Product Research · Blog post · source May 28, 2026
Per-Brand Ledgers from Amazon Payouts and Ad Charges with Claude Code
Match Amazon payouts and ad charges to the right brand and write them to each brand's ledger
He says the job went from two or three hours a month to a couple of minutes answering questions.
Per source Self-reported by the source, not independently verified.
Copy-paste prompt
Here are my bank and card transactions since [date] [attach], and Amazon's settlement payouts and ad invoices for each brand [attach]. Match each Amazon payout and ad charge to its brand by amount and date. List everything else you can't attribute, with your best guess and why. Output a table per brand (date, description, amount) I can paste into its ledger. Don't include anything already on the ledger [paste last 10 rows].
The source didn't publish its exact prompt. We wrote this starter so you can try the same workflow. Replace anything in [brackets].
What they did
Alex Willen's brands share bank and card accounts, so every Amazon payout and ad charge has to be matched to a brand by hand. His Claude Code skill finds the last entry on each brand's Google Sheets ledger, pulls Mercury transactions and Amazon payments and ad charges by API, opens the Amex page for him to log in and downloads the CSV, matches Amazon items by amount and date, attributes the rest from context notes (for example which manufacturer makes which brand), and asks him only about charges it can't place. It shows the list, writes in date order and then checks that every transaction is accounted for.
How it went
He says the job went from two or three hours a month to a couple of minutes answering questions. He no longer reviews the list before it writes because it has been 100% accurate for a long time.
We only report what the source says. Results are self-reported unless noted.
Worth knowing
He types his Amex login himself each run rather than giving Claude the password. Mercury has an API; most banks and cards only give you a CSV or PDF.