per box each. https://sell.amazon.com/programs/warehousing . Carried in Inventory Hero's reference-data registry so our articles and calculators agree. ↩↩2
Amazon-Fulfilled Shipments report, Amazon Seller Central help (Reports > Fulfillment > Sales > Amazon Fulfilled Shipments), accessed August 2026. https://sellercentral.amazon.com/help/hub/reference/external/200453120 . Report columns include sku, quantity-shipped, ship-city, ship-state, ship-postal-code, and ship-country; requests are limited to roughly one month of data at a time. ↩
US Census Bureau, 2020 Census regional populations: South 126.3M, West 78.6M, Midwest 68.9M, Northeast 57.6M of 331.4M total, giving approximate shares of 38%, 24%, 21%, and 17%. https://www.census.gov . Population share is a crude proxy for Amazon demand share and is used here only as a fallback when you do not have enough catalog history to build your own baseline. ↩
FBA monthly storage, standard size: $0.78 per cubic foot off-peak (January to September) and $2.40 per cubic foot peak (October to December), effective 2026-01-15, as carried in Inventory Hero's reference-data registry so our articles and calculators agree. Amazon Seller Central pricing, announced 2025-10-15. https://sell.amazon.com/pricing↩
Supply Chain Dive, "Amazon warns of West Coast capacity constraints," October 28, 2024. https://www.supplychaindive.com/news/amazon-west-coast-inventory-placement-capacity/731127/ . Reports that Amazon was "experiencing tight capacity at some West Coast facilities due to high demand," causing lengthier processing times, and that it temporarily reduced the inbound placement service fee for the minimal shipment splits option for Eastern US destinations, with "the fee's rate range now 5 cents lower per unit." This was a 2024 action; we found no equivalent Amazon announcement in 2026. ↩