Causal Inference

Also known as: Causal Analysis, Counterfactual Analysis, Econometrics, Causal AI

Statistical methodology for establishing cause-and-effect relationships between variables, beyond correlation.

Causal Inference is the set of statistical and methodological methods designed to establish cause-and-effect relationships between variables—not just correlations—from observational or experimental data.

The fundamental distinction: correlation ('A and B occur together') vs. causation ('A causes B'). Without adequate causal inference methods, spurious correlations can be confused with real causal relationships, leading to incorrect business decisions.

Main methods: Randomized experiments (the gold standard), Difference-in-Differences (DiD), Regression Discontinuity, Instrumental Variables, Propensity Score Matching, and Causal Graphs (DAGs). MMM and MTA also have causal inference components.

In marketing, causal inference is critical for evaluating the real impact of investments: what would have happened without the campaign? (the counterfactual). Without causal methods, marketing results are systematically overestimated.