Glossary · Marketing Engineering
CausalImpact (Bayesian Structural Time Series)
also: BSTS · Bayesian structural time series · synthetic control
Definition
CausalImpact is Google's open-source implementation of Bayesian structural time series for measuring the causal effect of a known intervention on a time-series outcome. It constructs a counterfactual from correlated control series and reports the posterior distribution of the cumulative effect with calibrated uncertainty.
Introduced by Brodersen et al. (2015), CausalImpact fits a state-space model to the pre-intervention period using correlated control variables, then projects the counterfactual forward into the post-period. The difference between observed and counterfactual yields the causal effect with full posterior distribution — not a point estimate. Widely used for measuring the lift of SEO content programs, TV campaigns, and feature launches where randomized experiments are infeasible.
Essays on this concept
- Business Analytics
Causal Discovery in Business Data: Applying PC Algorithm and FCI to Find Revenue Drivers Without Experiments
Correlation tells you that feature usage and retention move together. It doesn't tell you which causes which — or whether a third factor drives both. Causal discovery algorithms can untangle this from observational data alone.
- Marketing Engineering
Causal Impact of SEO on Branded Search: A Synthetic Control Method for Organic Channel Measurement
SEO is the only major marketing channel where practitioners still argue about whether measurement is even possible. Synthetic control methods borrowed from policy economics prove it is — and the results will surprise you.
- Marketing Engineering
Incrementality Testing at Scale: A Geo-Lift Framework for Measuring True Campaign Impact
Half your marketing budget is wasted. The classic joke, except now we can identify which half — geo-lift experiments measure what would have happened without the campaign, not just what happened with it.
- Marketing Engineering
Marketing Mix Modeling in the Privacy-First Era: Bayesian Structural Time Series Without User-Level Data
Cookies are dying. Deterministic attribution is shrinking. The irony: the measurement approach from the 1960s — Marketing Mix Modeling — is making a comeback, now powered by Bayesian inference that would have been computationally impossible when it was first invented.
- Marketing Engineering
Unified Measurement Architecture: Connecting MMM, MTA, and Experimentation Into a Single Source of Truth
MMM says Facebook works. MTA says Google works. The incrementality test says neither works as well as you thought. Three measurement systems, three different answers — here's how to reconcile them into one coherent picture.
Related concepts
Authoritative references