Amazon
Reading the Amazon settlement report for real profit
Every seller has a profit number. Very few have one that came from the settlement report, and the gap between the two is usually larger than people expect.
Why the dashboard number is not enough
Dashboard figures are built for reassurance and speed. They show revenue and a fee estimate. They do not show the adjustments, the reserve movements, the refund reversals or the fees that arrive late.
The settlement report shows all of it, in transaction-level detail, which is why it is harder to read and why it is the only one worth trusting.
The structure to expect
Rows are transactions, not orders. One order can produce a sale row, one or more fee rows, and later a refund row and a fee-reversal row. Aggregating naively by order id without understanding the types will double-count.
- Order payments - the gross sale
- Refunds - negative, often in a later settlement period than the sale
- Fees - referral, closing, and any programme fees, itemised
- Adjustments - reimbursements, corrections, and things you did not expect
- Transfers - the actual money moving to your bank
The period boundary is the main trap
Settlements cover a window, and orders do not respect it. A sale in one period is routinely refunded in the next, which means any single settlement looks either better or worse than reality.
Judging performance on one settlement is like judging a month by one week. Aggregate several and the picture stabilises.
Matching to supplier costs
The report tells you what Amazon did. It knows nothing about what you paid your supplier, and that is the other half of the margin. The join is the order id, and it only works if you recorded the supplier cost against the order at purchase time.
This is the point of all the record-keeping discipline elsewhere in your workflow. It exists so that this one calculation is possible.
What to compute
- Net Amazon proceeds per order, from the settlement rows
- Supplier cost per order, from your own records
- Difference, per order - that is real gross margin
- Roll up by product to see which lines actually earn
- Compare against the provisional margin you were using
The comparison is the payoff
That last step is what makes the exercise worth doing regularly. Once you know that a given product settles at, say, eighty percent of the provisional margin you estimated, the fast number becomes usable again - you just apply the discount.
Without it you are pricing, stocking and scaling on a number that has never been checked against reality.