Data Warehouse Migration Readiness Checklist
A readiness checklist for migrating an enterprise data warehouse to a cloud platform. It covers source-to-consumer mapping, SQL translation, target layout design, reconciliation, and cost forecasting. Use it to protect analytics pipelines and dashboards during the move.
When to Use This Checklist
Use this checklist when planning to move an enterprise data warehouse from a legacy system, such as on-premises Teradata, Netezza, or an older SQL Server warehouse, to a cloud platform like Snowflake, BigQuery, or Databricks. A data warehouse sits at the center of analytics: many pipelines feed it and many dashboards depend on it. Readiness work protects those connections.
How to Use This Checklist
Begin with a full discovery of what feeds the warehouse and what consumes it. The consumer map is as important as the schema inventory, because a migrated table with no reconnected dashboard delivers no value. Move through planning to translate SQL dialects and design the target storage layout for your real query patterns. Treat reconciliation and cost forecasting as required gates: a warehouse that returns wrong aggregates or blows the budget is not a successful migration.
What Good Looks Like
A ready migration has a complete source-to-consumer map, a sensitivity classification driving access controls, and a target schema designed around actual query patterns rather than a literal copy of the old layout. Reconciliation checks confirm that aggregates match between source and target. Costs are modeled before go-live, and a parallel-run period lets analysts trust the new platform before the old one is switched off. Lineage lets any consumer trace a metric back to its source.
Common Pitfalls
The biggest mistake is a like-for-like lift that ignores the target platform's strengths, leaving you with poor performance and high cost. Teams often forget downstream consumers and discover broken dashboards after cutover. Cost surprises are common because cloud warehouses bill on compute and scanned data, not fixed capacity. Skipping reconciliation lets subtle aggregation errors reach executives. Finally, decommissioning the legacy warehouse too soon removes your safety net before confidence is earned.
Related Resources
Design the target around medallion-architecture and data-lakehouse-architecture, and choose load patterns using elt-vs-etl-best-practices. Build trust with data-lineage and data-quality-management so every migrated metric is traceable and verified.