Operations: fulfilment, packaging & compliance
A retail buyer asking for a pallet is not asking for something your FBA inventory can supply. Those units are in the wrong place, in the wrong packaging, and under the control of a system that charges you to get them back and penalizes your account when you do.
The answer most brands land on, usually after one painful removal order, is to stop treating Amazon stock as a single pool you draw from. You split the production run at the factory and never mix the two streams again. Here is why, what a removal really costs, and how to size the first retail allocation without starving the listing that pays your bills.
Why FBA stock is the wrong pool to sell from
Three things make fulfillment center inventory unsuitable for a retail order.
The first is control. Once units are checked in, Amazon decides where they sit and how they move. You can sell them through Amazon, or you can pay to have them shipped somewhere. There is no third path where a distributor's truck backs up to a fulfillment center door.
The second is packaging, and this is the one that surprises people. Your FBA units are prepped to Amazon's rules: poly-bagged with a suffocation warning, an FNSKU label placed over the manufacturer barcode, loose inside a master carton that carries a shipment label. A retail distribution center wants close to the opposite. It wants the manufacturer UPC readable on the unit, a case label with a GTIN-14, a fixed and unchanging case pack, and cartons stacked to the retailer's pallet pattern. Converting FBA-prepped units into retail-ready cases means touching every single unit by hand.
The third is accounting. The units sitting in FBA are the units your restock limits, your sell-through and your Prime badge are calculated from. Pulling them out to serve a retail order takes something away from the channel that is currently funding the company.
What a removal order actually costs
The per-unit removal fee is the smallest line in the total. Budget for three other costs.
Time. A removal is not a next-day operation. The order has to be created, the units picked across however many fulfillment centers hold them, and the boxes shipped to your address. Plan in weeks. If a buyer gave you a ship window, that window will usually have closed before the last carton arrives.
Condition. Units come back in whatever state the building had them. Scuffed cartons, opened poly bags, and — if you are not careful about which condition you remove — customer returns mixed into what you assumed was clean stock. A DC receiving clerk will reject a case where the units look handled, and a rejected case turns into a deduction.
Labor. Every returned unit needs its FNSKU covered or removed, a decision made about the poly bag, and a count into a case pack. For a thousand units that is somebody's entire week, or an invoice from a prep house. Either way it is a cost you did not quote into the wholesale price.
Removals earn their place in one situation: you already want the stock out of Amazon. Slow movers, an old carton design, units drifting toward long-term charges. Then a retail order is a way to clear them at a better number than a liquidator would pay, and the storage math in amazon FBA storage fees tells you when that point has arrived. What removals cannot be is the foundation of a wholesale channel.
Split the production run at the factory instead
The clean version of this is a single purchase order to your manufacturer that produces two differently packed outputs from one run of the same product.
Stream one is Amazon-prepped: unit poly-bagged if the product needs it, FNSKU applied at the factory, packed loose into master cartons sized for a fulfillment center shipment. Stream two is retail-ready: unit in its retail carton or sleeve with the UPC printed or labeled, packed into a fixed case pack, case sealed and labeled with a GTIN-14 and the case contents, cartons palletized.
Same product, two packing specs, two sets of labels, two destinations from the same container. The incremental cost is the second set of printed cases and the factory's time to pack two ways, not a second production run. Ask your manufacturer to quote it as two pack styles on one PO and you will usually find it is cheaper than anything you could do domestically.
The point worth internalizing is that the retail spec is decided before the goods exist. Once a unit has an FNSKU sticker on its barcode, every downstream retail cost is a repair job.
A worked split on a run of 3,000 units
Take a stainless water bottle. Retail price 26 dollars, wholesale price 11.70, case pack of 12, so a case invoices at 140.40. The factory quotes 4.15 a unit landed, and the run is 3,000 units.
Amazon has been selling roughly 600 units a month and the lead time from PO to checked-in stock is about 90 days, so a straight Amazon-only run of 3,000 would be five months of cover. That is the slack you are working with.
