Amazon Seller Wholesale Accounting: A Two-Channel Setup

Most Amazon sellers keep books that are really a bank feed with categories on it. Money lands from Amazon every two weeks, that number gets called revenue, and the year adds up. It works because on Amazon the deposit is close enough to the truth. The moment a brand invoices a retailer on terms, it stops working, and the reason is worth understanding before touching the chart of accounts.

The settlement deposit is not revenue. It is revenue minus referral fees, minus fulfilment fees, minus storage, minus advertising, minus refunds, plus or minus reserve movements, for a period that does not line up with a calendar month. On a single channel that is a tolerable simplification. With a second channel it produces books that cannot answer the only question that matters: which channel is actually making money.

Why the settlement deposit stops being revenue

Wholesale has a gap that Amazon does not. You ship goods in week one, invoice on net 30, and get paid in week six. Nothing lands in the bank at the moment the economics happen.

Run cash basis books through that and the distortions compound. A month where you shipped four large purchase orders and collected nothing looks like your worst month of the year. The month the checks clear looks like a triumph, even though no goods moved. Inventory that has already left the building is still sitting in your asset number because nothing triggered the entry. And because the wholesale gap is long, the two channels drift apart until comparing them is meaningless.

Then there is the compounding version of the problem: you cannot see a receivable that is late if receivables do not exist as a concept in your books. The account that has not paid in seventy days looks exactly like the account that paid on time, because neither one shows up until money arrives.

Cash basis to accrual, in practical terms

Accrual sounds like an accountant's word for complexity. In practice it means two rules.

Revenue is recognised when the goods ship, not when the money arrives. Ship a purchase order on the fourteenth, the revenue belongs to that month, and the unpaid amount sits in accounts receivable as an asset until the customer pays. When payment lands, it moves from receivables to cash and touches no revenue account at all.

Costs are recognised in the same period as the revenue they produced. The cost of the units on that purchase order leaves inventory and becomes cost of goods sold in the month you shipped, regardless of when you paid the factory. Fees that relate to a sale belong in that sale's month.

That is the whole shift. It costs more bookkeeping discipline and it buys the ability to say, at the end of a month, what each channel earned, without waiting six weeks for the picture to resolve. If the cash timing itself is the worry rather than the recognition, Amazon payouts against net 30 works through the gap directly.

A chart of accounts for a two-channel brand

The single most useful change is splitting revenue and cost of goods by channel from the first wholesale order, rather than retrofitting it later. Here is a skeleton that covers what a two-channel brand needs.

Two entries in that table carry most of the value. The first is that Amazon revenue is gross, with fees as their own line, so the fee load is visible as a number rather than hidden inside a net figure. The second is that allowances and chargebacks are contra-revenue, not expenses.

That second one deserves a sentence. When a retailer deducts money from your invoice for a late shipment or a labelling error, it is tempting to book it as an admin expense and move on. Booked that way, it disappears into overhead and nobody ever sees it. Booked as contra-revenue, it sits directly against the channel that generated it, and your net wholesale revenue tells the truth. A brand that is quietly losing several points of margin to deductions finds out in the first month rather than the second year.

Landed cost, and valuing inventory across channels

Landed cost is what a unit really cost to have sitting in a warehouse ready to sell: the factory price, inbound freight, duty, customs brokerage, any inspection fee, and prep or labelling done before it reached the shelf. Not the invoice from the factory on its own.

Get this number per unit per production run, not as an average across all time. Two runs of the same SKU can differ by a meaningful amount because one came by sea and one by air, or because the freight market moved. If the books use one blended cost, every margin comparison is slightly wrong in a direction you cannot see.

The cross-channel rule is that landed cost does not change by channel. A unit costs what it costs. What differs after that point is the cost to serve: Amazon fees and storage on one side, outbound freight and any retailer-specific packaging on the other. Keeping those separate is what lets a brand compare unit economics across channels honestly instead of assuming wholesale margin is worse because the price is lower.

The AR aging report and what to do with each bucket

Accounts receivable aging is a report that lists every unpaid invoice by how long it has been outstanding. Once a brand sells on terms, this becomes the single most important report in the business, and it should be looked at weekly, not monthly.

Current, meaning not yet due, needs nothing beyond confirming the invoice was actually received by accounts payable. That sounds trivial and it is the cause of a large share of late payments; the invoice went to the buyer, and the buyer is not the person who pays.

One to thirty days past due is normal friction in retail and deserves a short, non-escalating email with the invoice attached again. Thirty-one to sixty gets a phone call to accounts payable rather than to the buyer, and a direct question about whether the invoice is approved for payment and in which run. Beyond sixty, stop shipping. Continuing to ship into an unpaid balance is how a brand turns a collection problem into a write-off, and the conversation to have is about a payment plan on the old balance before anything new goes out.

Track one number off this report every month: the average days it actually takes an account to pay, as against the terms on the invoice. An account on net 30 that reliably pays in 48 days is a net 45 account, and knowing that changes how you plan cash rather than how you feel about them.

Reconciling the Amazon settlement against the ledger

With gross revenue and separate fee accounts, the settlement report stops being a deposit and becomes a reconciliation.

Pull the settlement for the period, and split it into its components: product sales, refunds, referral fees, fulfilment fees, storage, advertising, reimbursements, and reserve movement. Each component posts to its own account. The sum of those postings has to equal the deposit that hit the bank. When it does not, the difference is almost always a reserve movement or a settlement period that straddles month end, and the fix is an accrual entry rather than a hunt.

Do this every settlement, not every quarter. The job takes fifteen minutes on current data and hours on stale data, and a bookkeeping template makes it repeatable. This is also the point where Amazon seller cash flow stops being a feeling and becomes a forecast, because you know exactly what the next deposit is made of.

Inventory in three places at once

A two-channel brand holds stock in at least three physical positions and the books need to know about each: units in Amazon fulfilment centres, units in your own or a partner warehouse for wholesale shipments, and units in transit from the factory.

Give each location its own inventory sub-account, valued at landed cost. Every transfer between them is a movement, not a sale, and the total across all three is what appears on the balance sheet. Units in transit are real inventory the moment the title passes, which depends on the shipping terms you agreed with the factory, and a brand that books them only on arrival will show a hole in its asset number every time a container is at sea.

Count physically at least once a quarter in your own warehouse, reconcile against Amazon's inventory reports, and investigate a variance rather than adjusting it away. A consistent shortfall in one location is usually a process problem, most often samples leaving without an entry, and finding that early is worth more than the units.

The monthly close checklist

A close that takes a morning and produces numbers you trust looks like this.

• Reconcile every bank and card account, and every Amazon settlement that fell in the period
• Post revenue for all purchase orders shipped in the month, whether invoiced or not
• Relieve inventory and post cost of goods sold for those shipments at landed cost
• Post the period's fees, freight in, freight out and advertising to their own accounts
• Book allowances, chargebacks and credits as contra-revenue against the right channel
• Reconcile inventory across all three locations to the balance sheet number
• Run the AR aging, work the buckets, and write down the true average days to pay
• Produce a profit and loss by channel and compare it to the previous two months

The last line is the point of all of it. A brand that can lay Amazon net margin next to wholesale net margin, after fees, freight and deductions, can decide where the next production run should go. Without it, the decision gets made on gut feel, and the gut usually favours whichever channel paid most recently. Feeding that comparison into the forecast for both channels is what turns a set of books into a plan.

If the blocker is that there are no wholesale accounts to account for yet, paste a product link into WholesalePilot and the preview shows which retailers and distributors would plausibly stock it.