Cloud to Warehouse Native Migration
Migrate from Statsig Cloud to Statsig Warehouse Native by setting up a new Warehouse Native project side by side, validating results, and switching over when you're confident.
Migrating from Statsig Cloud to Statsig Warehouse Native changes where Statsig computes your metrics: it sources them from your data warehouse instead. If all (or most) of your metrics already come from your warehouse, migration is relatively simple.
The recommended migration path is a side-by-side migration into a new Warehouse Native project, not an in-place conversion of your existing project. Your current project stays unchanged, so you can validate the new project at your own pace, compare results directly between the two, and switch over only when you're confident. The same approach applies if you're moving from one Warehouse Native project to another.
Migration overview
- Prepare your warehouse. Make sure your data warehouse is ready and contains the historical data you care about, including metric data and assignment (exposure) data, backfilled to the depth you need.
- Create a new Warehouse Native project and connect it to your warehouse.
- Set up your data foundations under the Data tab: metric sources, assignment sources, and your metrics catalog.
- Recreate your configurations (gates, experiments, and dynamic configs) manually or programmatically with the Console API.
- Run both projects side by side and compare results between your existing project and the new Warehouse Native project. Re-analyze experiments you've already run to confirm the numbers line up with what you expect.
- Switch your client-side SDKs to the new project's keys once you're satisfied with the comparison.
Why a separate project
Because your original project is never changed during the migration:
- No cutover window or change freeze: Your team keeps working in the existing project until you're ready to switch.
- Easy rollback: If anything looks off, keep using your existing project, which stays unchanged.
- Historical data stays accessible. Your past experiment results remain available in the original project.
- Direct validation: You can make an apples-to-apples comparison between the two projects before committing.
Before you start
- Review the feature comparison between Cloud and Warehouse Native so your team knows what changes. A few features behave differently on Warehouse Native, and it's best to identify those gaps before you migrate rather than after.
- Plan for the setup work in the new project up front: warehouse readiness, data backfill, and the Data tab configuration are the bulk of the migration effort.
- If you're an enterprise customer, work with your Statsig contact to evaluate feasibility and plan the migration.
Was this helpful?