Causal Lab overview

Some of the most important questions cannot be answered by a simple A/B split — "did this campaign actually drive incremental sales, or would they have happened anyway?" The Causal Lab (Experiment → Causal Lab) brings rigorous causal-inference methods to those questions.

What it is for

• Incrementality — the true causal effect of a campaign, channel or change, net of what would have happened regardless.
• Retrospective analysis — measure the impact of something that already happened, where you cannot run a forward test.
• Geo and time-based questions — where the unit of analysis is a region or a period, not a user.

When to use it instead of an experiment

Use a standard experiment when you can cleanly split traffic. Use the Causal Lab when you cannot — for example measuring a brand campaign, a channel''s incrementality, or the effect of a change you already made.

How it works

You choose a method that fits your situation, define the intervention and the comparison, and the platform estimates the causal effect with a measure of confidence. Results are saved so you can revisit and report them.