Overview
Incrementality Testing helps you measure the real, causal impact of your marketing spend, not just correlated performance. By comparing matched test and control markets, Sona shows you how much of your conversions and revenue are actually driven by a given channel, rather than what would have happened anyway. This guide walks you through reading your testing dashboard and launching a new experiment.
Incrementality Testing Overview
When you open the page, you'll see a summary of your current testing activity at the top, followed by your test vs. control markets and a list of your experiments.
KPI Summary
Four cards summarize your testing program at a glance:
Avg. Incremental ROAS: the average return on ad spend attributable specifically to your test markets, with period-over-period change shown below it.
Active Tests: the number of experiments currently running.
Geo-Groups: the number of geographic groups configured across your tests.
Total Spend Tracked: total ad spend covered by active incrementality tests.
Test vs. Control
This section shows which markets are assigned to your Test Markets and which are held out as Control Markets, along with each market's confidence level (High, Medium, or Low).
Toggle between List View and Map View in the top right to see markets as a table or plotted geographically.
Tip: Control markets should closely mirror test markets in size and behavior. Low-confidence matches can skew your incremental lift results.
Experiments
Below the market breakdown, the Experiments section lists every test you've created.
Switch between the Active and Archived tabs to filter by status.
Click Archive to move a list of experiments out of the active view.
Use Filters or the search bar to narrow the list.
Creating a New Experiment
Click + New Experiment in the top right.
On the Basics step, enter:
Name: a label for the experiment.
Channel: the ad platform to test (e.g., Google Ads).
Goal: the conversion event you're optimizing for (e.g., Purchase).
Campaigns: the specific campaigns included in the test.
Hypothesis (optional): your reasoning for running the test.
Click Next to continue through Markets, Parameters, and Review.
On the final Review step, confirm your settings and launch the experiment.
Note: Hypothesis is optional but recommended. It gives your team context when reviewing results later.
FAQs
What's the difference between a Test Market and a Control Market?
What's the difference between a Test Market and a Control Market?
Test markets receive the marketing treatment (e.g., ad spend on a given channel). Control markets don't, and act as the baseline used to calculate what would have happened without that spend.
What does the confidence level (High/Med/Low) on a market mean?
What does the confidence level (High/Med/Low) on a market mean?
It reflects how well a market is matched to its counterpart based on size, behavior, and historical performance. Higher confidence means a more reliable incrementality read.
Can I edit an experiment after launching it?
Can I edit an experiment after launching it?
Once an experiment is active, its markets and campaigns can't be changed. Archive it and create a new experiment if your setup needs to change.
What does "Total Spend Tracked" include?
What does "Total Spend Tracked" include?
The total ad spend across all campaigns currently included in active incrementality experiments.