Ready to take your Amazon Vendor data ops to the next level?
Request access today and let InsightLeap help grow your business!
September 7, 2026 by InsightLeap
An ASIN reads unavailable on its detail page while pallets of it sit in your own warehouse, because Amazon's last purchase order for it came in short and nobody looked hard at the confirmation before it went back. Three weeks later a deduction lands against the shipment that did go out. Both problems are already downstream of your warehouse by the time you see them, which is why counting units on your own floor does nothing for either one.
That is a consequence of how the 1P model works. Amazon owns the stock, generates the orders, and holds the units in its own fulfillment centers, which leaves you responsible for the quality of what you feed into that process and the accuracy of what you send back. The rest of this walks the cycle in the order you live it: how a PO gets generated, what confirming and shipping it commits you to, which chargebacks attach to the freight calendar rather than the warehouse, which fields show you any of it, and how this week's behavior lands in next week's forecast.
Worth clearing up first, because most of what gets written about Amazon inventory management describes a different selling model. The Inventory Performance Index scores how well a third-party seller manages stock they own inside FBA, so it leans on sell-through and punishes excess and aged units. As a 1P vendor you have no IPI score, because you are not the party deciding what sits in a fulfillment center; Amazon is. If you run a Seller Central account alongside your vendor account, our piece on improving an IPI score covers that side of the house. Everything below is the vendor side.
Amazon generates purchase orders against its own forecast, assembled from sales history, current inventory levels, and demand signals on the ASIN. You can read the output of that model directly: the forecasting data returns weekly customer demand forecasts at ASIN level in four flavors, a mean forecast plus P70, P80, and P90.
What you can influence is narrower than most vendors assume:
What you cannot do is edit the forecast. Inventory planning on the vendor side means combining Amazon's data with your own sales and supply-chain data and then acting on the difference early, which is why telling your vendor manager about a demand event before that period's POs generate is worth more than arguing about an order quantity after one lands.
The operating cycle runs in four steps: the PO arrives, you confirm it, you prepare and ship it, Amazon receives it. Most accounts move that traffic over EDI, where the order, the acknowledgment, and the shipment notice are exchanged as documents instead of screens. Our EDI breakdown covers what that integration commits you to and what each late document costs.
Each of the steps you own carries its own penalty for getting it wrong.
Confirmation. Late acceptance of a PO draws a chargeback on its own. What you accept, backorder, or reject also lands permanently in your fill rate. "Fill Rate - Accepted Rate" is the percentage of items confirmed as Accepted during PO confirmation, and "Fill Rate - Backordered Rate" is the percentage of quantity confirmed as Back Ordered; the same two values show up on the Operational Metrics report as "Accepted confirmation rate" and "Backorder confirmation rate". A team that lets POs sit and then confirms them as backorders is telling Amazon's next forecast that the demand it predicted cannot be filled.
Shipment. Ship more units than the PO authorized and Overage PO Units applies; ship fewer than you confirmed and the incomplete shipment draws its own chargeback. Since the number you confirmed is the commitment, the work that prevents both of those happens at confirmation time, well before anything reaches the dock.
The ASN. The advance ship notice tells Amazon what is arriving and in which cartons, before it arrives. ASN Accuracy and Carton Content Accuracy chargebacks exist for precisely this step: inaccurate shipment detail, or carton quantities that do not match what shows up. The ASN is also the document your shortage disputes will rest on months later, so it deserves a check against what physically shipped before anyone files it.
Your standing across all three reads in the Operational Performance dashboard, which tracks ODR, POP, and OTD.
Three further chargeback categories have very little to do with what you shipped and everything to do with when, and all three are measured against the shipping window Amazon assigned:
All three are decided on a calendar, well before anything is picked or packed. The booking deadline is the earliest date on that calendar, which makes it the one to put a standing reminder against: an ocean shipment booked inside fourteen days is late before a single container is packed. Vendors dispute chargebacks through the Operational Performance section of Vendor Central, and a dispute you win still cost you the hours it took to file. Navigating the vendor experience goes further into how chargebacks and deductions behave once they are on the account.
