Adstock & Saturation: Impact on Budget Planning

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

In our last lesson, we decoded the primary outputs of an MMM, learning how to interpret contribution charts and response curves. We briefly touched on the idea that an MMM doesn't just use raw spend; it first transforms the data to reflect how marketing actually works in the real world. You learned that these transformations are called adstock and saturation.

Today, we will dive deep into these two fundamental concepts. Our goal is to explain exactly what adstock and saturation are, how they are modeled, and most importantly, how a firm grasp of these concepts is critical for making strategic budget planning decisions. This lesson will equip you to understand the "physics" behind the response curves and use that knowledge to guide your team's spending strategy.

1. Adstock: Modeling the Lingering Effect of Ads

Let's start with a simple truth: the impact of an advertisement doesn't vanish the moment the ad finishes airing or the impression is served. A memorable TV commercial or a compelling social media campaign can influence a customer's decision days or even weeks later. This lingering impact is known as the carryover effect, and in MMM, we model it using a technique called adstock.

The purpose of adstock is to transform a single point-in-time spend into a continuous effect that decays over time.

To see how this is calculated, let's revisit the PyMC-Labs guide we consulted in the previous lesson.

The key parameter in the adstock formula is the decay rate, often represented by lambda ($\lambda$) or alpha ($\alpha$). This single number determines how "sticky" a channel's advertising is.

• A high decay rate (e.g., $\lambda =0.9$) means the effect lingers for a long time. This is typical for channels like TV or out-of-home, which build brand memory.
• A low decay rate (e.g., $\lambda =0.2$) means the effect fades quickly. This is common for direct-response channels like Branded Search, where the impact is immediate.

The image below visualizes this concept.

The Strategic Impact of Adstock on Budgeting

Understanding a channel's adstock is crucial for budget planning. It helps you answer questions about campaign frequency and flighting. Let's watch a short clip that explains the strategic implications.

As the speaker highlights, a channel with a long decay offers an opportunity for efficiency. You might not need to spend heavily every single week. Instead, you could "pulse" your spending, letting the carryover effect bridge the gaps, thereby "squeezing more efficiency" out of your budget. Conversely, for a channel with a very short decay, a consistent, "always-on" presence is necessary to maintain its impact.

2. Saturation: Modeling Diminishing Returns

The second critical concept is saturation, which is the formal term for diminishing returns. You know this intuitively: the first $10,000 you spend on a new campaign reaches the most receptive audience and yields a strong return. The next $10,000 is still effective, but slightly less so. As you keep spending, you start reaching people who are less interested or have already seen your ad multiple times, and each additional dollar brings back less and less.

To capture this, MMM applies a non-linear saturation function to the adstocked spend. This creates the "S-shaped" response curve we discussed in the last lesson.

Let's look at a clear written explanation of this.

Just like adstock, saturation curves are controlled by parameters. These parameters define how quickly the curve rises and where it flattens out.

The Strategic Impact of Saturation on Budgeting

Saturation is the single most important concept for budget allocation. The saturation point on a response curve is a clear signal that a channel is becoming inefficient. It tells you where to stop spending.

This next video segment provides a perfect visual guide for how to interpret saturation curves from a strategic standpoint.

As a marketing leader, your job is to use these curves to guide investment:

• Low Saturation Channel: This is a growth opportunity. You can confidently increase the budget here because you are still on the steep, efficient part of the curve. Your marginal ROI is high.
• High Saturation Channel: This is a signal for optimization. While the channel might have a good average ROI, its marginal ROI is low. Continuing to pour money here is wasteful. The correct move is to cap spending at the "knee" of the curve and reallocate the excess budget to less saturated channels.

The documentation for Google's own MMM tool, Meridian, reinforces this exact point.

3. Synthesis: From Theory to Actionable Budgets

We've now examined adstock and saturation separately. In practice, they work together. An MMM estimates the adstock decay rate and the saturation curve parameters for every channel simultaneously. These parameters are then used to build the final response curves and drive an optimization engine that recommends a new budget allocation.

The case study of how Bolt, the ride-sharing company, uses these principles is a powerful real-world example.

Bolt's story encapsulates the strategic value of this lesson. By understanding the underlying physics of their marketing channels, they moved beyond simple ROI measurement to a dynamic, data-driven budgeting process that could adapt to different business goals. This is the level of strategic oversight you are aiming for.

Conclusion

Today we've unpacked the two most important transformations inside a Marketing Mix Model. They are not just technical details for data scientists; they are the conceptual engine for strategic budget optimization.

Key Takeaways:

• Adstock (Carryover Effect) models the delayed impact of advertising. Understanding a channel's adstock helps you plan campaign frequency and timing to maximize efficiency.
• Saturation (Diminishing Returns) models the "S-shaped" response of marketing spend. It is the key to identifying the point of inefficiency and knowing when to stop investing in a channel.
• Impact on Budgeting: Together, adstock and saturation provide a sophisticated framework for budget allocation. They allow you to decide not only how much to spend (from saturation curves) but also how often to spend it (from adstock decay).

Preview of the Next Lesson:
Now that you have a firm grasp of contribution, response curves, adstock, and saturation, we are ready to put it all into practice. In our next lesson, we will focus on how to use MMM-derived response curves to inform optimal budget allocation across channels, moving from interpretation to active scenario planning.