Your COGS method decides which purchase price gets attached to the unit you just sold. FIFO assigns the cost of your oldest inventory, weighted average blends every batch into one cost, and cash-basis expensing recognises the entire purchase order when you pay for it. The product is identical and the cash movement is identical — but the three methods produce different gross margins in the same month, different taxable income, and different answers to the question “is this SKU working?” On a catalogue with rising supplier costs, the gap between FIFO and weighted average margin regularly runs three to six points.
Most Amazon sellers never choose a method. They inherit one from whatever their bookkeeping software defaults to, or they use a fourth approach — dividing this month’s inventory spend by this month’s units sold — which is not an accounting method at all and produces margin figures that swing violently with purchase timing.
What do the three methods actually do?
You bought the same product three times as your supplier’s price moved:
| Batch | Date | Units | Landed cost per unit | Batch value |
|---|---|---|---|---|
| A | January | 1,000 | $6.00 | $6,000 |
| B | April | 1,000 | $7.20 | $7,200 |
| C | July | 1,000 | $8.40 | $8,400 |
In August you sell 1,200 units at $24.99, generating $29,988 in revenue.
| Method | COGS on 1,200 units | Gross profit | Gross margin | Ending inventory value |
|---|---|---|---|---|
| FIFO | 1,000 × $6.00 + 200 × $7.20 = $7,440 | $22,548 | 75.2% | $14,160 |
| Weighted average ($7.20) | 1,200 × $7.20 = $8,640 | $21,348 | 71.2% | $12,960 |
| Cash basis (July PO paid in August) | $8,400 expensed in full | $21,588 | 72.0% | $0 on the P&L |
Four margin points separate FIFO from weighted average on the same sales, and the cash-basis figure is an accident of when a purchase order happened to be paid rather than a measure of anything about the product.
Which method should an Amazon seller use?
| Method | Best for | Main weakness |
|---|---|---|
| FIFO | Sellers with batch-level cost tracking, dated inventory, perishables, or products where supplier cost moves materially between orders | Requires tracking which batch each sale came from; overstates margin during cost inflation |
| Weighted average | Commodity-like SKUs, high reorder frequency, sellers wanting stable month-to-month margin | Smooths over real cost changes; a tariff shock takes months to show up in reported margin |
| Specific identification | High-value, serialised, or low-volume items where each unit is individually tracked | Impractical above a few hundred units |
| Cash-basis expensing | Very small sellers below IRS thresholds who prioritise simplicity | Makes monthly SKU profitability effectively unreadable |
For most FBA sellers with rising landed costs, FIFO matches the physical reality — Amazon does broadly ship older stock first — and gives the most accurate picture of what the units sold this month actually cost you. Weighted average is defensible and much easier to maintain if your reorder cadence is fast and your cost variance is small.
LIFO is a real option in US tax law but is a poor fit for Amazon sellers: it is prohibited under IFRS, it conflicts with how FBA physically rotates inventory, and electing it for tax purposes carries a conformity requirement that forces the same method into financial statements given to lenders.
Why is cash-basis expensing so dangerous for SKU analysis?
This is the most common setup among sellers under about $1M, and it is where the worst decisions get made.
Suppose you pay a $42,000 supplier invoice in March. Under cash-basis expensing, March’s P&L absorbs the entire $42,000 against whatever you happened to sell in March. April, May, and June sell that inventory with almost no COGS recorded.
| Month | Revenue | COGS recorded | Apparent gross margin | Reality |
|---|---|---|---|---|
| March | $38,000 | $42,000 | −10.5% | Normal month, big PO paid |
| April | $41,000 | $0 | 100% | Normal month, no PO |
| May | $39,000 | $0 | 100% | Normal month, no PO |
| June | $40,000 | $0 | 100% | Normal month, no PO |
Nothing changed operationally across those four months. The seller sees a disaster followed by three spectacular months and makes decisions accordingly — cutting ad spend in March, scaling aggressively in April. Both decisions are responses to an artefact of payment timing.
The correct treatment is to capitalise inventory on the balance sheet when purchased and release it to COGS as units sell. Whether you can do this for tax purposes is a separate question from whether you should do it for management reporting. You can, and should, run accrual-style COGS internally regardless of what your tax return does.
What do the tax rules require?
