Magento and Adobe Commerce
Magento migration: legacy containment, Magento 2, and moves off it. The platform with the deepest footprint in our work.
Services
Page
Lazy Ants · eCommerce
eCommerce Replatforming & MigrationChange platforms while the store keeps selling.
eCommerce replatforming in staged cutovers. Each one is validated before the next opens, each one can be rolled back on its own, and every system that touches an order has a named owner.
Replatforming & Migration
Step 1 · Architecture Review
$3,850Fixed price
2 weeksFrom day zero
Credited toward the migrationUp to 10% of its value, for 60 days.
Then
- Verdict: migrate, optimise or wait
- Migration at a fixed price per phase
- 30 days of hypercare
Replatforming & Migration
Step 1 · Architecture Review
$3,850Fixed price
2 weeksFrom day zero
Credited toward the migrationUp to 10% of its value, for 60 days.
Then
- Verdict: migrate, optimise or wait
- Migration at a fixed price per phase
- 30 days of hypercare
- No planned downtime
- NAF NAF: about 13,000 products moved in two months, store live throughout
- Rankings held
- Rental 24, SEO preserved under staged exposure
- Legacy contained first
- Adampolis, before anything was moved
What actually goes wrong
Moving the data is the part everyone budgets for
Tools will copy products, customers and orders between platforms for a few hundred dollars, and they do it well. None of that is where migrations lose money. The losses land in the week around the switch, and they land in four places.
The cutover is one event instead of many
Everything moves on a Saturday night, and there is no way back by Sunday morning. A migration that ships in units can be stopped after any one of them.
Two systems write the same record
Orders live in the storefront and in the ERP, stock in the ERP and in a cache. During the move both are live at once, and nobody owns reconciliation.
SEO signals change silently
Routing, rendering and preferred-URL behaviour all shift with the platform. Cutover day looks normal, and the traffic drop arrives six weeks later, when the cause is hard to prove.
Nobody is named
When checkout breaks at 9am, the question that costs money is who decides whether to roll back, and how long that decision takes.
Five phases, each with a gate you can stop at
Every phase is priced fixed and ends in a condition that is either met or not. You can end the engagement at any gate and keep everything produced up to it.
- Discovery2 weeks
The eCommerce Architecture Review: current state, integrations, order data and the platform's real limits, read from your own systems. Ends with the verdict, the scope and the budget range.
Gate: the verdict is migrate and the scope is agreed, or we stop here and you keep the report.
- Foundation
The cutover units are set from the review's scope, and the target platform stands up with the integrations that carry orders and stock. Nothing is exposed to customers yet.
Gate: orders flow end to end on staging, reconciled against the old system.
- Staged cutover
Traffic moves unit by unit: catalog and search, product pages, cart, checkout and payments, customer accounts, then the order flows into the ERP, CRM and PIM. Each unit is validated before the next opens and can be rolled back on its own. The first units can go live while the foundation for the later ones is still being built.
Gate: each unit meets its acceptance criteria before the next opens.
- SEO preservationRuns alongside
Preferred URLs, crawl behaviour and indexing signals are checked at every exposure step, so a drop is caught while it is still one unit wide.
Gate: indexing and routing verified at each exposure step.
- Hypercare30 days
Checkout and payment callbacks, order export to the ERP, stock, indexing, 404s and redirects, and the speed of key templates are watched every day. In the first week, orders are reconciled between the storefront and the ERP daily.
Gate: the 30 days end, and the store moves to Support & Maintenance with every open issue named.
Scope
What an eCommerce replatforming includes
Plan
The migration plan and a risk map by cutover unit, with the rollback point for each one.
Build
The target platform built in stages next to the live store. Nothing reaches customers before it passes its gate.
SEO
A replatforming SEO package: the redirect map, and a parity check of URLs, metadata and indexing before and after every unit.
Integrations
ERP, PIM and CRM integrations rebuilt against the new platform, with a named owner for every field that two systems write.
After launch
Thirty days of hypercare, then the store moves to our Support & Maintenance.
Platforms
What we move, and what we will not pretend to
The method is the same wherever the store is going. The platforms we have actually carried to production are below. If yours is not here, say so on the first call and we will tell you honestly whether we are the right team.
Shopify and Shopify Plus
In both directions: onto Shopify, and off it when the operating model outgrows it.