Inside Vendor Central, the inventory view lives under Reports, then Retail Analytics, then Inventory. Five fields carry most of the operational signal:
| Field | What it counts |
|---|---|
| Sellable On-Hand Units | Units in Amazon FCs still in sellable condition |
| Unsellable On-Hand Units | Units in Amazon FCs in unsellable condition |
| Open Purchase Order Quantity | Confirmed quantities pending receipt |
| Aged 90+ Days Sellable Units | Sellable units at least 90 days old |
| Procurable Product Out-of-Stock Rate | How often the product was unavailable, against total glance views |
Sellable on-hand read alongside open PO quantity is the pair worth checking first, because a thin sellable position sitting next to a large open PO quantity points at a receiving or transit bottleneck instead of an ordering problem, and those two get fixed in different places.
Two things about this screen are easy to get wrong. The first is timing: refreshed inventory data lands 72 hours after the close of the period, so a weekly dataset running Sunday through Saturday is readable by the end of Tuesday. The inventory view always describes last week, which is why PO confirmation discipline, where you are acting in real time, moves your stock position more than watching this screen does.
The second is provenance. Most material on the open web about "Amazon inventory reports" is describing Seller Central's Amazon Fulfilled Inventory Report, which breaks stock into sellable, unsellable, reserved, and inbound units with warehouse location, condition, and age. Those field names read close enough to the vendor ones to get copied into a vendor SOP by mistake, and they belong to the other selling model.
Two more places are worth a standing look. The Out of Stock Items metric lists out-of-stock items carrying a replenishment code of Planned Replenishment (PR) or New Product (NP), drawn from the Vendor Catalog Listing report, which makes it the fastest read on items that should be in stock and are not. Net Received, on the Inventory Health report, is the net amount Amazon received after subtracting what was returned to you. If you want the full field-by-field on which report holds what, every Vendor Central report, defined covers all of them.
Shortage deductions arise from a discrepancy between what you shipped and what Amazon's receiving process recorded. That gap is a receiving-side event, so your case is built entirely from documents you generated at ship time, and disputes turn on two of them: the accuracy of the ASN and the package weight evidence.
Shortage disputes are filed through Vendor Central's Dispute Management section. The supporting evidence that carries weight is specific:
Open the claim while those records are still easy to retrieve, because a shortage claim that sits gets harder to support as the carrier records age out and the people who packed the shipment stop remembering it. Shortages are one line in the reconciliation work that also covers payments, deductions, and invoices, and it is the line where waiting costs you the most, since the supporting evidence is the first thing to go.
Your PO confirmation behavior and shipment accuracy this cycle become inputs to Amazon's forecast next cycle. Backorder a confirmed quantity and the forecast learns that the demand it predicted cannot be filled at that level. Ship short and it is the receiving record, not your invoice, that becomes the history Amazon plans against. The chargeback shows up on the account where you can at least argue with it, while the forecast adjusts down without telling you.
Auditing that loop takes one habit, because Amazon does not keep the evidence for you: only the most recent forecast data is available, and historical forecast data cannot be requested. Refreshed forecasting data lands within 72 hours of week end, so if nobody on your side saves that weekly file when it publishes, forecast-versus-actual analysis is gone, along with any read on whether your confirmation discipline moved the forecast at all. A weekly export into a folder someone owns takes a few minutes and is the only version of that history you will ever have.
The fields above sit in different corners of Vendor Central on different refresh cycles, and lining them up week over week is manual work. That reconciliation is what InsightLeap automates: the fill rate and chargeback states, the inventory position, and the out-of-stock items with their replenishment codes arrive together on a fixed schedule instead of one pull at a time. Weeks of Coverage, which estimates how many weeks it will take Amazon to sell through its current stock, is calculated in InsightLeap and has no equivalent field in Vendor Central. Every metric mentioned here is spelled out in the metric descriptions glossary, including which Vendor Central report each one comes from.
Request access today and let InsightLeap help grow your business!