What's Working?: The Marketing Mix Modeling Playbook: Build, Validate, and Optimize Marketing Mix Models in Robyn, Meridian, and PyMC — Without Tracking a Single User.

Attribution is broken. Here's what actually works—and now it lives on your desk.When Apple's iOS 14.5 and the death of third-party cookies shattered digital attribution, marketers lost the ability to see what their spending actually did. Click tracking went dark, reported ROAS stopped matching real revenue, and budget meetings became guesswork in a spreadsheet costume.Marketing Mix Modeling (MMM)—also known as media mix modeling—is the privacy-proof, board-ready answer. Built on aggregate, first-party data, MMM measures what truly drives revenue across every channel—TV, paid search, paid social, email, and out-of-home—without tracking a single user, capturing both immediate and long-term impact.What's Working? The Marketing Mix Modeling Playbook is the comprehensive, keep-it-on-your-desk guide you'll annotate, dog-ear, and reach for again and again. Follow analyst Vera Goodwin and her mentor Clara Vance from broken dashboards to a marketing measurement system the CFO trusts—then run the same playbook on your own data. Story and substance: every formula, diagnostic, and line of code, in plain English. Most MMM resources are academic papers or vendor demos; this is the practitioner's end-to-end reference."Experiments tell you what happened once. A good model tells you what to do next."Across 20 chapters and a deep appendix set, this hands-on reference shows you how to:Build a credible marketing mix model from scratch—adstock, saturation, and diminishing returnsMeasure marketing attribution and incrementality without cookies or user-level trackingValidate models with holdout testing and Decomp.RSSD so every number holds upRun budget optimization, media planning, and scenario planning to move spend toward what worksProve marketing ROI and return on ad spend (ROAS) to your CMO and CFOCalibrate with experiments and report the marketing metrics and KPIs leaders trustImplement MMM in Robyn, Google Meridian, and PyMC-Marketing—plus running MMM conversationally with AIFrom adstock and Bayesian regression to media planning and executive communication, this is end-to-end marketing analytics and marketing science—rigorous when you need it, readable throughout. You'll also weigh build vs. buy across open-source, SaaS, and consulting, and adapt MMM for B2B, lead-based, higher-education, and nonprofit models.Who it's for: analysts and data scientists building real models; CMOs, VPs, and marketing managers defending the budget; CFOs demanding accountability; performance and growth marketers living beyond GA4; and consultants and students entering the field.Written by Joshua Dodson, Vice President of Analytics & Performance Marketing at BVK—two decades turning messy data into decisions executives act on.Stop guessing. Start proving what's working. Scroll up and add What's Working? to your cart. Read more