Be clear about why you are moving
Replatforming is expensive, slow and risky to revenue, so the reason has to be worth it. Write the reasons down as problems, with evidence, and check each against a simple question: would the new platform solve this, or would the problem move with us?
Reasons a new platform solves
- Features the business needs are impossible or absurdly costly on the current one
- The platform or version is no longer supported
- Running costs or transaction fees no longer make sense at your volume
- It cannot connect properly to the ERP or other systems
- It fails under peak load despite tuning
Problems that follow you
- Poor product data
- Weak photography and descriptions
- Slow fulfilment and customer service
- An unclear offer or pricing
- No one responsible for the store day to day
If most of your reasons are in the second column, fix those first. They cost less to solve, and a new platform will inherit them otherwise.
Take an inventory of what the store does today
A store that has traded for years does far more than its owners can recite. Features were added, apps installed, workarounds built. Each one is either rebuilt, replaced or knowingly dropped. The ones nobody listed are dropped by accident, and discovered by customers.
| Area | What to list | Easily forgotten |
|---|---|---|
| Storefront features | Search, filters, bundles, wish lists, reviews, gift cards, loyalty | Back-in-stock alerts, store locator, size guides |
| Pricing and promotions | Discount types, customer groups, coupons, tax rules by region | Automatic discounts, staff discounts, legacy coupon codes still in circulation |
| Checkout and payment | Payment methods, delivery options and rates, address validation | Saved cards, instalment options, cash on delivery rules |
| Apps and extensions | Every installed add-on and what depends on it | Scripts pasted into the theme years ago |
| Integrations | ERP, accounting, warehouse, couriers, marketplaces, email marketing | Scheduled exports someone downloads by hand |
| Emails and notifications | Order, shipping, refund and account messages | Abandoned basket sequences, internal alerts to staff |
| Customizations | Custom code in theme and backend | Fixes for edge cases, such as a special tax rule |
| Reports | What finance, marketing and operations pull regularly | The spreadsheet that feeds the weekly meeting |
Build the list from the admin screens, the app list and the code, then walk it past each department. For every line, mark keep, replace, improve or drop, with a name beside the decision.
Choose the destination against the inventory
With the inventory in hand, platforms can be compared on facts. For each candidate, mark every kept line as built in, available as an add-on, needing custom work or not possible. The pattern shows the true cost and the real gaps.
Our comparisons of Shopify and custom ecommerce, headless and traditional ecommerce and Salla, Zid and Shopify cover the main choices. Whatever the shortlist, build a trial store with a sample of your real products and run your three most awkward orders through it before deciding.
Decide what data moves, and how
- 01
Products
What to plan forUsually move, after cleaning. Variants, attributes, images and categories rarely map one to one, so expect restructuring.
- 02
Customers
What to plan forAccounts and addresses move. Passwords are stored in a protected form that often cannot be transferred, in which case customers set a new one. Tell them in advance.
- 03
Order history
What to plan forDecide how much history customers and staff need to see in the new store, and where the rest remains available.
- 04
Open orders and returns
What to plan forOrders in progress at the switch need a rule: finish them in the old system, or move them.
- 05
Gift cards, store credit and loyalty points
What to plan forThese are money you owe customers. Balances must move exactly, and be reconciled.
- 06
Subscriptions and saved payment methods
What to plan forOften tied to the payment provider. Find out early whether they can be transferred.
- 07
Reviews and content
What to plan forProduct reviews, blog posts and pages, with their addresses.
- 08
Marketing consent
What to plan forWho agreed to what, with dates. Losing consent records can make your mailing list unusable.
Protect search visibility and links
Every platform structures addresses differently, so most pages will change address. Search engines, adverts, emails, bookmarks and other sites all point at the old ones. Without redirects, those visitors meet an error page and the rankings behind them fade.
- 01
Export every current address
From the sitemap, analytics, search console and a crawl. Include products, categories, pages, articles and images that attract traffic.
- 02
Map each to its new address
One to one where a page continues. To the closest relevant page where it does not. Not everything to the home page.