Under Section 471(c), businesses whose average annual gross receipts over the prior three years are at or below the inflation-adjusted threshold — approximately $32 million for tax year 2026 — can elect out of the normal inventory rules. Eligible taxpayers may treat inventory as non-incidental materials and supplies, deducting it when used or sold, and skip the uniform capitalisation (UNICAP) rules of Section 263A that would otherwise require capitalising indirect costs into inventory.
Three practical implications:
- Nearly every independent Amazon seller qualifies for the small-business exception. It is a genuine simplification and often produces a one-time timing benefit.
- The exception does not mean you should abandon proper COGS internally. Tax treatment and management reporting are separate systems and can legitimately differ.
- Changing methods after you have adopted one is a formal change in accounting method. It generally requires IRS Form 3115 and produces a cumulative Section 481(a) adjustment. Pick a method deliberately rather than drifting into one.
Talk to a CPA before electing anything — this section describes the mechanism, not advice for your situation.
How do tariffs and freight change the picture?
Whichever method you use, the cost you attach must be the landed cost: supplier invoice plus inbound freight, duty and tariffs, customs brokerage, and prep. Sellers who capitalise only the supplier invoice and expense freight separately as an operating cost systematically overstate gross margin, then wonder why net profit never matches.
Cost volatility is exactly where method choice becomes material. When a tariff change raises landed cost from $6.00 to $8.40 per unit:
- FIFO keeps reporting the old $6.00 cost until pre-tariff stock is exhausted. Margin looks unchanged for two or three months, then drops sharply. Sellers who reprice off FIFO margin during that window are pricing against a cost base that no longer exists.
- Weighted average blends the increase in gradually. Margin drifts down smoothly, which is easier to live with but obscures the size of the shock.
The practical answer for pricing decisions is to use replacement cost — what the next unit will cost to land — rather than either historical method. Use FIFO or weighted average for the books and replacement cost for pricing. Conflating the two is how sellers end up with a profitable P&L and no cash.
Common mistakes
- Dividing monthly inventory spend by monthly units sold. This is not a costing method. It makes COGS a function of purchase timing and produces margin swings that have no relationship to performance.
- Excluding freight and duty from the capitalised cost. Landed cost, not invoice cost. This alone can misstate gross margin by 20% or more of the cost base.
- Switching methods mid-year to make a period look better. Beyond the tax consequences, it makes year-over-year comparison meaningless. Consistency is the point.
- Using FIFO cost to set prices during a cost increase. Your books say $6.00; your next container says $8.40. Price against the replacement cost.
- Ignoring inventory value on the balance sheet. Under cash-basis expensing, inventory disappears from your financial picture, which distorts lender conversations, valuation, and any assessment of how much capital is tied up in stock.
- Applying one blended cost across variations. If a parent ASIN has variations with different supplier costs, a single blended figure hides which variation is carrying the others.
Maintaining batch-level FIFO across a multi-SKU catalogue by hand is where most sellers give up and default to averages. Analytics platforms including sellerboard support per-batch COGS entry with FIFO assignment, so the cost attached to each sale reflects the batch it came from rather than a catalogue-wide average.
Frequently asked questions
Does Amazon calculate COGS for me?
No. Amazon knows your fees and your revenue; it has no visibility into what you paid your supplier, your freight forwarder, or customs. Every COGS figure in any dashboard comes from data you entered.
Can I use different methods for tax and management reporting?
Generally yes, and many sellers do — a simplified treatment on the return and full accrual COGS internally. Keep clean reconciliation between the two. LIFO is the exception: electing it for tax carries a conformity requirement.
Which method gives the highest reported profit?
When costs are rising, FIFO produces the lowest COGS and therefore the highest reported profit and the highest tax bill. When costs are falling, the relationship reverses. Choosing a method to maximise a headline number is the wrong basis for the decision.
What about inventory I have already sent to Amazon but not sold?
It remains your asset and stays on the balance sheet at cost until sold, regardless of physical location. FBA units, AWD units, 3PL units, and stock in transit all belong in ending inventory.
How do I handle unsellable returns under FIFO?
A unit that comes back unsellable is written off at its assigned batch cost. It is not a COGS entry against a sale — the sale was already refunded — so it belongs in inventory shrinkage. Burying write-offs inside COGS makes your unit economics look worse than they are and hides the returns problem.
