Doing Amazon Reimbursements Yourself vs Hiring a Service

What DIY actually requires, week to week

Filing your own claims isn't complicated in the sense of being technically hard — it's a report-reconciliation habit. You need to pull your inventory adjustment report, your FBA returns report, and your removal shipment status regularly, cross-reference them against what's already been reimbursed, and file cases against genuine gaps inside their specific windows. The catch is the windows: fulfillment-center losses and removal shortfalls run 60 days from the event, removals lost in transit run 15 to 75 days from shipment, and customer returns that never came back can't even be filed until 60 days after the refund, closing at 120. A quarterly review misses claims in at least one of those categories almost by definition, because the eligible window opens and closes between checks.

The real cost of DIY isn't the filing itself — it's the discipline of doing this every week, indefinitely, on top of everything else running a catalog requires. That's a genuinely different ask than “know how to fill out a form.”

What a service actually buys you

A reimbursement service's core value isn't superior knowledge of the policy — the policy is public, and a diligent seller can read the same pages we do. It's the standing cadence: someone (or some system) running the reconciliation every week regardless of whether anyone remembered to, and building the evidence trail — sourcing-cost invoices, shipment IDs, removal confirmations — before a claim deadline forces it. For a catalog where the founder or ops lead is already stretched across sourcing, advertising, and customer service, that cadence is the actual product, more than the filing mechanics.

What a service doesn't automatically buy you is a higher approval rate on any individual claim than a careful DIY filer would get — the evidence standard is the same either way. The advantage is coverage and consistency at scale, not a secret trick for winning individual cases. There's also a second, less obvious thing a good service should be buying: restraint. Amazon's own policy treats bulk, premature, or under-researched filing as grounds for account monitoring, and a provider paid on volume has an incentive problem here that a careful in-house filer, who has no reason to file more than the evidence supports, simply doesn't.

A worked breakeven, with real numbers

Say a catalog runs $6,500 in recoverable reimbursements in a fortnight when checked properly — a real figure from one account we track, running the check weekly. At a 25% contingency rate, that fortnight costs $1,625 in fees; the seller nets $4,875 for effectively no time spent. Doing it yourself at the same recovery rate costs nothing in fees but requires the hours to build the reconciliation and file the cases — call it four to six hours for a catalog that size, done properly. If your time is worth less than roughly $270–$400 an hour to you, the arithmetic favors doing it yourself, assuming you'll actually do it every single week without fail. If your time is worth more than that, or if the weekly habit is realistically going to slip after the third month, the math flips toward a service — not because DIY doesn't work, but because a process that runs three weeks out of four recovers a fraction of one that runs all four.

The scale where the calculation changes

Below a few hundred SKUs and modest loss volume, DIY is genuinely the more efficient choice for most sellers — the reconciliation is small enough to run as a Friday-morning habit, and a 25% fee on a small recovery base isn't worth the convenience. As catalog size and claim volume grow, the weekly reconciliation stops being a Friday habit and starts being close to a part-time job on its own, at which point the fee on a service starts buying real hours back rather than convenience. There's no universal SKU count where this flips — it depends on loss rate per SKU, which varies by category and packaging — but the honest signal is whether the reconciliation is still getting done every week, on time, three months in. If it isn't, the DIY math was theoretical, not real.

What to do when DIY reimbursements have already lapsed

If a self-run process has already gone quiet — nobody's checked in six weeks, or longer — the first move isn't panic, it's triage: pull the last three months of refund, removal, and inventory-adjustment reports and check what's still inside its window before anything else. Customer-return claims from more than four months ago are likely already outside the 120-day close and genuinely gone; fulfillment-center losses from the last two months may still be filable. Recovering from a lapsed process is mostly about accepting the losses that are truly gone and refocusing effort on the window that's still open, rather than trying to reconstruct everything retroactively.

This is also the moment most sellers actually decide between continuing DIY and hiring a service — not at the theoretical breakeven calculated in a spreadsheet, but at the point a lapse has already cost real, unrecoverable money. That's a legitimate way to make the decision. It's worth building the habit either way of asking, at the moment of lapse, whether the thing that broke was time, attention, or genuine complexity — the fix is different for each, and only the first two are solved by paying someone else to do it.

Neither answer is universally right

A seller who genuinely runs the weekly check, keeps sourcing invoices organized, and has the time to spare is well served doing this in-house — there's no fee to pay and no case-craft advantage a service adds on top of a disciplined internal process. A seller whose time is better spent on sourcing or advertising, or whose reconciliation has already slipped once, gets more from paying a percentage for continuity than from a free process that runs inconsistently. Full Circle has managed more than $500M in Amazon spend across 100+ brands, and Dr. Shield exists for the second case — priced on the call, as a contingency, with no fee if nothing comes back — running the exact weekly cadence described above as a standing practice rather than a resolution that fades by month three.