A first retail order from a regional chain is 40 cases, which is 480 units, or 5,616 dollars. Allocate like this: 2,160 units Amazon-prepped, 840 units retail-ready. The retail-ready side covers the 480-unit opening order and leaves 360 units — 30 cases — as a reserve for the reorder, a second account, or samples.
Amazon now has 2,160 units, which is three and a half months of cover instead of five. That is still comfortable, and the next PO gets placed on the same cycle it always was. What you have not done is take a single unit out of a fulfillment center.
Run the same arithmetic on margin before you commit. At 4.15 landed, the retail case at 140.40 returns 90.60 of gross margin across twelve units, before freight to the DC and before deductions. Whether that beats the Amazon unit after referral and FBA fees and ad spend is the calculation in ACoS versus wholesale margin, and it is worth doing honestly, because a retail account that loses money per case is not a channel, it is a hobby.
How do you decide the split for a first retail order?
Four inputs, in this order.
Start with the buyer's actual opening order, not the annual potential they described in the meeting. First orders are small on purpose — the buyer is testing sell-through in a handful of stores before committing chain-wide. Allocate against the PO in front of you.
Add a reserve of roughly the same size again. The reorder, if it comes, arrives faster than your factory lead time. A brand that wins a test and then cannot ship the follow-up order has lost the account, and it happens constantly.
Subtract from Amazon only down to your comfort floor, which is the weeks of cover you need to survive a lead-time slip. If the split would push Amazon below that floor, increase the size of the run rather than cutting Amazon's share. The extra units cost you 4.15 each. Running out on Amazon costs you rank.
Finally, check the cash. A 3,000-unit run at 4.15 is 12,450 dollars leaving your account now, against a retail invoice that pays on terms weeks after delivery. The gap between Amazon's settlement cycle and a retailer's payment cycle is the single most common thing that breaks a first wholesale year, and Amazon seller cash flow walks through how to size it before you sign.
What happens to your Amazon metrics when you divert units
Diverting at the factory rather than from the fulfillment center protects you from most of this, but not all of it.
Restock limits are driven by sales and by historical sell-through at the ASIN and account level. Sending fewer units into FBA does not reduce your limit by itself, but if lower on-hand stock leads to a stockout, the lost sales do eventually show up in what Amazon lets you send.
Sell-through actually improves when you send less. Holding 2,160 units against the same demand turns inventory faster than holding 3,000, which is the direction storage and low-inventory charges reward. The risk is overshooting: going too lean triggers the fee for running thin on stock, described in Amazon's low inventory fee, and one stockout on a top ASIN costs more than a quarter of storage charges.
The Buy Box and organic rank do not care where your other units are. They care whether the listing is in stock and converting. Keep it in stock and the retail channel is invisible to the algorithm.
Where should the retail pool physically sit?
Not in your garage, once the second account lands. The reserve needs a dock, a pallet jack and somebody who can build and wrap a pallet to a routing guide.
The usual answer is a third-party warehouse that handles business shipping rather than parcel ecommerce. What you are buying is not storage — storage is cheap — it is the ability to receive a PO, pick full cases, build a compliant pallet, print carton and pallet labels, book the carrier the retailer's routing guide specifies, and send back a clean bill of lading. Ask any warehouse you are evaluating whether they have shipped to the retailer in question before. The honest ones will tell you, and the answer decides more than their rate card.
Keep the retail pool and the Amazon pool physically and mentally separate. Different packaging, different labels, different location, different reorder trigger. The brands that get into trouble are the ones that keep one pile and decide what it is on the day an order arrives.
Is a retail channel worth the second packaging spec?
For a product with a real shelf presence, usually yes, and the reason is not the first order. It is that a retail account reorders on its own rhythm without ad spend attached, which changes what your revenue is worth and how much of it depends on one listing staying live. The broader case is in Amazon FBA to retail.
The honest prerequisite is that someone actually wants to stock the product. Before you pay for a second set of cases, find out which chains and distributors carry things like yours — paste the listing into WholesalePilot and the preview shows who would plausibly stock it, which is enough to tell you whether the packaging spend has a buyer waiting at the end of it.
Then split the next run. Not the stock you already sent to Amazon — the next run.