Export your current products first, then build a small ten-product CSV that mirrors the real file's columns and encoding. Import it into a development or duplicate store, or into a clearly named test set, and compare the result field by field against the export. Only after the ten products match should you run the full catalog update.
Why ten products is the right test size
A ten-product test is large enough to expose column mismatches, encoding damage and variant problems, but small enough that you can inspect every field by hand. It also keeps the destructive part of an import contained. Shopify's own guidance notes that when you import a CSV file, you can select the option Overwrite products with matching handles, and that if a non-required column in the import CSV file is blank, then the matching value in the product list is overwritten as blank (Shopify Help: Using CSV files to import and export products). That overwrite behaviour is exactly why a rehearsal matters.
Use ten products that represent the awkward cases in your catalog: one simple product, one with two variants, one with a long title, one with accented characters, one with a blank optional field, and one whose image already exists in your store.
Step 1: Export before you change anything
Export your current products from Shopify and keep that file untouched. This is your rollback reference and your field-by-field comparison source. Do not edit the export in place; copy it and work on the copy.
If your store uses multiple locations and you need inventory quantities, Shopify's documentation points to the separate inventory CSV rather than the product CSV (Shopify Help). Mixing those two jobs in one file is a common source of confusion.
Step 2: Build the ten-product test file
Download the sample product CSV from Shopify and use it as a column template, but delete the example products before adding your own. The sample contains two example products with variants, and Shopify warns that you should remove them if you build your own file from it (Shopify Help).
Keep the column headers exactly as exported. Do not rename, reorder or translate them. If your real import will include metafield columns, include them here too: Shopify supports product metafields in bulk import and export, with headers in the format <name> (product.metafields.<namespace>.<key>) (Shopify Help).
Encoding: save as UTF-8
Accented characters, currency symbols and typographic apostrophes are the first things to break. Save the file as UTF-8 and open it again in a plain text editor to confirm that characters like é, ü or £ still look correct. If they appear as é or similar, the encoding was changed somewhere in the chain.
Blank fields versus absent columns
These behave differently and the difference matters. If a non-required column in the import CSV file is blank, then the matching value in the product list is overwritten as blank. A missing column means the import has no instruction for that field at all. Before importing, list every column in your test file and mark it as present-and-filled, present-and-blank, or absent. That list becomes your checklist after the import.
URL handle and Title
When you import a CSV file, you can select the option Overwrite products with matching handles. If the handle does not match an existing product exactly, the import may create a new product instead of updating the intended one. Test this deliberately: include one row with a handle you know exists and one with a handle that does not.
Step 3: Run the import somewhere safe
If you have a development store or a duplicate store, import there first. If you do not, import the ten products into your live store but give them a clearly identifiable test handle prefix, and record exactly which products you touched. Never run the full catalog file as your first import.
Example (hypothetical)
Imagine a UK store selling ceramics. The test file contains ten products: six simple mugs, two mugs with a colour variant, one bowl with a blank Compare-at price, and one item whose title contains "Café". After import, the store checks whether the accented title survived, whether the blank Compare-at price stayed blank rather than becoming zero, and whether the two variant products kept both variants. These are illustrative checks, not measured results.
Step 4: Compare the result against the export
Open the exported file and the imported test products side by side. Check, in this order:
- Product count: did ten rows create or update exactly ten products?
- Titles and handles: any duplicates or unexpected new products?
- Prices and Compare-at prices: did blanks stay blank?
- Variants: correct number, correct option names, correct order?
- Images: do they load, and are the URLs still valid?
- Metafields: do custom fields appear where expected?
- Encoding: any broken characters anywhere?
Shopify cautions that if you removed products or images from your store after exporting, and then import those products without updating their image URLs, the import will fail (Shopify Help). Test at least one image URL that still resolves.
Step 5: Decide, then scale
If all ten products match, run the full import. If any field is wrong, fix the test file and repeat the ten-product cycle. Do not scale a file that has failed once.
Keep the pre-import export until you have confirmed the full catalog is correct. If something goes wrong at scale, that export is your reference for restoring values.
Common mistakes
- Editing the original export instead of a copy.
- Renaming column headers to "make them clearer".
- Assuming a blank cell deletes a value when it may leave it unchanged.
- Testing with products that are all simple, then importing variants at scale.
- Forgetting that a mismatched URL handle can create a duplicate product.
- Using the product CSV for inventory quantities when a separate inventory CSV is required for multi-location stores.
A reusable worksheet
For each of the ten test products, record: handle, title, price, compare-at price, variant count, image status, metafield status, encoding status, and pass or fail. Nine passes and one fail means the file is not ready.
Follow-up questions
Can I test a Shopify CSV import without a development store?
Yes, but contain it. Import ten products with clearly identifiable test handles, record exactly what you changed, and compare against your pre-import export. The risk is lower than a full import but not zero, so keep the export until you are satisfied.
What should I do if accented characters break during import?
Re-save the file as UTF-8, reopen it in a plain text editor to confirm the characters look correct, and re-run the ten-product test. If they still break, check whether your spreadsheet software changed the encoding on save.
