Home » Insights » What is adstock in marketing mix modeling (MMM)?

Data & Analytics

# What is adstock in marketing mix modeling?

Adstock is a concept that is commonly used in marketing mix modeling (MMM) to account for the carryover effects of advertising

Adstock refers to the amount of residual impact that an advertisement has on consumer behavior even after the advertisement has stopped running. It is important to understand adstock when building MMM models. This is because adstock in marketing mix modeling can help marketers accurately estimate the long-term impact of their advertising efforts.

## 1. Why is Adstock Important in MMM Models?

Marketing mix modeling involves analyzing the impact of various marketing activities on consumer behavior, such as sales or website traffic. Adstock is important in MMM models because it helps to account for the long-term impact of advertising. For example, an advertisement for a product may continue to influence consumer behavior even after the ad has stopped running, due to residual brand awareness or other factors. By including adstock in MMM models, marketers can more accurately estimate the long-term impact of their advertising efforts and optimize their marketing mix accordingly.

## 2. Example of an Adstock: Calculating using Weibull PDF, Weibull CDF, and Geometric Functions:

### 1. Weibull PDF:

One way to calculate adstock is by using the Weibull probability density function (PDF). The Weibull PDF is a commonly used distribution function in reliability engineering, and it can also be applied to adstock modeling. The formula for the Weibull PDF is:

f(t) = (k/λ) * (t/λ)^(k-1) * e^(-(t/λ)^k)

Where:

– f(t) is the probability density function of the Weibull distribution at time t

– k is the shape parameter, which controls the shape of the distribution

– λ is the scale parameter, which controls the location of the distribution

Pros:

– The Weibull PDF is a flexible and versatile distribution function that can be used to model a wide range of phenomena, including adstock.

– It provides a detailed description of how the probability of residual impact changes over time, which can be useful for understanding the dynamics of adstock.

Cons:

– The Weibull PDF requires fitting two parameters (k and λ) to data, which can be computationally intensive and may require a large amount of data to obtain accurate estimates.

– The Weibull PDF assumes that the probability of residual impact decreases monotonically over time, which may not always be the case in practice.

### 2. Weibull CDF:

Another way to calculate adstock is by using the Weibull cumulative distribution function (CDF). The Weibull CDF describes the probability that residual impact will occur before or at a given time t. The formula for the Weibull CDF is:

F(t) = 1 – e^(-(t/λ)^k)

Pros:

– The Weibull CDF provides a simple and intuitive way to model adstock, as it describes the probability that residual impact will occur within a given time frame.

– It is relatively easy to fit the Weibull CDF to data, as it only requires estimating the two parameters (k and λ).

Cons:

– The Weibull CDF assumes that the probability of residual impact decreases monotonically over time, which may not always be the case in practice.

– It provides less information about the dynamics of adstock compared to the Weibull PDF.

### 3. Geometric Function

The geometric function is a simple and commonly used way to model adstock. The geometric function assumes that the probability of residual impact decreases exponentially over time. The only parameter that needs to be estimated for the geometric function is λ.

The formula for the geometric function is:

f(t) = exp(-λt)

Pros:

• The geometric function is straightforward to estimate and interpret, requiring only a single parameter (λ).
• It assumes that the probability of residual impact decreases exponentially over time, which is a reasonable assumption in many cases.

Cons:

• The geometric function assumes that the probability of residual impact decreases exponentially over time, which may not always be the case in practice.
• It provides less information about the dynamics of adstock compared to the Weibull functions.

Adstock is an essential concept in MMM that helps marketers accurately estimate the long-term impact of their advertising efforts. We have explored how adstock can be calculated using three different functions.