Geo Incrementality
Causal measurement using geographic holdouts. Estimate the counterfactual with time-based regression, synthetic control, augmented synthetic control (ASCM), synthetic difference-in-differences (SDID), or Bayesian structural time series (BSTS) models—then compare results side-by-side with uncertainty intervals for every counterfactual-derived metric.
Centralized Truth
All your tests, one database
Methodology & Algorithm
We utilize a custom, proprietary algorithm built from the ground up to identify optimal treatment and control regions. This is the most powerful design algorithm available, performing:
- Market StratificationClustering of geographic units based on historical conversion volume and trend correlation.
- Synthetic Control GenerationConstruction of a counterfactual baseline using a weighted combination of control markets.
- Power Analysis SimulationMonte Carlo simulations to estimate Minimum Detectable Effect (MDE) across various duration and budget scenarios.
Experimental Design Parameters
Advanced Statistical Models
Select the counterfactual model(s) you want to run—or run them all—and compare every result. We show every model run and the uncertainty intervals for all metrics that incorporate the counterfactual baseline.
Time-based Regression
RegressionA transparent baseline that learns time-based relationships between treatment and control markets.
Synthetic Control
SCConstructs a synthetic counterfactual from a weighted combination of control markets.
Augmented Synthetic Control
ASCMAugmented synthetic control method that improves robustness when perfect controls are hard to find.
Synthetic Difference-in-Differences
SDIDDouble-robust approach that combines unit and time weights for counterfactual estimation.
Bayesian Structural Time Series
BSTSState-space time-series modeling that produces a probabilistic counterfactual and credible uncertainty.
Multi-Model Validation
Run one model or run them all—then compare results side-by-side to stress-test your lift estimates.
Full Transparency
We don’t hide models or cherry-pick outputs. Every model run is visible so you can understand what each approach implies.
Uncertainty Intervals
View wide and narrow uncertainty intervals for all metrics that incorporate the counterfactual baseline (not just the counterfactual series itself).
Use Cases
- • Validating pre-test design assumptions
- • Finding synthetic controls when treatment is already selected
- • Comparing and validating model outputs
Enterprise-Grade Test Management
A centralized platform to design, execute, and analyze all your geo incrementality tests. All stored in one database for future reference.
Geo Tests
Manage your geo incrementality tests
| Test Name | Status | Type | Start Date | Results (iROAS) | |
|---|---|---|---|---|---|
| Q4 Holiday Blitz - Holdout | Running | Geo Holdout | Oct 15, 2025 | Pending | |
| YouTube Scale Test | Analyzed | Scale Test | Sep 01, 2025 | 3.42 | |
| TikTok vs Reels Interaction | Analyzed | Multi-Cell | Aug 15, 2025 | 1.85 | |
| Summer Promo - Low Spend | Completed | Scale Test | Jul 01, 2025 | Calculating... | |
| Brand Awareness Lift | Draft | Geo Holdout | - | - |
Calibrate & Validate MMM
Geo lift tests create causal ground truth you can use to calibrate and validate our future MMM model implementation.
Supported Geo Channels
Test any channel where you can control spend or delivery by geography. Store results centrally to build your calibration database.
User-Level Tests
Coming soon: Run housefile and prospect list tests for channels where you control the audience.
Supported Experimental Designs
Flexible configuration for various causal inference needs.
Geo Holdout (Exclusion)
Standard exclusion testing where a specific region is suppressed to measure the incremental contribution of a channel.
Multi-Cell Interaction
Factorial designs to measure interaction effects between multiple channels or tactics.
Scale & Spend Tests
Response curve modeling by testing varied spend levels across comparable regions.
Incrementality Measurement
Pure lift measurement for channels with poor attribution visibility (e.g., TV, OOH).
Measurement Ops Make the Product Repeatable
The models and tests matter, but the workflow around them matters too. Shako Stats is designed to become the operating system around experiment planning, metadata, documentation, and cross-test learning.
Data Management
Centralize datasets, mappings, and historical records so experiments and models always start from the same source of truth.
Tags and Parameters
Organize tests by audience, creative strategy, bidding logic, or business objective so learnings remain searchable and reusable.
Gantt Timeline View
See what tests are planned, in-flight, or completed so overlapping interventions and measurement conflicts are easier to manage.
Documentation
Turn methodology, definitions, and experiment design guidance into an internal operating system instead of leaving them scattered across decks.