Moët & Chandon — SFCC backend development for a luxury champagne house
Salesforce Commerce Cloud backend work for the Moët & Chandon storefront — custom SFRA cartridges, OCAPI integrations and scheduled job workflows.
At a glance
- Brand
- Moët & Chandon
- Sector
- Luxury wines and spirits
- Engagement
- Dec 2021 – Nov 2023, at OSF Digital
- Platform
- Salesforce B2C Commerce (SFRA)
The brief
Moët & Chandon is one of the best-known champagne houses in the world, and its storefront runs on Salesforce B2C Commerce. Working at OSF Digital, I developed and maintained the backend of that storefront: the cartridges, the integrations and the scheduled jobs that keep catalogue and order data moving.
Luxury retail puts unusual pressure on an eCommerce backend. Product ranges are small but heavily merchandised, campaigns are tightly scheduled around seasonal peaks, and the tolerance for a broken checkout on a gifting weekend is effectively zero.
What I built
The work sat in custom SFRA cartridges layered over the base platform. Keeping customisation inside cartridges rather than modifying core code is what makes an SFCC implementation upgradeable — the cartridge path resolves overrides in order, so platform updates can land without unpicking bespoke changes.
Day to day this meant:
- Controller and model extensions over SFRA for storefront behaviour specific to the brand
- ISML template work where backend data had to reach the presentation layer
- Business Manager configuration — custom preferences, site settings and the job schedules behind them
Integrations
The storefront exchanged data with systems outside Commerce Cloud through OCAPI, the platform's REST layer. That covered reading and writing commerce data from external services and exposing what downstream systems needed.
Automated job workflows
A large share of the work was scheduled jobs: the recurring processes that import catalogue data, synchronise inventory and export orders. Jobs are where SFCC implementations tend to fail quietly — a feed that silently imports nothing looks identical to a feed with nothing to import until someone checks the storefront. Building them with explicit logging and failure states is what makes that difference visible.
Outcome
The storefront ran on this backend throughout my time on the account, from December 2021 to November 2023.
<!-- TODO(gizem): if you have shareable figures — release cadence, job runtimes, incident reduction, catalogue volume — this is where they belong. I have deliberately not estimated any. -->