Skip to content
Senior web & app development · 20 capabilities
oakstone.digital
Case study / Atlas Commerce

A storefront built for considered growth.

A Shopify migration and modular commerce build gave a fast-moving team a clearer path from interest to order.

Industry
Ecommerce
Duration
9 weeks
Division
Websites & Commerce
Year
2026
Atlas Commerce — ecommerce web development project by Oakstone
Websites & Commerce / delivered
+63%
mobile conversion rate
11 min
cutover window
1,240
URLs redirected, zero 404s
The challenge

Atlas sold technical equipment on an ageing Magento installation that cost more to host than it returned in flexibility. Every promotion required a developer, the catalogue had outgrown its navigation, and mobile conversion was less than a third of desktop.

A previous migration attempt had been abandoned after a test import lost order history.

The move

We ran the migration as a data project: full audit, field-level mapping, a complete dry run into a staging store, then reconciliation against the source system signed off by their finance lead before anything went live.

On the storefront, we rebuilt the catalogue taxonomy around how customers search, gave merchandisers editable content blocks, and rebuilt the mobile product page to answer specification questions in order.

The result

Cutover took eleven minutes in a Tuesday morning window, with 1,240 URLs redirected and no loss of order or customer history. Organic traffic recovered to baseline within three weeks and exceeded it by week seven.

Mobile conversion rose 63% against the pre-migration baseline, and promotions now go live without engineering involvement.

Stack
ShopifyLiquidHydrogen componentsAlgoliaMatrixifyKlaviyo
Team

One commerce lead, one designer, two engineers, their finance lead for reconciliation

Constraints
  • A previous migration attempt had lost order history, so trust was low
  • Cutover had to avoid the pre-season trading window
  • 1,240 indexed URLs to preserve
What we would do differently

The dry run took nine days rather than the five we estimated, almost entirely in reconciling historical order edge cases. We now quote reconciliation separately so it does not get squeezed.

Filter taxonomy should have been agreed with the merchandising team before the catalogue import, not alongside it. Two collections had to be restructured after go-live.

A better brief starts here

Your next proof point is waiting.

Bring us the hard thing. We will bring a point of view, a senior team and a clear next step.

Start a conversation