Media/Marketing Mix Modeling & Experimentation - Specialist/Data Scientist

Role Overview

We run a statistically rigorous media measurement practice built on Bayesian MMM, causal experimentation, and decision science. Were hiring a Specialist/Data Scientist to strengthen our modeling foundation and ensure budget allocation recommendations are defensible, uncertainty aware, and actionable across retail media (CPG),high-growth DTC, and enterprise brands.

Final designation and compensation will be aligned with your specific experience level and performance during our evaluation. Even if you dont fit all the criteria, we encourage you to apply.

We are fully remote. The working hours are 4 AM - 1 PM Eastern Standard Time.

Key Responsibilities

Bayesian MMM (core ownership)

• Build and refine hierarchical Bayesian MMMs across a portfolio of brands/channels.
• Calibrate priors/constraints (ROI priors, intercept bounds, response curves) using diagnostics + business context.
• Implement and evaluate adstock/carryover, saturation, elasticity, and (when supported) interactions.
• Quantify uncertainty (credible ranges) and translate it into budget allocation risk.
• Stress test recommendations with posterior predictive checks and sensitivity to priors.

Incrementality, Calibration & Scenario Support

• Design and evaluate A/B and geo incrementality tests with rigorous power/MDE planning.
• Translate lift into incremental revenue, CPA, and marginal efficiency; communicate error tradeoffs clearly.
• Use experiment results to validate/calibrate MMM outputs and improve external validity.
• Support budget optimization and forecast scenarios with transparent assumptions and uncertainty framing (expected value + plausible range).

Technical Standards & Mentorship

• Build in Python (e.g., PyMC/Bambi/statsmodels/scikit-learn) and productionize in a vendor MMM platform.
• Own MMM QA standards (convergence/PPCs, identifiability, spend correlation/collinearity, sensitivity analyses).
• Mentor analysts on Bayesian interpretation, uncertainty communication, and QA discipline.

Required Skills & Qualifications

Technical Skills

• Strong Python and SQL, plus comfort wrangling multi-source marketing data.
• MMM experience (build, audit, or operate) and ability to turn outputs into decisions.
• Statistical depth: uncertainty quantification, model validation, and experiment literacy (power/MDE, error tradeoffs).
• Clear communication with non-technical stakeholders; comfortable defending assumptions.

Nice to have Skills

• Deep Bayesian modeling (PyMC ecosystem), retail media/performance marketing measurement.
• Geo-experiment design & execution in practice (market selection, spillover/interference, and power under real-world constraints).

What Success Looks Like in 60 Days

• Deliver at least one MMM build or meaningful model improvement that informs a real budget decision with uncertainty-aware ROI guidance.
• Publish and apply a repeatable MMM QA checklist (Convergence, identifiability, collinearity, sensitivity to priors).
• Produce an incrementality calibration plan (power/MDE standards + readout template) aligned with upcoming test(s).
• Coach analysts on Bayesian interpretation and uncertainty communication so outputs are consistent and defensible.

Why This Role Matters

This is not a reporting role. Youll influence how millions of dollars are invested by grounding decisions in rigorous modeling, calibrated incrementality evidence, and honest uncertainty.