Guide
Five checks before a bulk CSV import
A CSV can import with zero errors and still put the wrong values in the wrong places. Check headers, identifiers, blanks, duplicates, and a small test batch.
“Import successful” is not the same as “data correct”
Bulk imports are dangerous precisely because valid-looking mistakes scale well. A shifted price column or stripped leading zero may be accepted without complaint.
Protect values that only look numeric
SKUs, postal codes, GTINs, and account IDs may contain leading zeros. Treat them as identifiers, not numbers you intend to calculate with.
Blank cells need a meaning
Depending on the platform, a blank can mean leave unchanged, clear the value, or use a default. Test before uploading thousands of rows.
Use a representative 5–10 row test
Include awkward values: Japanese text, blank fields, long strings, prices, inventory, and URLs. Verify the result in the destination before you run the full file.
A 30-second preflight
Check the header, row count, a few rows at the top and bottom, and the columns most likely to be auto-converted: IDs, dates, prices and blanks.
If anything looks odd, stop before the full import. That pause is much cheaper than undoing a bulk update.