Skip to main content

Getting Started with Statsig Warehouse Native

Statsig Warehouse Native brings Statsig's to your warehouse. Assess the impact of new features on-demand, on your compute, and mix and match use Statsig's SDKs with your own sources of metric and assignment data.

This is an appealing option for customers who are privacy sensitive or who are already using other solutions for experiment assignment. Our support team can help you decide if Statsig Warehouse Native is a good fit for your experimentation needs.

Walkthrough Guides​

Connect Your Warehouse

Connect Statsig to your warehouse to start analyzing experiments

Create Metrics

Connect to a metric source and configure metrics for analysis in Statsig

Start an Experiment

Connect to your assignment data in minutes, and start testing hypotheses

Run Pulse

Calculate and explore experiment results in your warehouse

Use Statsig SDKs

Use Statsig's SDK to get easy event logging and rich features like holdouts and mutual exclusion. Export events to your warehouse for analysis

Tools​

Statsig Warehouse Native integrates with Statsig's development tools:

Feature Gates

Control access to new functionality and automatically measure how it’s performing with an A/B test before you ship

Dynamic Configs

Dynamically configure user experience with configuration parameters instead of hard-coded constants

Segments

Define a reusable set of rules to target a group of users

Experiments+

Run sophisticated experiments with multiple variants (A/B/n tests) or mutual exclusion

Users

View users exposed to all gate checks and events logged with Statsig

Pulse

View how your experiments impact your metrics

Insights

View all experiments that have impacted a metric

Holdouts

Measure the cumulative impact of all features shipped over a period of time

What Statsig Warehouse Native Does​

With Statsig Warehouse Native, you set up connections to your source data using basic SQL queries and define Metrics as measures on top of that source data. Statsig will run a SQL-based data pipeline in your warehouse to run experimental analysis on your raw data. At the end, Statsig downloads a small result set with no user identifiers - just group-level summary data.

Simple-view

This process is fully transparent. The queries and cost data are available in the Statsig console, and algorithms like CUPED and winsorization are run in SQL to avoid black boxes and inconsistent results.

Nonetheless, the Statsig Warehouse Native pipeline supports many of the advanced features offered in Statsig, such as:

  • Support for multiple unit types - the same event can be used in experiments for a user_id, company_id, etc.
  • Timeseries and days-since-exposure breakdowns
  • Dimensional breakdowns - set up metrics to calculate dimensional breakdowns as well as a topline view
  • Custom queries - filter and group results by user-level attributes
  • Advanced metrics and corresponding statistics tools like Ratio Metrics, Funnels, and the Delta Method to adjust for metric covariance
  • CUPED, Sequential Testing, Outlier Removal, and more

You can also integrate Statsig's feature gates, rollout scheduling, holdouts, and layers (mutual exclusion) with Warehouse Native and still store and process the data in your warehouse.

info

Statsig Warehouse Native is only available for Enterprise contracts. Please reach out to our support team, your sales contact, or via our slack channel if you are interested in evaluating Statsig Warehouse Native.