Reconcile the Amazon deposit.
Explain profit by SKU.
Export a cleaner journal.

SkuSum sums your actual settlement per SKU (not an estimate) and flags any rows where Orders and Settlement don't match. You keep the raw inputs, so every number traces back to a report you can open and check yourself. The parser is validated against 161,344 real settlement lines from US, UK, German and Indian seller files, every line classified, and anything it cannot classify in your file is flagged for review instead of guessed. The full list of line types and where each one belongs is on the fee types glossary.

They stay on your machine. SkuSum runs inside your own browser, so your reports are never uploaded and nothing reaches our servers. We don't even read buyer personal data.

No. There is no AI in the product: no model, no LLM, no call to any AI service. SkuSum is a deterministic parser. It applies fixed mapping rules, does arithmetic, and the same file always produces the same journal. Every number traces back to a settlement line you can open and check, and anything the rules cannot classify is flagged for your review instead of guessed. A bookkeeper cannot trust a number a model guessed, so we kept models out on purpose.

I use AI models every day and my own operation runs on them, so this is not an anti-AI position. It is about what a model can promise. Ask one to classify the same settlement file twice and you can get two different answers, and on a fee code it has not seen before it will help you by guessing instead of stopping. The journal balances either way, which is exactly why a reviewer will not catch it. Books need the opposite property: same input, same output, every line traceable, and a hard stop on anything unrecognized. That is a parser's job, not a model's.

Dashboards keep your data in their cloud and ask you to trust their number. Accounting connectors like A2X summarize the detail away and charge more as you grow. SkuSum gives you the clean, joined, SKU-level file itself, reconciled to your own reports and yours to verify, analyze, or hand to your bookkeeper.

The SKU Economics report is a solid operating view, and for watching a SKU over the long run it is often enough. But it is built for the dashboard, not your books. It doesn't tie back to the settlement payout that actually reached your bank, and it gives you no journal to import into QuickBooks or Xero. SkuSum starts from the settlement, so every number reconciles to the money that arrived, categorizes each fee, refund and reimbursement, and hands you a balanced journal your bookkeeper can review and import. It also flags the rows that don't reconcile instead of averaging over them. Use the SKU Economics report to watch a SKU over time. Use SkuSum when the number has to match your bank and your books.

Yes. SkuSum exports a balanced QuickBooks / Xero journal per settlement (sales, referral and FBA fees, withheld tax, refunds folded into their home accounts, and COGS as proper entries), reviewed with bookkeepers who run Amazon clients, plus a clean profit-by-SKU sheet (CSV or Excel). It is a one-way, import-ready journal file, not a live API sync (by design: nothing connects to your accounts, nothing to break when Amazon changes). See the account mapping, or join below to help shape what comes next.

The core tool is free during validation. Paid licenses: SkuSum Founding Pro, $49 one-time (lifetime Pro, limited to 10 founding seats) and SkuSum Bookkeeper Founding, $199 per year (adds rights for paid client work). Full details on the pricing page. Prices are in USD; taxes may apply and will be calculated at checkout. No per-order pricing: flat, ever.

Amazon only keeps about 90 days of settlement data. SkuSum doesn't store anything for you, but every run gives you a dated file to save. Export regularly and your own saved files become the long history Amazon won't keep. (An optional on-device history is on the roadmap.)