Media Mix Modeling for Advanced Marketers

How is Incrementality Testing Redefining MMM?

Marketers around the world will spend more than $1 trillion on paid media this year — deployed with the singular objective of driving important business outcomes. But do marketers know how much of that spend is actually delivering on their desired goal?

The ability to measure the true business impact of every marketing dollar across the entire media portfolio is essential for today’s data-driven marketers. These insights not only allow marketers to prove the value of their ad spend, but also to understand how different marketing programs interact with one another. 

Without this capability, it’s impossible to confidently reallocate less effective ad dollars to higher ROI channels and construct a more efficient, revenue-driving marketing funnel.

Most measurement solutions founded in the past 15 years built their offerings in the pre-privacy era based on a user-level attribution methodology. But since the fallout of click-based attribution systems in recent years, Media Mix Modeling (MMM) has seen a notable resurgence.

While MMM is a powerful method for measuring media performance, it has significant limitations. MMM is correlation based, not causal. It’s expensive, it relies heavily on historical data, it provides only a relatively big-picture view — and MMM isn’t fast enough or granular enough for daily optimization across today’s extremely complex channel ecosystem. As such, MMM is not a complete or reliable portfolio optimization solution.

Instead, today’s leading marketers are taking a more progressive, modernized approach to MMM by combining the powerful attributes of MMM with incrementality testing for significantly more accurate, timely, and actionable insights into their marketing performance.

By triangulating MMM with incrementality tests for causality and platform attribution reporting for granularity, marketers have a more advanced way of ensuring that their brand is capturing the full impact of their entire media mix based on causal ground truth experimentation.

You can also download our free, ungated Guide for a deeper dive into how calibrating MMM with incrementality testing and platform data not only solves the modern measurement conundrum, it empowers forward-thinking marketers to drive substantially better business performance, improve sales, and achieve marketing measurement success. 

The Origins of Media Mix Modeling

Media Mix Modeling (MMM) has been a cornerstone of marketing measurement for more than 40 years, helping marketers evaluate the performance of their campaigns across various channels.

A long-trusted tool for marketers seeking to understand the effectiveness of their marketing efforts, MMM has evolved to remain relevant despite the rapid changes in media formats, advertising channels, and consumer behaviors. 

Media Mix Modeling has been used in some form or another by marketers since the 1950s. Major consumer packaged goods (CPG) companies such as Procter & Gamble and Unilever were pioneers in using statistical methods to understand the effectiveness of their TV advertising spend. With the advent of computer systems in the 1980s, MMM became a more widely adopted practice.

At its core, MMM is a data-driven analytical process used to evaluate the performance of marketing campaigns. It helps marketers uncover insights into how different media channels contribute to business objectives, such as sales, brand awareness, or website traffic. 

By analyzing historical data, MMM can identify patterns and trends, offering valuable insights into which marketing activities are most effective in driving desired outcomes.

MMM's strength lies explicitly in its ability to provide a holistic view of marketing performance by integrating various data points across multiple channels. This flexibility is also its greatest achilles heel, as these models rely on a large amount of human input and adjustment, which is subject to confirmation bias.

Despite its enduring history, MMM relies heavily on a correlation-based approach, making it difficult to differentiate between mere correlation and actual causation. This limitation can lead to inaccurate conclusions about which marketing activities truly drive sales. 

Essential Terminology

Understanding the terminology and key concepts discussed in this Guide is essential to grasping the nuances of MMM and incrementality testing. These are a few critical terms and definitions you should know:

• • Media Mix Modeling (MMM): A statistical regression-based analysis technique used to evaluate the impact various marketing activities have on business outcomes such as sales or brand awareness. MMM analyzes historical data across different media channels to determine how each contributes to overall performance.
  • Incrementality Testing: A methodology that isolates and measures the true impact of a specific marketing activity by comparing the outcomes of exposed groups (those who saw the marketing) with control groups (those who did not). This helps distinguish the actual effect of marketing from what would have occurred naturally.
  • Correlation vs. Causation: Correlation refers to a statistical relationship between two variables, where changes in one variable are associated with changes in another. However, correlation does not imply causation, meaning one variable does not necessarily cause the other to change. Causation, on the other hand, directly links one variable's change as the reason for another's change, which is the desired outcome in accurate marketing measurement.
• Priors: Using priors is a modeling technique where, in the case of MMM, industry-established marketing insights can be included in the model as “prior” knowledge. Prior is a technical term that refers to what and how additional information can be input into a model. The usage of priors in MMM is typically based on how plausible experts think an outcome is, relative to a model where this input is not provided.

