The challenge
Ecommerce ran on Google Cloud and BigQuery, stores on Snowflake, and finance on Azure. Every cross-channel question required a custom data pull. Merchandising couldn't see inventory and demand in the same place during peak season.
What we did
Instead of a big-bang migration, Tulan mirrored Snowflake and created shortcuts to BigQuery exports so OneLake became the single logical lake in six weeks. Gold-layer models joined channels for the first time. Workloads then migrated in waves, prioritized by cost and value, with Fabric Run taking over operations.
The outcome
Merchandisers now see demand, inventory, and margin by store and channel daily. A markdown optimization model built on the unified data added $2.7 million in annual margin. Two legacy ETL tools were retired.
“We got the unified view in six weeks and migrated when it made sense, not when a project plan said so.”
VP, Data and Analytics