Most accounting writing treats sales channel as a reporting dimension. Something you slice the P&L by after the fact.
In a business that manufactures and sells through e-commerce, retail, wholesale, and direct to consumer at the same time, channel is not a dimension. It changes the accounting itself: when revenue is recognized, what gets deducted from it, how cost is assigned, and what the word margin actually means.
The consequence is that consolidated margin, the number most people look at first, is the least informative number available. It averages four genuinely different economics into one figure that is technically correct and operationally useless.
Here is where the complexity actually lives.
Revenue recognition varies by channel
The five step model is the same everywhere. What differs is the answers it produces.
DTC and e-commerce. Control typically transfers at shipment or delivery, and the main complications are a returns reserve and, if you sell gift cards, a deferred revenue balance with a breakage estimate attached. Relatively clean.
Retail. This is where it gets interesting. Chargebacks, markdown allowances, and co-op advertising are not expenses. They are variable consideration, which means they reduce the transaction price and have to be estimated at the point of sale rather than recorded when the deduction eventually arrives. Treating them as expenses overstates revenue in the current period and understates it later, and it makes every channel comparison wrong.
Wholesale. Volume rebates and tiered pricing are also variable consideration requiring estimation. If a customer earns a better rate once they cross a threshold, and they are on pace to cross it, the expected rate is the one to accrue, not the rate currently being invoiced.
Marketplace arrangements. These raise the principal versus agent question, which determines gross versus net presentation. Getting it wrong does not change net income by a cent and changes reported revenue enormously. It is one of the few accounting decisions where the wrong answer is highly visible to anyone who looks.
Deductions are where the margin goes
If there is one thing to take from this, it is that deductions deserve more attention than they usually get.
Trade spend, allowances, chargebacks, returns. They arrive late. They are frequently disputed. They are often documented in a format that bears no resemblance to your invoice. And they relate to revenue recognized in a period that has already closed.
The failure mode is predictable: a strong month followed by a weak one, where neither number was real. The strong month was missing its deductions and the weak month absorbed two months of them.
Accrual discipline is the only defense, and it is not glamorous work. It means estimating deductions at the time of sale based on the actual terms in the actual contract, maintaining the estimate as evidence accumulates, and tracking the difference between accrued and settled by customer so the estimate improves. A deduction accrual that is never compared against what actually settled is a number somebody made up once.
Cost assignment
Revenue is only half the margin. The other half varies by channel just as much.
Standard costing gives you a stable basis for comparison, which is exactly what you need when the same unit flows four different ways. It also drifts, and a standard that has drifted far from actual has not stopped working, it has started lying quietly while the variances absorb the gap.
Inventory valuation has to hold up across every location the goods sit in. A unit at a third party fulfillment center and a unit in your own warehouse should carry the same cost, and reconciling those systems is what keeps that true.
Fulfillment cost differs enormously by channel. A DTC unit carries pick, pack, individual shipping, and a much higher return rate. A wholesale pallet carries almost none of that. If those costs sit in operating expense rather than being assigned to the channel that caused them, DTC will look more profitable than it is and wholesale less.
Which is the argument for reporting margin by channel rather than in aggregate. It is more work to build and it answers questions the aggregate cannot: whether this product is worth making, and whether this channel is worth serving for it.
Inventory in places you do not control
Multichannel selling usually means inventory sitting in a third party fulfillment center, in transit, on consignment at a retailer, and in your own building, all at once.
Ownership does not follow physical possession. Goods at a 3PL are yours. Goods shipped FOB origin become the customer's when they leave. Consignment stock at a retailer stays yours until it sells, which also means it is still your obsolescence risk.
The practical difficulty is counting what you cannot walk up to. A cycle count program that only covers your own warehouse is not covering the majority of the inventory. That requires either reconciling to the 3PL's reported counts on a schedule and investigating differences, or accepting that a meaningful share of the balance sheet is unverified.
Returns
Return rates differ by channel by a wide margin, and so does recoverability. A DTC return usually comes back in sellable condition. A retail return often comes back as part of a bulk reverse shipment in unknown condition.
Presentation matters too. Under current guidance, a return expectation produces a refund liability and a return asset rather than a simple net revenue reduction, and the return asset has to be measured at the expected recoverable value, not the original cost.
A single blended return rate applied across all channels will be wrong for every one of them. The blended rate is an average of four distributions, and no channel actually behaves like the average.
What the close needs to look like
All of the above has sequencing consequences.
The deduction accrual is a gating item. It cannot be completed until sales are final by channel, and revenue cannot be finalized until it is done. That makes it an early, high priority task, not something to work on while waiting for other things.
Inventory valuation depends on the period's receipts being costed, which depends on supplier invoices that frequently arrive after the goods. That dependency needs an explicit accrual rather than a delay.
And the reconciliations that catch channel specific errors, sub-ledger to general ledger, system to physical, receiving to invoice, are the ones that surface a problem while there is still time to fix it rather than after the statements are issued.
The short version
Channel changes revenue timing, revenue amount, cost, and inventory ownership. Report margin by channel, accrue deductions when the sale happens rather than when the deduction arrives, keep standard costs current, and count the inventory you cannot see.
Do that and you can answer the question a multichannel business actually needs answered, which is not "how did we do" but "which of these four ways of selling this product is worth continuing."