Looking Under the Hood at Media Mix Modeling

Traditional MMM typically follows a structured four-step process, which includes:

1. Data Collection: The foundation of MMM is the collection of extensive historical data. This data spans various marketing activities and external factors, such as economic indicators, competitor actions, and seasonal trends. By compiling a comprehensive dataset, marketers can ensure that the model accurately reflects the diverse influences on business outcomes.
2. Model Building: With the data in hand, the next step is to construct a mathematical model that represents the relationships between marketing efforts and sales outcomes. Techniques like multiple linear regression are commonly used to determine how different marketing channels contribute to overall performance. The model is built to account for the interplay between various channels and external factors, providing a nuanced view of how different elements interact to drive business results.
3. Insight Generation: Once the model is built, it is used to generate insights into the effectiveness of different marketing activities. By examining the relationships identified in the model, marketers can understand which channels and tactics are driving sales, brand awareness, or other key metrics. These insights are critical for making informed decisions about future marketing strategies and budget allocations.
4. Optimization: The final step in the MMM process is to use the insights generated to optimize the marketing mix. Marketers can explore "what if" scenarios to test different strategies and predict their potential impact on business outcomes. By allowing an optimization model to produce an optimal mix, you can compare various user-driven media mixes to mathematically optimal mixes based on response functions and adjustment factors determined throughout the MMM process.

Through this structured approach, MMM provides a systematic way to understand and optimize the relationship between marketing efforts and business outcomes, making it an invaluable tool for marketers.

Drawbacks of Traditional Media Mix Modeling

While MMM has proven to be a powerful tool, it is not without its challenges, particularly in the context of today’s complex and rapidly evolving marketing landscape. Understanding these limitations is crucial for marketers who want to fully leverage MMM's capabilities.

• Correlation-Based Approach: One of the most significant challenges of traditional MMM is its reliance on correlation-based data. While correlation can indicate relationships between variables, it does not establish causation. In the context of marketing, this means that traditional MMM might identify a relationship between a particular marketing activity and an increase in sales, but it cannot definitively prove that the activity caused the sales increase. This limitation can result in misleading conclusions, where marketers might attribute success to the wrong channels or tactics.
• High Cost: Implementing traditional MMM can be prohibitively expensive, especially for large enterprises. The process often requires extensive historical data collection and integration, which can be both time-consuming, costly, and not necessarily repeatable with the fidelity required for accurate modeling. For enterprise-sized brands, the costs can easily exceed $1 million, making it a significant investment that may not always yield proportionate returns.
• Inaccurate Models: Without the incorporation of incrementality testing, traditional MMM models are often built on assumptions rather than grounded in real-world data. These models may be calibrated by data scientists who lack a deep understanding of marketing or business operations, leading to potential inaccuracies. Furthermore, traditional MMM models often rely on priors from other models, creating a cycle of assumptions that don’t always reflect the true impact of marketing activities.
• Long Time to Deploy: Building a traditional MMM from scratch is a time-intensive process, often taking over six months before actionable insights are available. This long lead time can be a significant drawback in fast-paced industries where market conditions and consumer behaviors can shift rapidly. The delay in generating insights means that marketers might miss opportunities to make timely adjustments to their strategies.
• Lack of Granularity: Traditional MMM tends to focus on high-level insights, often analyzing performance at the channel level rather than at the campaign level. This lack of granularity makes it difficult to distinguish the performance of closely related campaigns within the same channel, reducing the model's usefulness for making detailed strategic decisions. For marketers who need to understand the impact of individual campaigns, this can be a significant limitation.

In summary, while traditional MMM offers valuable insights, these challenges highlight the need for a more modern approach that can address these limitations and provide more accurate, actionable insights for today’s marketers.

Taking a Modernized Approach to Media Mix Modeling 

The limitations of traditional MMM can be addressed by integrating incrementality testing and leveraging platform attribution data, transforming MMM into a more powerful, comprehensive measurement solution. This modernized approach overcomes traditional challenges in the following ways:

