A catalog import is successful only when the right products, variants, prices, and images appear correctly. A completed upload message does not prove that the store preserved the meaning of the source data.
Related service: Website and ecommerce development- 01Profile source records
- 02Map fields and identifiers
- 03Run a trial import
- 04Reconcile exceptions
- 05Publish approved products
SECTION 01
Agree what an import may change
Ask the ecommerce agency to document product and variant identifiers, categories, prices, units, publication states, and media references. Decide whether the import creates records, updates records, or both. Define what happens to omitted fields and records no longer present in the source. Never assume that deleting a row from a spreadsheet safely removes the corresponding live product.
Shopify's CSV documentation provides a concrete example of field dependencies and different overwrite behavior. Its variant guidance illustrates the distinction between a product and a sellable combination. These are reference models, not confirmation that a specific platform migration is included. Verify the selected store's supported import behavior before preparing the production file.
SECTION 02
Illustrative example: sizes become separate products
Imagine a supplier exports workwear with one row per size. The receiving store expects those rows to belong to a shared product with distinct variants. A naive import creates several nearly identical product pages, loses the size selector, and attaches every image to the first row. The record count still appears plausible.
A trial batch includes this product, a simple product, a missing image, and an existing SKU. The reviewer checks grouping, options, price, stock identity, and public status. Incorrect rows return to a visible correction list. This hypothetical example shows why representative edge cases are more valuable than importing an arbitrary first handful of clean records.
SECTION 03
Make reconciliation part of acceptance
Compare the source and destination by meaningful totals and samples: products, variants, active listings, missing prices, unavailable images, and rejected records. Check decimal separators, units, duplicate identifiers, and text encoding. Keep products unpublished until the responsible business reviewer approves the result. A production rerun should not create duplicate records or clear fields unexpectedly.
Ask for a correction plan and a record of the actual file used. Reverting an import may require more than restoring a spreadsheet, especially if orders were accepted against newly created identifiers. Separate catalog work from inventory and order migration where the platform treats them differently. The handoff should explain future import rules so staff can repeat a safe process instead of relying on an undocumented one-time script.
SECTION 04
Frequently asked questions
Is a CSV upload enough for migration?
Only when it covers the agreed data and its behavior has been verified. Accounts, history, inventory, and other records may need separate handling.
Should missing prices become zero?
Not by default. Flag the record for review unless an approved business rule explicitly permits that value.
How many records should a trial include?
Choose enough to represent the actual product patterns and failure cases, rather than an arbitrary number of easy rows.
PRIMARY REFERENCES
Sources and further reading
These references cover the standards, platforms, or published prices discussed in the guide. Worked examples and checklists are our editorial guidance.
- Product CSV import and exportShopify Help Center
- Product variantsShopify Help Center
EDITORIAL METHOD
About this guide
We use AI to assist with drafting and editing. Catapult AI Work is responsible for the published content. Examples illustrate possible approaches; they are not client case studies unless identified as such.
Budget examples are not Catapult package prices. Check linked provider pages for current fees and plan limits before making a purchase.
Read the editorial policy