Skip to main content

Warehouse Native

ABsmartly's Warehouse Native mode lets you run experiment analysis directly inside your own data warehouse. Instead of sending event data to ABsmartly's cloud for processing, you point ABsmartly at your existing warehouse tables and it computes results where your data already lives.

Why Warehouse Native?

Most experimentation platforms follow a cloud-first model: events are sent to the platform, metrics are computed there, and results are served back. This works well — it's fast to set up, provides real-time monitoring, and has no dependency on your data infrastructure.

But as experimentation programs mature, new requirements emerge:

  • Single source of truth — Experiment metrics should match what finance and BI teams report. When experimentation data lives in a separate system, numbers can diverge, eroding trust in results.
  • Data residency & compliance — Regulations like GDPR, SOC2, and HIPAA may require that user-level data stays within your own infrastructure. Warehouse Native means no user-level information leaves your environment.
  • Auditability — When results are computed inside your warehouse, you have a full audit trail using your existing governance tools — who accessed what, when, and how metrics were defined at the time.

How it works

With Warehouse Native, ABsmartly handles experiment assignment, management, statistical analysis, and metric governance. But the underlying data — exposures, goals, and attributes — stays in your warehouse.

You configure a data source that connects ABsmartly to your warehouse, map your tables to ABsmartly's expected schema, and set a refresh schedule. ABsmartly then queries your warehouse on that schedule to compute experiment results.

Supported warehouses

WarehouseStatus
BigQuery✅ Available
Snowflake✅ Available
ClickHouse✅ Available
Redshift✅ Available
Databricks✅ Available

Cloud vs. Warehouse Native

CloudWarehouse Native
SetupQuick — no data infrastructure neededRequires existing warehouse and ETL pipelines
Data speedReal-time monitoringBatch-processed (minutes to hours, depending on your refresh schedule)
Metric alignmentSeparate from BI/finance dataSame source of truth as the rest of your business
Data residencyData processed in ABsmartly's cloudData never leaves your warehouse
Cost modelIncluded in subscriptionYou control (and pay for) warehouse compute

Next steps

Ready to set up Warehouse Native? Head to the Get Started guide to connect your warehouse, map your tables, and configure data freshness.