1. Establish the destination and category
Decide which marketplace and category the file is for before mapping the product data. A template for one article type may not describe another accurately. Preserve the expected structure rather than changing column headings to make them easier to read. Use the current seller portal to confirm the appropriate template for the upload.
2. Separate facts from generated copy
Brand, product identifiers, size, dimensions and material composition should come from your product records. AI can turn those facts into a readable draft, but missing factual fields still require review. Keep a source sheet or record so the team can trace a questionable value back to its origin.
- Check identifiers as text if leading zeroes matter.
- Keep units explicit and consistent.
- Do not fill unknown factual fields with plausible guesses.
3. Check rows and images together
Count the sellable variations and compare them with the workbook rows. Match every row to the correct product and colour. Amazon and Flipkart exports can carry hosted image links; the EcommerceListing Myntra workflow uses a separate image upload. Assign that handoff explicitly so a completed spreadsheet is not mistaken for a fully submitted catalog.
4. Close the loop after upload
Keep the submitted workbook and processing feedback together. A portal can reject a value even when the spreadsheet is structurally valid. Identify whether the correction belongs to the source product record, a category choice or the export mapping. Make the smallest accurate correction and review the affected rows before resubmitting.