Sellforte Experiments: Overview
Measure the causal impact of specific marketing activities and feed the results back into MMM.
Sellforte Experiments helps you measure the causal impact of specific marketing activities. While Marketing Mix Modeling answers the broad question — what is the overall ROI of our media mix? — Experiments answers a narrower and more actionable one: did this specific campaign, promotion, or test actually drive incremental results?
Experiment results are also designed to feed back into MMM calibration, creating a stronger and more evidence-based measurement foundation over time.

Three experiment types
Sellforte Experiments currently supports three types of experiment. Each is designed for a different situation.
Conversion Lift
For importing and visualising the results of a conversion lift study run inside Meta. Meta conducts the experiment; Sellforte structures, displays, and summarises the output.
Use this when you have already run a Meta conversion lift study and want to review and store the result in Sellforte.
If Meta is connected through Sellforte's native Meta connector, your studies can be imported automatically instead. See Sellforte Experiments: Conversion Lift.

Geo Lift
For measuring the incremental impact of a campaign or spend change using geographic groups already in Sellforte. No data upload is required. Sellforte builds a synthetic control from untreated regions and estimates the causal effect of the treatment.
Use this when a campaign ran in some regions but not others, and the relevant data is already in Sellforte.
A/B Test
For measuring incremental impact using data you upload directly. Uses the same methodology as Geo Lift, but accepts any grouping — customer segments, product categories, store tiers, or geographies — as long as the treatment is clearly isolated.
Use this when the relevant data is not in Sellforte, or when the experiment does not fit a standard geographic split.
Choosing the right experiment type
A simple way to decide:
- You ran a Meta conversion lift study and have the export → Conversion Lift
- Your data is in Sellforte and the split is geographic → Geo Lift
- Your data is outside Sellforte or the split is non-geographic → A/B Test
A consistent workflow across all types
Regardless of experiment type, Sellforte presents results in a consistent structure: headline metrics, visual charts, an AI-generated summary, and a comments section for capturing conclusions. All experiments are stored in the Experiment Library, making it easy to review, compare, and build on previous work.


Choosing what to test next
Test Prioritizer ranks testing opportunities across your media mix with a priority score based on Impact, Confidence and Ease, and shows how much of your media spend has recent test evidence. See Test Prioritizer: Overview.
How Experiments supports MMM
Experiments is a complement to MMM, not a replacement for it. MMM gives a broad view of marketing performance across channels. Experiments gives a focused read on whether a specific intervention caused an incremental outcome.
High-quality experiment results can be used to calibrate MMM with real-world evidence, improving the reliability of model outputs over time. Not every experiment is suitable for calibration — the quality of the design and the strength of the result both matter.
Further reading
Detailed guidance for each part of the Experiments toolset is available in the following articles:
- Sellforte Experiments: Conversion Lift — what it is, when to use it, and how to set it up
- Sellforte Experiments: Geo Lift — what it is, when to use it, and how to set it up
- Sellforte Experiments: A/B Test — what it is, when to use it, and how to set it up
- Sellforte Experiments: Methodology — how Geo Lift and A/B Test estimate incremental impact
- Sellforte Experiments: How to Read the Results Dashboard — how to interpret every metric, chart, and section
- Test Prioritizer: Overview — how to choose which test to run next
- Calibrating your MMM with experiment results — how test results are used as evidence in your MMM