Handle challenges in MMM/MMX

You are handed a weekly marketing mix model (MMX) spanning 156 weeks, with inputs for TV GRPs, paid search spend, display spend, email sends, price, promotions, and a competitor index; TV and search have a correlation of 0.90, and a pandemic-driven structural break occurs in week 70. Which factors make this model fragile, and how would you address them? Be specific about endogeneity and omitted variables, remedies for multicollinearity (priors, ridge/LASSO, hierarchical Bayesian), adstock/lag and saturation decisions, non-stationarity and change points, promotion cannibalization, privacy-induced measurement error (such as ATT), and calibration with randomized geo-tests. Describe your validation plan (out-of-time fit, lift alignment, posterior predictive checks) and how you would generate robust ROI and budget recommendations that account for uncertainty.

Overview: This question assesses a data scientist's skill in applied machine learning and causal inference for marketing mix modeling, with emphasis on diagnosing model fragility caused by multicollinearity, endogeneity, non-stationarity, measurement error, and choices about adstock, saturation, and promotion effects.