• Establishing Causality: By using incrementality test results as inputs — Bayesian priors — marketers get a mix model anchored in the causal relationship between media and sales rather than correlation, which assumes media exposure prior to a sale was the true driver. This gives marketers performance insights that reflect the true impact of marketing activities, enabling more effective, confident budget allocation changes.
• Cost Efficiency: By leveraging existing testing input data, marketers can streamline the modeling process and quickly begin generating actionable insights. Incorporating incrementality to control for the media-driven impact on sales prevents you from having to include many non-media variables and data sources that can be manual, costly, and that drive delays. Incrementality tests allow you to control those factors in a meaningful way.
• Improved Accuracy Across All Channels: Using incrementality test results to calibrate the MMM (where the results are inserted as Bayesian priors) ensures that the model reflects real-world business performance based on the causal relationship between media and sales. The benefits of this process, however, are not limited to the tested channels. The portfolio of untested media will also be more accurately measured, as the remaining media contribution to sales will be more accurately allocated. This increases the reliability of the insights and allows for more precise optimization of marketing strategies, avoiding the pitfalls of reliance on unvalidated assumptions.
• Timeliness and Granularity: Overlaying MMM data with platform-reported attribution data allows for more frequent updates — potentially daily. This enables marketers to make timely adjustments to campaigns based on the most current data, improving responsiveness to market changes. Additionally, this approach provides detailed insights at the campaign level, allowing for more targeted and effective optimization of marketing efforts.

Triangulated Media Mix Modeling Explained

The Measured Incrementality Model brings the concepts outlined in this FAQ into action by integrating MMM with ongoing incrementality testing and platform attribution data. This unique approach allows marketers to accurately measure and optimize their media spend across all channels with precision, speed, and granularity.

The Measured Incrementality Model provides a comprehensive solution that integrates the best aspects of MMM with ongoing incrementality testing. This model runs weekly MMM updates that are unique to each brand and conversion type and calibrates it with real-world test results on an ongoing basis. This model is then used to generate incremental adjustment factors for every channel, tactic, and campaign in a brand’s portfolio. By doing so, it adjusts each ad platform’s reported KPIs, such as ROAS, to reflect true incremental performance, providing actionable recommendations for reallocating media budgets to maximize ROI.

Transforming Data Into Value

In the rapidly evolving world of marketing, the challenge isn't just measuring performance — it's interpreting the ever-growing volume of complex data and taking decisive action. Marketers often find themselves swamped with metrics, and while measurement is an essential first step, it is what follows that drives true value. This is where Measured stands apart, not only in gathering insightful data but also in enabling actionable optimization. Measured's innovative approach transforms data into value through a seamless integration with three key capabilities:

• Cross-Channel Dashboard: This offers a holistic view of all media channels, delivering transparency into marketing performance at every level. Marketers can use the dashboard to quickly assess the effectiveness of each channel, tactic, and campaign. The dashboard’s visual summaries and detailed reports help pinpoint areas for improvement and spotlight which channels are performing best.
• Media Plan Optimizer: Once performance is assessed, Measured's Media Plan Optimizer takes the insights a step further by providing granular recommendations. Using the calibrated model outputs, this tool identifies specific opportunities for reallocating budgets across channels to enhance media efficiency and maximize ROI. It helps marketers optimize their spend in real-time, ensuring that every dollar is strategically invested.
• Competitive Intelligence: Industry benchmarks play a critical role in helping marketers understand their position within the market. The Competitive Intelligence application contextualizes performance by comparing it against industry standards, historical results, and peer benchmarks. This allows marketers to see how they stack up against competitors and find areas where they can push for further improvement.
• Optimization Report: This essential report makes it easier for marketers to prove the value of their media investment decisions and demonstrate the business impact of those changes to key stakeholders such as CMOs, CFOs and Board Directors, helping build cross-functional alignment. Unlike other Measured reporting applications that track static media performance KPIs over time, like incremental ROAS or CPO, the Optimization Report uniquely helps marketers understand the true incremental revenue driven specifically by actions they’ve taken, holding Measured to ultimate accountability for the performance of our recommendations.

A Clear Roadmap for Success

By integrating the applications listed above, the Measured Platform not only measures marketing performance but actively helps marketers move from data to action. We empower brands to optimize their media mix, experiment with new strategies, and make data-driven decisions to reach their business goals more effectively. 

The insights provided by these capabilities turn raw data into a clear roadmap for success — guiding marketers on how to improve performance, explore new growth opportunities, and maximize the return on every marketing dollar spent.

Measured is the pioneer and leader of incrementality-based measurement and optimization and our world-class team of measurement experts have more than 100 collective years of experience working with MMM. 

We show up for you as a dedicated partner, fully invested in your success, offering tailored solutions, and continuous support and guidance to uncover hidden opportunities that maximize your ROI. To find out more about our partnership and how it can benefit your brand, reach out to a Measured expert today.