Experiment Result Exports
Export Statsig experiment results into your data warehouse or BI tool for custom reporting, executive dashboards, and longitudinal experiment analysis.
How experiment result exports work
Statsig lets you export both the reports and the raw data your feature rollouts and experiments generate.
Export methods
- Download experiment results from the Console as a CSV file - including a summary view, exposures and the raw data. This is for one-off downloads and analysis.
- Programmatically export the data underlying Pulse.
- For ongoing data exports, Statsig supports data integrations through customer data platforms like Segment, RudderStack, and mParticle. There is also a generic webhook if you want to build your own integration. To set up a daily export into your data warehouse, go to Data warehouse exports.
Validate exported data
Many teams audit and compare their data in Statsig with what they have in other systems. Statsig uses well-recognized statistical methods and industry best practices, so you can reproduce results yourself.
Tips for validation:
- Start small: Use a single day of data to reduce variables. When comparing experiments, use a full day (not days the experiment started or stopped, where there is partial data).
- Third-party ad/tracking blockers can block events sent to third-party services. A server-side integration or proxying requests through your domain removes this issue.
- Check for time zone conversion issues to confirm a consistent definition of "day" is being used.
- Statsig applies winsorization on metrics to remove outliers.
- Statsig applies CUPED to reduce variance on metrics using pre-experimental data. Disable CUPED when comparing console results to external data.
Was this helpful?