Experiment measuring the causal impact of a channel or campaign.
Geo-split holdouts or ghost-ad experiments quantify the true uplift of paid media beyond what would have happened organically. Essential before scaling budget.