The problem
Seller Central's sales and traffic report by child item has a row for every product, with sessions, conversion and Buy Box share, but going through it line by line is the job that gets skipped. For a seller with 60 listings, a proper monthly look takes about 3 hours in a spreadsheet, and a listing with plenty of visitors but few sales goes unnoticed.
What the AI does
ChatGPT's data analysis reads an uploaded CSV or Excel file, runs the calculations in Python and returns tables and charts you can question. Download last month's report by child item, upload it, and ask which products have above-average sessions but below-average unit session percentage, and which lost Buy Box share. You get a short list to work on and can ask follow-ups in plain English. Full data analysis is on Go (about $9£7€8A$13C$13 a month) and above; Free is limited.
A worked example
A seller in Stockport with 60 kitchenware listings on Amazon.co.uk downloads the child-item report on the 1st of each month and uploads it to a ChatGPT project that holds the same set of questions.
Before
About 3 hours a month in a spreadsheet, often skipped
After
About 45 minutes a month, with one weak listing picked out to fix
What it's worth
About $90£68€81A$131C$128 to $250£188€224A$361C$353 a month, against $9£7€8A$13C$13 to $27£20€24A$38C$38 for ChatGPT Go or Plus. The time saved is 2.25 hours a month, about $90£68€81A$131C$128 at an assumed $40£30€36A$58C$56 an hour; if fixing one listing lifts its conversion from 8% to 10% on 1,200 sessions a month at $27£20€24A$38C$38 a sale, that's 24 more sales ($638£480€571A$922C$902), about $160£120€143A$230C$226 profit at an assumed 25% margin.
How to set it up
- In Seller Central, open Reports, then Business Reports, and under By ASIN choose the detail page sales and traffic report by child item. Set last month's dates and download the CSV.
- Use ChatGPT Go (€8, about $9£7€8A$13C$13 a month) or Plus (€23, about $27£20€24A$38C$38 a month); Free's data analysis is limited.
- Create a project called 'Monthly Amazon review' and save your standard questions in its instructions.
- Upload the CSV and ask: which ASINs have above-average sessions and below-average unit session percentage, which lost Buy Box share, and a chart of the top 10 by sales.
- Ask it to list the columns and figures behind each conclusion, and check two of them against Seller Central.
- Pick 1 or 2 listings to work on: images, price, title or reviews.
What goes wrong
- Report layouts change and it can pick the wrong column. Ask it to name every column it used.
- It only knows the file. Stock-outs, ad spend changes or a rival's price cut can explain a dip and aren't in the report.
- Free data analysis is limited and may not finish the job. Go handles it; Plus adds the Excel and Google Sheets extensions, which this job doesn't need.
- Go and Plus are personal plans: turn off 'Improve the model for everyone' in Data Controls if you don't want sales data used for training. The Business plan (about $24£18€21A$35C$34 per user a month billed annually, from 2 users) doesn't train on business data by default.