Interactive
studies.
Built for marketing analysts who need more than dashboards. Each study is an interactive deep dive into the methods behind incrementality, attribution, and forecasting — with real data and real code.
Causal Inference in Marketing Attribution
Last-click is wrong. MTA is biased. This study walks through geo experiments, PSM, and difference-in-differences to measure what your campaigns actually caused — not just correlated with.
Double Machine Learning for Treatment Effect Estimation
When budgets, seasonality, and audience quality are all correlated, linear controls aren't enough. Use ML to partial out high-dimensional confounders and isolate the true effect of your ad spend.
Web Traffic Forecasting with GA4 and BigQuery
Forecast sessions and conversion rates directly from GA4 exports in BigQuery. Compare ARIMA, Prophet, and gradient boosting — and see which holds up when campaigns spike the data.