Openlift

Open-Source Marketing Incrementality

Open Source

Analytics

Marketing

GitHub

Visit WebsiteSee on Product HuntGithub ⧉Twitter ⧉

Marketing measurement is undergoing a structural failure, not a temporary disruption. The identity and user-level tracking infrastructure that powered a decade of performance marketing is being dismantled by privacy regulation, platform restrictions, signal loss and fragmented customer journeys. More importantly, much of that infrastructure was never causally reliable. Attribution models measured correlation within observable conversion paths, then presented it as incremental impact. OpenLift replaces attribution-based inference with a statistically rigorous causal measurement system. It uses Bayesian Synthetic Control to estimate the counterfactual outcome that would have occurred without an intervention, while preserving full posterior uncertainty rather than collapsing results into a misleading point estimate. Dynamic Time Warping identifies structurally comparable geographies based on pre-intervention behaviour, while simulation-based power analysis determines whether an experiment has sufficient sensitivity to detect commercially meaningful lift before capital is committed. The output is engineered for decision-making, not statistical theatre. A five-tier evidence framework converts posterior distributions, credible intervals and probability-of-lift estimates into explicit business actions: reject, hold, continue learning, validate or scale. An AI translation layer then converts complex causal outputs into operational recommendations, removing the requirement for every decision-maker to be an applied statistician. Because OpenLift operates at the geographic level, it does not depend on cookies, device identifiers or deterministic user matching. The methodology remains applicable across markets, channels and privacy environments. It is open source under the MIT Licence and designed for production deployment. When platform-reported ROAS can overstate true incremental return by three to five times, causal measurement is not an analytical luxury. It is the control system that separates businesses scaling genuine demand creation from those repeatedly funding conversions that would have happened anyway.