Media Mix Attribution Calibration Plan

Role: Lead Econometrician & Media Attribution Modeler

Context

• Entertainment Studio: {{entertainment_studio_name}}
• Release Campaign Type: {{release_title_type}}
• Evaluated Media Channels: {{campaign_media_channels}}
• Historical Calibration Window: {{historical_lookback_window}}
• Prior Adstock Assumptions: {{adstock_decay_assumptions}}
• Core Performance Metric: {{target_performance_metric}}

Task

Author a Bayesian marketing mix modeling (MMM) calibration and validation plan for {{entertainment_studio_name}} to estimate channel elasticities and optimize budget allocation toward {{target_performance_metric}} for {{release_title_type}} releases.

Method

1. Specify the hierarchical Bayesian regression model structure incorporating seasonal trends, box-office/broadcast momentum, and baseline release buzz.
2. Translate {{adstock_decay_assumptions}} into parametric Weibull or geometric adstock decay and Hill saturation functions for each channel in {{campaign_media_channels}}.
3. Formulate informative prior distributions for channel coefficients using historical spend data across {{historical_lookback_window}}.
4. Design Markov Chain Monte Carlo (MCMC) sampling diagnostic protocols, establishing target $\hat{R}$ thresholds and effective sample size targets.
5. Detail the calibration strategy comparing model-estimated incremental lift with empirical randomized geo-experiment results.
6. Establish decomposition procedures to isolate organic release awareness from paid media spend contributions.
7. Formulate out-of-sample cross-validation benchmarks and mean absolute scaled error (MASE) performance criteria.

Constraints

• MUST define explicit prior parameterizations (e.g., Half-Normal, Gamma) for channel return elasticities.
• MUST validate model outputs against at least one empirical calibration or holdout dataset.
• MUST NOT treat digital and broadcast channels with identical lag and carryover decay assumptions.
• Exclude general marketing commentary; restrict focus to econometric specifications and statistical validation.

Output format

Deliver an econometric modeling plan with the following sections:

1. Model Specification & Transformations (functional forms of adstock and saturation)
2. Bayesian Prior Specification & Elicitation Protocol
3. MCMC Diagnostics & Calibration Framework
4. Validation Strategy & Goodness-of-Fit Targets (under 550 words)

Self-review

• Confirm that adstock parameter bounds match {{adstock_decay_assumptions}}.
• Verify all channels listed in {{campaign_media_channels}} have designated transformation treatments.
• Check that convergence criteria ($\hat{R} < 1.05$) are explicitly articulated.