The Test Configurations tab is where you define the parameters of an incrementality test in Funnel. This is the first step in measuring if your marketing campaigns are causing incremental performance beyond the baseline.
Each test configuration in Funnel compares a group of regions that receive marketing exposure, which is the test group, against a similar group that does not, which is the control group. Funnel applies a statistical model to estimate the lift caused by the campaign.
Types of tests
Before creating a test in Funnel, it is important to understand the two types of incrementality tests available: inverse and uplift. Each test type measures incrementality in a different way, depending on whether you introduce or remove marketing activity in a given region or audience. Choosing the right test type ensures accurate measurement and reliable results. Let’s look into each of these tests.
Inverse test
An inverse test measures the effect of pausing or reducing media in the test group. Use this test when you want to measure the effect of removing or pausing media investment. In an inverse test, you stop or reduce marketing in the test group, and continue with the normal campaign exposure for the control group. An inverse test helps you understand how much value a campaign provided before it was turned off.
Uplift test
An uplift test measures the effect of increasing or introducing media in the test group. Use this test when you want to measure the effect of adding or increasing media investment in a region. In an uplift test, the test group receives additional or new marketing, and the control group receives no exposure or maintains current levels. An uplift test is ideal when you are launching a new campaign or scaling an existing one in selected areas.
Each test uses campaign metrics, such as conversions, revenue, or CPA, across defined dates and regions to determine whether the media investment caused a measurable impact.
Guidelines for incrementality testing
Ensure that the data you exported into Funnel Measurement is available in the View Input Data tab before creating a test.
Make sure that the test and control groups are similar in demographics and behavior.
Funnel does not support overlapping groups. Each region can appear in only one group per test.
Maintain consistent tracking and attribution setup across all regions.
The test groups must have sufficient volume and duration for analysis. Ensure that the training data is available for at least a year before the start date of the test.
Short test durations or small audiences may lead to inconclusive results.
Uplift and inverse tests cannot run on the same media source during the same time window.
You cannot edit a test configuration after you save it. Delete and recreate it if you want to modify it.
Create an incrementality test
Complete the following steps to create an incrementality test.
In Funnel Triangulation, go to Incrementality > Test Configurations.
A list of previously created tests appears. The Test Configurations screen has the following fields:
Name: Name of your incrementality test.
Test Type: Mentions the type of incrementality test: inverse test or uplift test.
Metric Name: The metrics for which a test was run.
Source: Sources of the metrics for the campaign.
Training Start: The date from when the training or historical data is available. This is the date when Funnel Triangulation starts modeling the data.
Test Start: The date from when you start a test.
Test End: The date when the test ends.
Actions: You can Preview Test or Analyze Test.
Click Create Test.
The Test Configuration window appears.
Fill in the details for the test.
The Test Configuration window has the following fields:
Test Name: Enter a unique and descriptive name for the test. Use a naming convention that clearly reflects the purpose of the test.
Test Type: Select Inverse Test or Uplift Test based on the purpose of the test.
Metric Name: Select a metric. The options are populated from the data you export into the Funnel Measurement destination.
Source: Enter a source name of the metrics. It could be an ad platform or any paid social prospecting.
Training Start: Enter the date from when Funnel Triangulation should model the training data.
Start Date: Enter the date from when you want to start the test.
End Date: Enter the end date of the test.
Note: Funnel Triangulation will create a model based on the Training Start date and the control group to predict for the future. This data is then compared with the data that Funnel Triangulation models for the test group for the defined period. The difference in the outcomes between these groups is the lift.Planned Spend: Enter the amount you are spending on the test regions. You can enter only a numerical value in this field. The currency is the default value you set for the campaign.
Description: Enter a description for your reference.
Test Group: Select the regions for the test group. The options are populated from the data you export into the Funnel Measurement destination.
Control Group: Select the regions for the control group. The options are populated from the data you export into the Funnel Measurement destination.
(Optional) You can click Add All Remaining States to add all the remaining regions as control groups.
Click Save Configuration.
The new test appears in the Test Configurations screen.
(Optional) Under the Actions column, click Preview Test to review the test you created.
Preview test configuration
The Preview Test option allows you to review a saved test configuration before proceeding with analysis. This screen summarizes test setup details and displays visualizations of metric trends for the test and control groups. Use this preview to verify that the configuration is correct and that data is available for all selected regions. The Test Configuration Preview screen has three sections.
Test Configuration
This section gives you the configuration summary of the test you created. You can also start analyzing the test by clicking Analyse Test.
Test Configuration Preview
This section helps you visualize time-series data for the selected metric for both the test and control groups in a chart.
This chart helps you confirm:
Both groups have stable data
Data exists for the entire test period
Distinction between test and control trends
All states metric over time chart
This chart shows the metric trend line for each region, color-coded individually.
Use this view to:
Identify irregularities in individual regions.
Check that no region is contributing disproportionately.
Verify that all regions are consistently represented.