StoreBuilt checked Shopify’s current product CSV specification and common import-error guidance for this article. The important distinction in that review is between a file that passes import validation and a catalogue change that matches the merchant’s intention. Those are separate checks, and the second requires comparing what changed with what was supposed to stay the same.
A Shopify product CSV update can be useful for a controlled catalogue change, but a spreadsheet should be treated as an instruction to modify live data. UK ecommerce teams need a clear scope, a fresh starting export and a way to verify the result. This guide focuses on updating existing products, rather than moving an entire store or choosing a product-information platform.
Contact StoreBuilt with the fields you need to change and the product structures involved before a complex catalogue update.
Table of contents
- Write a field-level change brief
- Understand matching and blank values
- Protect identity and spreadsheet formatting
- Compare the proposed file with the export
- Run a representative pilot
- Walk through a price-update scenario
- Verify the result beyond the import message
- Prepare a realistic recovery route
- StoreBuilt point of view
Write a field-level change brief
Describe the intended change precisely. Updating selected retail prices is different from replacing product descriptions, changing options or rebuilding image associations. If the brief contains several unrelated operations, separate them where practical so the result of each can be checked independently.
List the products in scope, the fields allowed to change and the fields that must remain stable. Assign an owner who understands the merchandising decision, not just the spreadsheet format. Someone needs to resolve whether an unexpected difference is a mistake or an intentional commercial change.
Take a fresh export close to the planned update and preserve an untouched copy. If colleagues continue editing products during preparation, the file may become stale before import. Agree a short coordination window or compare against a newer export before proceeding, especially where another system also updates catalogue information.
Understand matching and blank values
Shopify’s product CSV specification explains matching-handle overwrites and the difference between omitted columns and included blank values. Read the requirements for the fields involved rather than assuming that an empty cell means leave unchanged.
Do not assume that a supplier SKU or barcode is the native product overwrite key. Map the supplier’s data to the correct existing products before constructing the import file. A plausible product name is not enough when several sizes, finishes or pack formats share similar descriptions.
| File situation | Risk to consider | Review action |
|---|---|---|
| Included field left blank | existing value may be cleared | identify every new blank |
| Handle unexpectedly changed | matching intention may fail | compare product identity |
| Option structure altered | variants may change unexpectedly | separate structural edits |
| Supplier file used directly | columns may not match native format | map to current specification |
| Old export reimported | recent changes may be replaced | compare with fresh data |
Omitting a column is not a universal shortcut. Required columns and relationships between fields still matter. Keep the current specification beside the preparation process and test the actual file you intend to use, not a simplified example with a different structure.
Protect identity and spreadsheet formatting
Spreadsheet software can transform identifiers in ways that are difficult to notice. Leading zeros, long numeric strings and text resembling dates deserve explicit checks. Preserve product identifiers as text where appropriate and compare the saved file after export, not only the values displayed in the editing application.
Check punctuation, line breaks and encoding in descriptions. A table view can make a multiline description appear to be several records even when the CSV is structurally valid, or hide broken quoting that shifts values between columns. Use a CSV-aware inspection method rather than counting commas manually.
Keep formulas out of the final data unless their evaluated values are deliberately exported and verified. A calculation working in one workbook is not evidence that the resulting CSV contains the intended final price. Review a sample of saved values independently from the formula cells that produced them.
Compare the proposed file with the export
Produce a difference report before importing. For each product in scope, identify the changed fields and flag unexpected blanks, new identifiers or removed rows. The merchandising owner should be able to read the report without inspecting hundreds of untouched columns.
Distinguish a row count from a product count. A catalogue with variants and images can use multiple rows for one product, so matching total rows does not prove matching product coverage. Compare the relevant product and option structure, including a few records with deliberately different complexity.
If the change brief says prices only, an altered description or image reference should stop the review until explained. This rule is more useful than asking someone to visually scan the entire spreadsheet and confirm that it looks fine. It gives the approver a specific condition to evaluate.
Run a representative pilot
Choose a small group that represents the structures being changed. Include a straightforward product and, where relevant, one with several option combinations. A pilot containing only the easiest record can pass while leaving the most important risk untested.
Follow the documented import workflow and inspect its preview and messages. Shopify’s common import issues guidance can help interpret validation failures. Do not repeatedly alter unrelated columns to silence an error without understanding the underlying requirement.
After the pilot, review the product in the admin and on the public storefront. Confirm the intended values and several unchanged controls. For a price change, that includes the selected option and cart line, not merely the default product-card price. Keep the evidence with the exact pilot file.
Walk through a price-update scenario
Consider an illustrative homeware retailer updating prices for a selected range of lamps. The spreadsheet contains both the intended price column and an accidentally emptied vendor column. The file may still look plausible because the review focuses on the new prices.
A field-level difference report reveals the additional blanks before import. The team restores the unintended values, checks the documented requirements and pilots a single-format lamp plus a lamp with several finishes. The storefront review compares each selected finish with its cart price.
The lesson is operational rather than numerical: define the change boundary and verify the saved file against it. This example does not describe an actual client incident or claim a measured saving. It illustrates why import success alone is not an adequate acceptance criterion.
Verify the result beyond the import message
After the main import, compare a new export with the approved change brief. Check that all intended products changed and that products outside scope remained stable. Investigate partial success or unexpected skips before starting a second upload that could obscure the original result.
| Verification | Pass condition | Evidence to retain |
|---|---|---|
| Intended values | match approved changes | before and after comparison |
| Product identity | same intended products updated | handle mapping |
| Option combinations | expected choices remain | representative product review |
| Unchanged fields | no unexplained differences | exception report |
| Storefront purchase | selected option and price agree | product and cart check |
| External owner | no immediate conflicting overwrite | sync owner confirmation |
Where a separate system owns part of the catalogue, confirm how it reacts. A successful manual update can disappear when the next scheduled sync restores an older value. Solve ownership and source-of-truth disagreements before treating repeated uploads as a sustainable process.
Use the product taxonomy and navigation guide if the project also changes category structure. That is a separate merchandising decision and should not slip into a routine data upload unnoticed.
Prepare a realistic recovery route
Keep the original export, approved import, difference report and execution time together. A product CSV is useful recovery evidence, but it is not a complete backup of the store or all app-managed data. Identify which affected fields can be restored through the same mechanism and which require another route.
If a problem appears, stop subsequent changes and establish its scope first. Reimporting an old full export can overwrite legitimate edits made since it was taken. A focused correction may be more appropriate, but it still needs comparison, approval and a representative check.
For recurring updates, evaluate whether a controlled integration or another catalogue workflow would reduce repeated manual risk. The decision should follow field coverage, volume, ownership and recovery needs. A larger tool is not automatically better, but an unreviewed spreadsheet is not a long-term governance model either.
StoreBuilt point of view
The safest catalogue update is one whose intended differences can be explained and whose unintended differences can be detected. StoreBuilt would prioritise field-level scope, representative pilots and verified recovery over a fast import of a large file. Completion means the catalogue behaves as intended after the upload.
Explore Shopify design and development for catalogue-connected storefront work, or Contact StoreBuilt with a description of the update and the systems that own its data.