The situation
Magento 2 does the job, but it charges rent for it: dedicated hosting, a developer for changes that should take minutes, and an upgrade cycle nobody looks forward to. For a catalogue in the tens of thousands of SKUs, the cost of simply keeping the lights on had stopped being justifiable.
The complication was Brightpearl. The ERP — not the storefront — was the system of record for inventory, pricing and fulfilment, and it was working. This was never a project to replace it. It was a project to move the storefront out from under Magento without the ERP noticing.
What we did
Migrated the catalogue with reconciliation, not just import. Products, variants, options, images and category structure moved across, then got counted back against the source until the totals agreed. On a catalogue this size the import is the easy half; proving nothing was quietly dropped is the half that takes the time.
Preserved the URL structure. Every existing product and category URL was mapped to its Shopify destination before launch. Replatforming and redesigning the information architecture at the same time is the single most common way merchants lose their rankings in a migration, so we deliberately did not do both at once.
Kept Brightpearl authoritative. The integration was rebuilt so stock levels, pricing and order flow continue to originate in the ERP, with Shopify as the presentation layer. Sync runs on a schedule, backs off cleanly when the API rate-limits, and reports when it cannot complete — because an ERP integration that fails silently produces overselling, and overselling produces refunds.
Moved merchandising rules into Shopify Functions. Logic that had lived in Magento templates — which products are excluded from discount codes, which ship free and which do not — became structured metafields plus Functions that read them. The rules are now data the merchant controls, not code a developer has to change.
Automated what the migration exposed. Replatforming surfaced the manual routines that had been propping up the old store: availability maintained by hand, promotion eligibility maintained by hand, shipping profiles corrected by hand after the fact. Each became a scheduled job that derives the answer from the source of truth instead of relying on someone remembering.
Why it mattered
The headline is a cheaper, faster storefront. The durable win is that the day-to-day maintenance the old setup demanded largely stopped existing — and the ERP the business actually runs on never had to change.
Migrations get judged on launch day. They should be judged six months later, on how much manual work the new setup quietly requires.