You connect once. We do the rest.
No software to learn, no dashboard you have to remember to check, no claims for you to write. Ten minutes of setup, then a monthly report.
Connect — 10 minutes, two pieces
Data access. You authorize FeeBulldog through Amazon's official Selling Partner API with read-only permissions. Standard Seller Central flow: you click authorize, we never see your password, and you can revoke it yourself at any time.
A scoped user invitation. Amazon provides no API for filing a claim — claims are submitted as cases inside Seller Central. So you also send us a normal Seller Central user invitation, limited to case management and reporting permissions.
We would rather tell you this on a website than surprise you on a call. Any service that claims it recovers money through the API alone is either not filing claims, or not being straight with you about how it works. The full explanation is below.
Audit — the first sweep
We pull everything Amazon will give us: fee preview and fee reports, inventory ledger and adjustments, returns and reimbursements, removal orders, storage and surcharge billing, and settlement reports.
Then we reconcile it. Every unit that entered a fulfillment center should have left as a sale, a return, a removal, or a documented write-off. Every fee charged should match the dimensions and weight on file. Where the numbers do not reconcile, there is a claim.
Our first sweep covers everything still inside the claim window. Anything older is unrecoverable — for us and for anyone else — which is exactly why the sooner you start, the more there is.
File — evidence, not volume
A claim without documentation gets valued by Amazon's internal estimate, which is consistently lower than what you are actually owed. So we build the package: supplier invoices showing per-unit cost, corrected dimension and weight evidence, order-level and shipment-level detail.
We file inside the deadline, we file claims we can substantiate, and we track every one through to resolution. When a claim is wrongly denied, we appeal it with better evidence rather than resubmitting the same thing.
We do not carpet-bomb Seller Support with weak claims. That gets accounts flagged, and it is not how you get paid.
Watch — and fix the cause
The 60-day window turned recovery from an annual cleanup into a monitoring problem. A monthly check already loses money. We re-run the reconciliation daily, and new discrepancies become claims immediately.
When we find a fee error, we do not stop at the refund. We push to get the dimensions and weight corrected on file, so the overcharge stops going forward. On a fast-moving SKU that forward saving routinely dwarfs the refund itself.
Once a month you get a plain report: what we found, what we filed, what Amazon paid, what we fixed, and what we are still chasing.
The cost documentation vault
The least glamorous part of this service, and probably the highest-return.
Since March 2025, Amazon reimburses lost and damaged inventory at your manufacturing cost. If you cannot document that cost, Amazon applies its own internal estimate — and that estimate is consistently below what sellers actually paid.
So the difference between a well-documented claim and a fast one is not five percent. On a catalog with real COGS behind it, it is routinely the difference between being made whole and being made half-whole. Every time. On every unit.
What we collect
- Supplier invoices showing verifiable per-unit cost
- Purchase orders and payment records that corroborate them
- Bills of materials where you manufacture rather than source
- Updated costs whenever a production run changes them
How we hold it
- Collected during onboarding, before the first claim goes out
- Stored encrypted, isolated to your account, never pooled
- Attached automatically to every claim that needs it
- Yours to export or delete whenever you want
The trap this avoids: a service that files immediately looks responsive. But a claim submitted without cost documentation gets valued low and settled, and a settled claim is difficult to reopen. Speed without evidence is not service, it is just a faster way to be underpaid.
How a claim actually gets filed
Reimbursement is not an API transaction. It is a support case, read by a human or a model at Amazon, that either persuades or does not.
Why there is no "file claim" button
Amazon's Selling Partner API is generous with data. You can pull settlements, fee events, the inventory ledger, adjustments, returns, removals and storage billing. That is everything needed to find what you are owed.
What it does not offer is any endpoint that submits a reimbursement claim. Every claim in FBA is raised as a case in Seller Central's case log — the same system you use when something goes wrong with a shipment. There is no way around this, for us or for anyone else.
What a case looks like
Opened
We open a case against a specific issue — a shipment, an order ID, a FNSKU, a fee. Vague cases get closed. A case that says "please check my reimbursements" is worth nothing; a case that says "shipment FBA15X9J2 shows 500 units shipped, 480 received, no reconciliation, see attached BOL" is worth money.
First response
Usually fast and often automated. A meaningful share of legitimate claims get an initial denial or a request for information that has already been provided. That is normal, and it is where most sellers give up.
Reply with evidence
Supplier invoices proving per-unit cost, shipment documentation, order-level detail, corrected dimensions. Documentation is the entire game. Without cost documentation on file, Amazon values inventory claims using its own internal estimate, which is consistently lower than your actual cost.
Escalate, or accept
Where a denial is wrong, we reply with the specific policy and the specific numbers rather than resubmitting the same case. Where a denial is correct, we close it. Re-filing dead claims wastes the one resource that actually matters here: Amazon's willingness to take your account's cases seriously.
Paid
Approved reimbursements post directly to your account as a credit. The money never touches us. You see it in your own settlement report, and we invoice afterwards for our 10%.
Why we file fewer cases than you would expect
Volume is not the strategy. Amazon tracks case quality per account, and an account that floods support with weak, undocumented claims gets slower responses and more denials on the claims that are good. Some services optimize for cases filed. We optimize for dollars recovered per case, because the two work against each other.
You can watch all of it. Every case we open appears in your own Seller Central case log, in your account, with the full correspondence. There is no hidden dashboard where the real activity happens. If you want to read every word we send Amazon on your behalf, it is already there.
What we can and cannot do in your account
You are handing a stranger access to your business. Here is exactly what that means.
What we have
- Read-only API access, via Amazon's official authorization flow
- A Seller Central user invitation scoped to case management and reporting
- The ability to open, answer and escalate cases on your behalf
What we do not have
- Your password — we never request or store it
- Any ability to edit listings, prices or advertising
- Any access to your payouts or bank details
- Permission to add or remove other users
- Your data mixed with any other seller's, ever
You revoke both the API authorization and the user invitation yourself, from your own account, in about thirty seconds, without asking us. Read the privacy policy.
Find out what Amazon owes you.
A free audit takes ten minutes to set up. You get a findings report with real dollar figures, and no obligation to do anything with it.