- 03
Carry titles, descriptions and structured data
Do not let the new theme's defaults replace text that has been refined over years.
- 04
Test the redirects before launch
On the full list, automatically, on a staging copy.
- 05
Watch after launch
Error reports, rankings and traffic for the pages that matter most, daily at first.
Our guide on migrating a website without losing important URLs describes this work in full.
A worked example
A sports retailer moves platform
This example is invented to show what an inventory turns up. A sports retailer is leaving a self-hosted platform whose version is no longer supported.
- ReasonsUnsupported software and a failing ERP connection, both of which a new platform solves. Thin product descriptions are noted as a separate job.
- InventoryThe list reaches 74 items. Fourteen were unknown to the ecommerce manager, including a nightly stock export used by a marketplace and a script that blocks delivery of bulky goods to islands.
- DataProducts, customers, three years of orders, reviews and gift card balances move. Customers will set new passwords. Loyalty points are converted to store credit after reconciliation.
- Addresses11,200 old addresses are mapped. A test run finds 600 colour variant pages with no destination, which are pointed to their parent products.
- CutoverThe switch is set for a Tuesday in the quietest month, eight weeks clear of the nearest sale. The old store stays online in read-only mode for a fortnight.
The island delivery rule would have been lost without the inventory, and the first sign would have been a courier's surcharge. Search traffic dips for a few weeks and recovers, which is the normal pattern when redirects are complete.
Plan the switch and the weeks after
- Timing. Well away from peak season, promotions and public holidays, early in the week, with the whole team available.
- Rehearsal. A full practice run of the data move and the switch on a staging copy, timed, with test orders through every payment and delivery option.
- Freeze. A short period in which product and content changes stop on the old store, so that nothing is lost between the final data copy and the switch.
- Go or no-go criteria. Agreed in advance: payments work, orders reach the ERP, stock is right, redirects pass.
- The way back. The old store kept intact, with a deadline for deciding to return. After real orders arrive in the new store, going back becomes hard.
- Customer communication. A message before the switch about new passwords and anything that changes, and clear help on the sign-in page.
- Close attention afterwards. For the first weeks, someone watches orders, payments, errors, search reports and customer messages every day.
Replatforming checklist
Ecommerce replatforming checklist
0 of 11 checkedWhat to do next
Start the inventory now, even if the move is a year away. Open the list of installed apps and extensions and write beside each what it does and who relies on it. That exercise alone often changes which platform makes sense.
Our e-commerce platform development page explains how we plan and carry out these moves. Tell us what you run today and what is holding you back, and we will say whether a move is the answer.
Common questions
How long does an ecommerce replatforming take?
It depends on catalogue size, the number of integrations and customizations, and how much is being improved along the way, so a general figure would mislead. The inventory gives a realistic basis for an estimate. Projects that try to redesign everything, restructure the catalogue and change platform at once take longest and carry most risk.
Will we lose search rankings when we change platform?
A temporary dip is common even when everything is done well, because search engines need time to process changed addresses. Lasting losses usually come from missing redirects, changed titles and content, or slower pages. A complete address map, tested before launch, and unchanged key content are the main protections.
Can customer passwords be moved to the new platform?
Often not. Passwords are stored in a protected form specific to each platform, and many platforms cannot import another's. In that case customers are asked to set a new password at first sign-in. Tell them beforehand, make the reset simple and brief your support team, since this is the most common question after a switch.
Should we redesign the store at the same time?
Some redesign is unavoidable, because themes do not transfer. Keeping the structure, navigation and content close to the old store reduces risk and makes any change in sales easier to interpret. Larger design changes can follow once the new platform is stable. If a full redesign is essential, allow more time and test with customers.
What is the biggest risk in replatforming?
Losing something the old store did that nobody listed: a tax rule, a delivery restriction, an export another system depends on, a discount some customers still use. These rarely appear in a specification because nobody remembers them. A thorough inventory, checked with every department, is the best defence.