Interpreting Key MMM Outputs

Hello! Welcome back to our module on Marketing Mix Modeling (MMM).

In our last lesson, we established the strategic purpose of MMM. We saw that it's a powerful tool for answering critical business questions about what drove past sales, how efficient our channels are (iROAS), and how to optimize our budget for the future.

Today, we're moving from the why to the what. Our goal is to equip you to confidently interpret the key deliverables of an MMM project. When your analytics team or a vendor presents you with a deck full of charts and numbers, you need to know exactly what you're looking at and what it means for your business.

We will focus on three core outputs that you will almost always encounter:

1. Contribution Charts, which tell the historical story of your sales.
2. Response Curves, which provide the forward-looking insights needed for optimization.
3. Channel Coefficients, the underlying model parameters that quantify each channel's effectiveness.

By the end of this lesson, you'll be able to look at these outputs and translate them directly into strategic insights.

1. Contribution Charts: The Historical Story

Let's start with the most intuitive output: the contribution chart (also known as a sales decomposition chart). This visual directly answers the question, "What drove my sales over the past year?"

It deconstructs your total sales over time into its constituent parts:

• Base Sales: The foundation of your revenue that you'd likely have even with zero marketing spend. This is driven by brand equity, distribution, and non-promotional demand.
• Incremental Sales from Marketing: The specific lift generated by each marketing channel (e.g., TV, Meta Ads, Google Search).
• Incremental Sales from Other Factors: The impact of promotions, pricing changes, seasonality, and other control variables included in the model.

Here is a typical example of what these charts look like:

To dive a bit deeper into the primary outputs of MMM, let's turn to a helpful article.

As you can see, the contribution chart is your historical report card. For a marketing leader, its value is in:

• Quantifying Marketing's Value: It provides a clear, defensible number for the total revenue driven by the marketing team's efforts.
• Understanding Drivers: It shows which channels have been the biggest historical growth drivers.
• Spotting Trends: By looking at the chart over time, you can see if the impact of certain channels is growing or shrinking.

2. Response Curves: The Key to Future Optimization

While contribution charts look backward, response curves look forward. They are arguably the most important output for strategic decision-making and budget allocation.

A response curve for a given channel visualizes the relationship between how much you spend and the incremental sales you can expect to get in return. Crucially, these curves are not straight lines; they flatten out, showing the effect of diminishing returns or saturation.

Let's return to the Sellforte article to get a more detailed explanation of why these curves are so powerful.

As the article highlights, response curves are the engine for optimization. They allow you to move beyond average ROI and embrace the concept of marginal ROI that we discussed in the previous module.

When you're presented with these curves, you should be asking:

• Where are we on the curve? Are our key channels operating on the steep, efficient part of the curve, or are they nearing the flat, saturated part?
• Where is the point of diminishing returns? At what spend level does the miROAS drop below an acceptable threshold?
• Where is the opportunity? Are there channels with low current spend but a steep initial curve, suggesting they are untapped opportunities for growth?

3. Channel Coefficients & Model Components: Under the Hood

So, how does the model generate these contribution charts and response curves? The answer lies in the model's underlying parameters, specifically the channel coefficients.

If you recall from our earlier statistics module, a regression model tries to find the relationship between a dependent variable (Sales) and one or more independent variables (like marketing spend). The coefficient ($\beta$) for each variable represents the magnitude of its effect.

Salest​=β0​+β1​⋅Transformed_SpendChannel_1t​+...

• ${\beta }_0$ represents the base sales.
• ${\beta }_1$ represents the coefficient for Channel 1. A higher $\beta$ implies a more effective channel, all else being equal.

However, in a modern MMM, the model doesn't just use raw spend. It first transforms the spend data to account for real-world marketing physics, namely adstock and saturation.

To understand these core components, we'll consult a more detailed guide. Your computer science background will make the logic behind this structure quite intuitive.

As a marketing leader, you won't be asked to calculate the adstock decay rate or choose a saturation function. However, understanding these concepts is crucial for two reasons:

1. Interpreting Results: You now know that a channel's contribution is a function of its raw spend, its carryover effect (adstock), its point of diminishing returns (saturation), and its overall effectiveness (coefficient).
2. Asking Smart Questions: When you're presented with an MMM, you can now ask informed questions like, "What adstock assumptions were used for TV versus digital?" or "How saturated does the model suggest our main channels are?" This shows a deeper level of engagement and helps you critically assess the model's outputs. This leads us to the topic of our next lesson.

4. Tying It All Together: A Case Study Revisited

Let's revisit the case study from the Funnel video we saw in the last lesson. Now, with your understanding of the key outputs, you can see how the analyst moves from the model's results to a strategic recommendation.

This case study perfectly demonstrates your role as a strategic leader. You take the core outputs—contribution, ROI (efficiency), and response curves (diminishing returns)—and synthesize them to create an actionable plan that optimizes budget and drives business growth.

Conclusion

Today we've demystified the primary outputs of a Marketing Mix Model. You are now equipped to read the story they tell about your marketing performance.

Key Takeaways:

• Contribution Charts decompose historical sales, showing you what has driven your business and quantifying marketing's total impact.
• Response Curves are your forward-looking tool. They visualize channel saturation and are essential for making marginal investment decisions to optimize your budget. The slope of the curve is your marginal ROI.
• Channel Coefficients, combined with adstock and saturation transformations, are the "under the hood" components that determine a channel's impact within the model.

Preview of the Next Lesson:
We can now read the outputs of an MMM. But how do we know if the model itself is any good? A model built on flawed assumptions or bad data will give you beautifully charted, but dangerously misleading, advice. In our next lesson, we will address this by learning how to critically assess the methodology and potential limitations of a proposed or completed MMM.