Purchase Order (PO)

A Purchase Order (PO) is the document Amazon issues to a vendor — a first-party seller operating through Vendor Central — committing to buy a specified quantity of a SKU at an agreed cost, with a defined ship-by window. The PO is the operational unit of vendor life: forecasting, production planning, shipping, invoicing and chargeback disputes all hang off it.

For sellers (3P), the analogous concept is "incoming orders," but the dynamic is fundamentally different: 3P sellers ship to fulfil customer orders Amazon already brokered; 1P vendors ship POs Amazon places against the vendor's catalogue based on Amazon's own demand forecast.

PO lifecycle

Amazon forecasts demand for vendor SKUs
  → Amazon issues PO (qty, cost, ship window, FC destination)
  → Vendor confirms (accept / partial / reject)
  → Vendor ships within window
  → Amazon receives & checks in
  → Amazon pays per agreed terms (often Net 60–90)

A typical Amazon PO ships a vendor 2–8 ASINs at 50–5,000 units each, with a 7–14 day ship window to a specified FBA destination.

PO confirmation: the highest-leverage step

When Amazon sends a PO the vendor has three responses per line:

• Accept full — commits to ship the full quantity on time.
• Accept partial — commits to a lower quantity (often used when production is constrained).
• Reject — cannot fulfil at all.

Acceptance rate is one of the metrics Amazon tracks against vendors. A vendor that consistently rejects POs loses forecast weight (Amazon orders less in future cycles) and risks scorecard penalties.

Chargebacks: the cost of getting POs wrong

Every operational miss on a PO triggers a chargeback — a fee Amazon deducts from the vendor's next payment. The big ones:

Chargebacks routinely consume 2–8% of a vendor's gross PO revenue. Disputing them is operational table-stakes for any serious vendor.

PO forecasting

Amazon shares two forecast horizons through Vendor Central / ARA (Analytics):

• 6-week forecast — short-term, near-deterministic; basis for production planning.
• 26-week forecast — long-term, directional only; basis for raw-material commits and capacity planning.

Both are Amazon's view of demand — not the vendor's. Vendors that build production around the 6-week without buffer get caught when Amazon revises forecasts upward late.

Common mistakes

• Treating PO ship windows as soft. Late = chargeback. The window is the contract.
• Accepting full when capacity says partial. A partial-accept is far cheaper than missing the ship window on a full-accept.
• Not disputing chargebacks. A meaningful share (often 40–60%) are disputable with evidence. Vendors that don't dispute leak margin permanently.
• Forecasting from Amazon's 26-week as if it were committed. It is directional; over-committing raw materials to it is how vendors get stuck with excess inventory.