Marketing Mix Modeling

Overview

Marketing Mix Modeling (MMM) measures the true incremental impact of every marketing channel using Bayesian statistics. Unlike click-based attribution, MMM considers both online and offline channels — including seasonality, weather, trends, and macroeconomic factors — to provide a complete picture of how your marketing mix drives revenue.

Key Features

• Revenue Decomposition: Waterfall chart showing how total revenue breaks down into base sales, seasonality, and each channel's contribution
• Historical ROI vs. Marginal ROI: Understand why a channel can be efficient overall but at its scaling limit
• Saturation Curves: Hill function curves showing diminishing returns for each channel, with your current operating point marked
• Budget Optimizer: Interactive sliders to simulate "what if" budget reallocation scenarios in real time
• Model Accuracy: R² and MAPE metrics to quantify model reliability
• Bayesian Time Series: 95% and 80% credible interval bands on revenue predictions
• Historical Run Comparison: Compare model results across different time periods
• 12-Week Forecast: Revenue forecast with confidence intervals

How It Works

MMM runs a Bayesian regression model (PyMC Marketing) on your aggregated weekly data. It separates organic demand from marketing-driven demand, then estimates each channel's contribution using logistic saturation functions. The model accounts for adstock (carry-over effects), seasonality, weather, and external factors.

Key Dashboard Sections

Contribution & Decomposition

The waterfall chart shows a step-by-step breakdown from organic base revenue through seasonality effects to each paid channel's contribution.

Efficiency & Marginal ROI

Side-by-side comparison of historical ROI (total return over total spend) with marginal ROI (what the next euro would generate). A channel with high historical ROI but low mROI is near saturation.

Saturation & Diminishing Returns

Each channel has its own S-curve showing the relationship between spend and response. The current spend level is marked as the operating point.

Budget Optimizer

Interactive sliders for each channel let you simulate budget reallocation. The summary shows total spend change, projected revenue delta, and the new Marketing Efficiency Ratio.

Use Cases

• Budget Planning: Make data-driven decisions about how to allocate budget across channels
• Channel Evaluation: Understand the true incremental value of each channel beyond click-based metrics
• Diminishing Returns: Identify channels that have hit saturation and where spend should be redirected
• Unified Measurement: Combine MMM with attribution and incrementality testing for triangulated insights

Getting Started

1. Connect your ad platforms and install the tracking pixel
2. Ensure at least 12 weeks of historical data
3. Contact your account manager to activate MMM for your workspace
4. Review model results and use the budget optimizer for planning