Building an Amazon vendor P&L that tells the truth
Most vendor P&Ls are wrong in the same direction: they start from purchase order value and end in optimism. A vendor P&L that tells the truth starts from what actually landed in the bank and works down to contribution per item — which is the only number that can support a pricing, assortment or negotiation decision.
Why PO value is the wrong starting point
Purchase order value is an intention. Between that number and your bank sit confirmation shortfalls, receiving discrepancies, agreed allowances, compliance chargebacks and unrecovered claims. Every one of them reduces the figure, and none of them appear in the PO.
Vendors who report on ordered revenue therefore run a business on a number that is systematically too high — and, worse, too high by an amount that varies by item. That variance is what makes assortment decisions go wrong: the items with the worst deduction profile look identical to the best ones in a PO-based view.
The fix is structural, not analytical: start from net receipts and build downward.
The line structure that works
A workable vendor P&L runs in this order, item by item:
- Net receipts — what was actually paid against invoices for that item
- Less allowances — the CoOp stack attributable to it
- Less compliance deductions — chargebacks traced to its shipments
- Less unrecovered claims — shortage and pricing deductions never won back
- Less cost of goods — production or purchase cost, landed
- Less your logistics — inbound freight, handling, packaging and preparation
- = Contribution per item — the number decisions are made on
Below contribution sit your fixed costs — people, systems, overhead. Those are managed at business level, not attributed item by item, because false precision there produces worse decisions than honest aggregation.
| Line | Source | Common error |
|---|---|---|
| Net receipts | Remittance advice, not your invoices | Starting from PO or invoiced value |
| Allowances | Vendor agreement and remittance detail | Treating them as a footnote instead of a line |
| Chargebacks | Deduction log by shipment | Leaving them in overhead, unattributed |
| Unrecovered claims | Claims log after dispute outcomes | Assuming disputed equals recovered |
| Cost of goods | Landed production or purchase cost | Using ex-works cost and ignoring inbound |
| Your logistics | Freight, handling, packaging, preparation | Attributing freight per unit rather than by cube |
The attribution problems, and honest answers
Three lines resist clean attribution, and pretending otherwise is how P&Ls lose credibility.
Allowances usually accrue on total revenue rather than per item. Attribute proportionally to revenue and label the method — the goal is comparability between items, not accounting perfection.
Chargebacks attach to shipments, which often contain many items. Attribute by the item that caused the failure where the record allows, and proportionally where it does not.
Freight should be attributed by cube or weight, not by unit count. Attributing freight per unit is the single most common way bulky low-price items are made to look profitable — precisely the items most at risk of being flagged as uneconomic.
Turning it into decisions
Rank items by total contribution — per-unit contribution times volume — rather than by margin percentage. A high-percentage item selling nothing funds nothing; a moderate-percentage item at volume may carry the business.
Then read the ranking in three bands. The top is what you protect: supply reliability, availability and content quality here matter more than anywhere else. The middle is where improvement pays: packaging, returns, deduction reduction. The bottom is a triage list: fix, restructure, move to another model, or retire.
That ranking is also your negotiation evidence base. Walking into annual negotiations with item-level contribution after deductions is a categorically different conversation from walking in with a cost-price request.
Keeping it honest over time
Rebuild the numbers on a regular cycle, ideally monthly, and always after any terms change. Deduction rates move, allowances change at renewal, freight costs drift and product costs shift — a P&L built once is a historical document within a quarter.
Reconcile it against remittances rather than against your own invoices, since the whole point is to measure what landed. And keep the method stable: comparability across periods is worth more than a slightly better model introduced halfway through the year.
Note: this article describes a reporting structure, not a target or a threshold. Which items to act on, at what level, and with what alerting logic is a commercial decision that belongs inside your own operating setup — not in a general guide.
FAQ
Why not build the P&L from purchase order value?
Because PO value is an intention, not a receipt. Confirmation shortfalls, allowances, chargebacks and unrecovered claims all reduce it, and by different amounts per item — which makes assortment comparisons unreliable.
What lines belong in a vendor P&L?
Net receipts, less allowances, less compliance deductions, less unrecovered claims, less cost of goods, less your own logistics, giving contribution per item. Fixed costs sit below that and are managed at business level.
How should I attribute allowances to items?
Proportionally to revenue in most cases, since allowances typically accrue on totals. Label the method and apply it consistently — comparability between items matters more than accounting perfection.
How should freight be attributed?
By cube or weight rather than by unit count. Per-unit freight attribution systematically flatters bulky, low-price items, which are exactly the products most at risk of being flagged as uneconomic.
Should I rank items by margin percentage?
No. Rank by total contribution, which is per-unit contribution times volume. A high-percentage item with negligible volume funds far less than a moderate-percentage item at scale.
How often should the P&L be rebuilt?
Monthly as a rhythm, and always after a terms change. Deduction rates, allowances, freight and product costs all drift, so a P&L built once becomes historical within a quarter.
What should I do with the weakest items?
Triage rather than delete: fix the physical or content driver, restructure the offer, move the item to another model, or retire it. The choice depends on whether the underlying problem is fixable at source.
Use it where it counts
Item-level contribution after deductions is the strongest evidence you can bring to annual terms discussions.