Sylius Plus
NAF NAF moved here from Shopify Plus with the store live throughout.
Shopware
BigCommerce
WooCommerce
When more than the platform changes
Headless and multi-market moves
An ecommerce platform migration is often the moment a team also goes headless or opens new markets. Both multiply what can break, so both are planned as cutover units of their own.
Headless
Separating the storefront adds deployable parts, APIs and release points, and search visibility starts to depend on rendering. We plan it layer by layer, only where it pays for itself, with an owner for every component and gates that check performance and SEO signals before exposure expands. Mantinga runs its frontend on this model.
Multi-market
Catalogs, currencies, tax rules and SEO surfaces multiply with every market. Each market goes live on its own, only after its pricing, catalog and SEO are checked, and it can be paused or rolled back without touching the others.
Four moves, four different reasons they could have gone wrong
Phased replatforming without downtimeShopify Plus to Sylius Plus in two months, about 13,000 products, with the store live. Staged cutovers, validation gates, ownership assigned to the riskiest areas first.Read the caseMagento legacy to a controlled migration planCoupling and fragile release paths contained before anything moved, so the blast radius was known in advance.Read the caseRental 24SEO preservation during platform changePreferred URLs, crawl behaviour and indexing controlled while the rollout happened in stages.Read the caseConfidential retailerMulti-market expansion without breaking the platformCatalogs, currencies, tax rules and SEO surfaces multiplied through staged exposure and measurable gates.Read the caseWhere to start
We do not quote a migration before we have read the system
A number produced from a call is a guess, and on an ecommerce replatforming a guess this size is either padded or wrong. Start with the review. Two weeks, fixed price, and it ends with a verdict, a scope and a budget range you can take to your board. It is the first phase of the migration, paid up front and credited toward the project. If the verdict is that you should not migrate yet, we say that too.
Questions we get before a migration
Can you do it without stopping sales?
Yes, and it is how every migration we run is set up. NAF NAF's store kept selling through its whole move from Shopify Plus to Sylius Plus: about 13,000 products, two months. Staged cutovers are what make that possible: traffic shifts in units, and any unit can be rolled back on its own.
How much does a migration cost?
We do not answer that before reading the system, and neither should anyone else. The review ends with a scope and a budget range built from your own data. It is the migration's first phase, and it is credited toward the project.
How long does it take?
NAF NAF took two months for about 13,000 products: two weeks of discovery and six weeks building the new platform and its integrations, with the first units going live while the last parts were still being built. Yours depends on the number of cutover units and how many systems write order data, both of which the review establishes.
Will we lose search traffic when we replatform?
SEO preservation runs alongside the migration from the first unit. Preferred URLs, crawl behaviour and indexing signals are checked at each exposure step, because the drop from getting this wrong arrives weeks later, when it is hardest to fix.
What if the review says we should not migrate?
Then you have saved a six-figure decision for $3,850, and we will tell you what to fix instead. The credit does not depend on the verdict, so we have no financial reason to prefer one.
What happens after launch?
Thirty days of hypercare. We watch checkout and payment callbacks, order export to the ERP, stock, indexing and redirects, and reconcile orders against the ERP every day in the first week. A broken checkout gets a response within an hour during the working hours we agree with you. Cover outside them is agreed around your season and peak traffic, and billed separately. Anything the migration broke, we fix at no charge. How long the old platform stays available for rollback is agreed with you before the last cutover. After thirty days the store moves to our Support & Maintenance, or hypercare is extended as paid work.
Can we go headless as part of the migration?
Yes, if the team that runs the store afterwards can own the extra moving parts. Headless adds components, APIs and release points, so we plan it as separate cutover units with gates of their own, and we will tell you if a platform-led storefront would serve you better.
What do you need from our team?
Access to the current platform, analytics and the systems that carry order and product data, and one named person on your side who answers our questions. Cutover windows also need someone who can make the business call to pause or roll back.
Do you rebuild our ERP and PIM integrations?
Yes, as part of the migration. Orders, stock and product data are where two systems end up writing the same record, so every integration gets a named owner per field and is reconciled against the old platform before any unit opens.
Tell us what you are running and what is forcing the move
If a migration is not the right answer, the review will say so, and that is a cheaper way to find out than starting one.
