AI-powered Incrementality Test

Incrementality Test Introduction

A GeoLift experiment (often referred to as Incrementality Test) is, in general, the most accurate way to measure the true impact of marketing without relying on individual tracking, which can bias results due to privacy restrictions.

It splits the target market into similar test and control regions, launches ads only in the test regions while keeping the control regions as the holdout, and then compares the business outcomes between the two to estimate the incremental impact of the marketing activity.

Marketing Measurement Overview summarizes its benefits and limitations and compares this approach with other common methods.

The Maxma Advantage

While incrementality testing is the gold standard for measuring marketing impact, it has traditionally been limited to big brands and tech companies due to its high cost and complexity. Even modern vendors charge around $10,000 per experiment, making it impractical for most businesses to run tests at scale.

Maxma is changing that. Every test runs on the same automated platform, with two ways to run it:

• Self-serve — an AI expert with deep knowledge of attribution, statistics, and marketing best practices guides you through the full process. The most cost-efficient way to run a rigorous test — ideal for budget-conscious teams.
• Guided — Maxma's managed service: our measurement experts design the test with you, run it end to end, and walk through the results together. Hands-on support at a fraction of what traditional vendors charge — and what most teams choose.

Whichever way you run it, the platform removes the cost and complexity at every stage of the test — not just one:

• Expert-guided design — every design decision is guided: goals, KPI selection, and the power analysis that sizes budget, duration, and regions. See How It Works.
• Small holdouts — the most common reason teams never test is not wanting to turn off ads in half their market. With ~600 Commute Zones to select from instead of coarse units, the control group can often shrink to just 10–20% of the market — plus more sensitive measurement and local testing that coarse units can't support.
• Connected data — conversion data flows in through a one-click Shopify connection or a one-time database integration, for this test and every test after — no CSV wrangling per experiment.
• Push-button launch and revert — one click applies the geo targeting changes across your connected ad platforms, and one click restores everything when the test ends.
• Check in any time — watch incremental conversions accumulate during the test and conclude poor performers early instead of spending out the full budget.
• Marketer-ready results — incremental impact, cost per incremental conversion, and lift over time for every metric you measure, explained in Deliverables.

This end-to-end automation is what turns testing from a one-off project into a habit. Outschool ran four causal tests across funnel stages in their first two months on Maxma — scaling a top-of-funnel tactic that attribution had undervalued, right-sizing lower-funnel spend, and skipping a channel expansion that showed no incremental impact — an always-on testing cadence with minimal operational overhead.

Use Cases

Different Testing Scenarios

Maxma Incrementality Testing supports a variety of testing scenarios:

• Current Budget: Measure incrementality at the existing spend level.
• Heavy-Up: Test the impact of increasing ad spend to prevent diminishing returns and saturation.
• Dial-Down: Reduce spend safely with potential missed opportunities measured.
• New Launch: Before scaling a new advertising initiative, test for detectable incremental impact at the lowest possible spend level to minimize risk.

Multiple Metrics

Maxma Incrementality Testing supports measuring multiple metrics within a single experiment. Beyond a primary conversion metric (e.g. orders, signups, app installs, etc.), you can measure additional metrics for a broader view of marketing impact. Examples include:

• Conversion Funnel Stages – Track multiple points in the funnel, such as web visits, signups, and purchases.
• Conversion Segments – Compare new vs. returning customers, or Shopify vs. Amazon orders.

Note: The experiment is optimized to detect lift for the primary metric, so additional metrics may have lower statistical power (i.e. ability to get statistically significant results). But the platform will tell you what's the minimum lift you can detect for the addtional metrics so you can adjust the design or promote the metric as the primary metric.

Lingering Effect

Incrementality Testing can capture short-term lingering effects of advertising. While it cannot fully measure long-term brand awareness effects that may take months to materialize, it can capture delayed impact in the weeks following the advertising.

The results of the experiment will show the incremental conversions continue to grow after the experiment concludes.

Offline Advertising

Incrementality Test is not limited to digital advertising—it can also be applied to offline marketing channels, such as out-of-home, TV, radio, and print. The requirement is you can target only some of the regions with the offline advertising.

Getting Started

You can now Sign up to run a test end-to-end for free. Additional tests are charged per test see pricing. Once you're in the app, go to "Experiment" -> "Geolift" for our AI Incrementality Test product. If you want to signup with personal email (e.g. gmail), reach out to support@maxma.ai.

See What You Get

Curious what a finished test looks like before running one? Deliverables walks through the results — incremental impact, cost per conversion, and the lift charts. For the end-to-end flow from design to launch to results, see How It Works.

Sample Data

To help you get started, we've prepared some sample data that you can use to test the tool:

• 📊 Download Primary Metric Sample Data (CSV)
• 📊 Download Secondary Metric 1 Sample Data (CSV)
• 📊 Download Secondary Metric 2 Sample Data (CSV)