The Situation
APAC expansion (Australia, New Zealand) required a 3PL that could reach markets far outside the US logistics network. Global-e was the business's chosen solution. It came with a fundamentally different operating model. Unlike GXO or CEVA, Global-e doesn't warehouse products directly, it connects to existing warehouses and handles international fulfillment from there. It also takes over payment processing and returns, sending reconciliation reports back rather than routing through Elastic Path like every other storefront. The result: a shorter order flow on our side, but a completely different shape.
The Decision
The path of least resistance was to add Global-e support into the existing order workflow. if/else conditions to branch behavior based on storefront/retailer. That approach had a serious problem: the existing workflow touches the US storefront, which handles the majority of order volume. Adding conditional logic into that flow meant any Global-e bug or edge case was one bad deploy away from affecting US orders.
And it made no sense structurally. The Global-e flow was actually simpler and shorter than the US flow precisely because it offloaded payment and returns. Embedding a simpler, different flow inside a complex one creates confusion in both directions.
The right answer was a clean separation. This was also the right moment to implement Fluent's new rules library which they claimed to be a faster, more robust foundation and were actively pushing their clients toward.
What We Built
Built on Fluent's new rules library, the core flow follows the same pattern as other storefronts: create order → create fulfillment. But workflows are also made up of custom rules, and this is where the separation paid off. Some rules could be shared across workflows, which we did. Others needed to be net new.
Global-e required a custom HOLD step that paused the order until a fraud check completed before fulfillment could proceed. That kind of workflow-specific logic would have been messy to bolt onto an existing flow. Built separately, it was clean and contained.
Impact
- APAC storefront launched cleanly with no disruption to existing US/UK order flows
- Germany integration followed the same pattern.
- Fluent's new library adoption proved out, the US flow eventually migrated to it as well
- Isolated failure surface: a Global-e issue stays in the Global-e workflow