Measurement

Geo-Testing

Validate channel impact with geographic experiments — the gold standard for causal measurement.

Geo-Testing & ExperimentationComing Soon

Overview

Geo-Testing enables controlled geographic experiments to measure the true causal impact of your marketing channels. By turning ads off in specific regions and comparing against control regions, you get definitive proof of whether a channel actually drives incremental revenue.

Key Features

  • Full Test Lifecycle: Create, configure, launch, monitor, and conclude geo-lift tests from a single interface
  • AI Test Designer: Get AI-powered recommendations for optimal test design — regions, duration, and expected statistical power
  • Candidate Finder: Automatically identifies test-suitable regions based on your conversion data patterns
  • Test-Twin Pairing: Matches test regions with statistically similar control regions for reliable results
  • Power Analysis: Know in advance how long your test needs to run for statistically significant results
  • Progress Monitoring: Track test progress in real time with confidence interval updates
  • Auto-Conclusion: Tests automatically conclude when statistical significance is reached

How It Works

Geo-Testing works by creating holdout experiments: you stop advertising in specific geographic regions (test group) while continuing as normal in similar regions (control group). After the test period, the platform measures the difference in conversions between groups, isolating the causal effect of your advertising.

Test Lifecycle

  1. Design: Choose channels, regions, and duration using the AI Test Designer
  2. Launch: Pause ads in test regions (manual action in ad platforms)
  3. Monitor: Watch real-time progress with daily lift estimates
  4. Conclude: Get final results with confidence intervals and statistical significance
  5. Calibrate: Feed results back into MMM for more accurate future models

Use Cases

  • Prove Incrementality: Get causal evidence that a channel drives real revenue lift
  • MMM Calibration: Use geo-test results as Bayesian priors to improve MMM accuracy
  • Budget Justification: Provide leadership with scientific evidence for marketing investments
  • Channel Assessment: Test whether turning off a channel would actually hurt revenue

Getting Started

  1. Open Geo-Testing from the sidebar
  2. Use the Candidate Finder to identify suitable test regions
  3. Create a test using the AI Test Designer
  4. Pause ads in the selected test regions
  5. Monitor progress until the test reaches significance