KNOWLEDGE BASE
How do you migrate e-commerce data and URLs?
Table of Contents — English
Which inventory should you compile before migrating?
List your product, variant, category, image, customer, address, order, and content records. For each record type, define the source, target, ID field, and the scope of history to be kept. Record integration connections and external system IDs separately. The store pages you can see are not the whole inventory.
Passwords and payment tokens require special consideration. You cannot automatically assume that the target system supports the existing password hashes or that the provider allows token transfer. Where needed, plan a secure account reset or a provider-supported migration path.
In what order is the data transfer carried out?
1. Take a read-only copy of the source data and prepare the field mapping. State the ID, date, currency, and unit-of-measure conversions explicitly.
2. Identify duplicates and gaps. Clean up identical SKUs, orphaned variants, and broken image URLs. Cleanup decisions should be documented.
3. Transfer in dependency order. Categories and products may be needed before the variants and relationships attached to them; follow the rules of the target model.
4. Run a rehearsal transfer in a test environment. Rerunning the same transfer must not create a second copy. The mapping between old and new IDs should be kept permanently.
5. Besides record counts, validate sample content and totals. Critical fields such as order amount, number of line items, date, and status should be compared one by one.
How is the final data delta handled?
New records appear between the rehearsal and the live cutover. You should choose between a full-downtime transfer and a controlled delta transfer. If change times are unreliable, selection based only on timestamps can skip records. Deleted records and retroactive changes should be assessed separately.
How do you map old URLs?
Collect your important old addresses from crawls, sitemaps, and analytics records. Link each address to its most relevant new counterpart. For URL changes, Google recommends using permanent server-side redirects and handling internal links and sitemaps that use the new URLs.
If a product has no real counterpart, plan a response appropriate to the content's status instead of redirecting all old addresses to the home page. Check redirect chains and loops. The canonical, the sitemap, and internal links must not point to different targets.
How do you validate the live cutover?
With controlled tests that take no real payments, try product selection, order recording, notifications, and transfer to external systems. Check that test settings that block indexing have not been carried over to the live environment. After the cutover, monitor broken links, redirects, order errors, and the status of important pages.
The conditions for a rollback decision and the person authorized to make it should be determined in advance. If the target system has received new orders, it must be explained how those records will be reconciled with the old system. Restoring the old backup alone does not provide a consistent rollback.
Frequently asked questions
What is the next step?
Prepare your current data and URL inventory, and review the rehearsal and live cutover plan with PROJX Digital.
