The saying goes that you cannot manage what you cannot measure. When it comes to choosing the right media mix to achieve your marketing efforts, how do you know what is working and what isn’t? Media mix modeling is a group of technologies and practices geared to identify the impact (in money and results) of your marketing tactics on your ROI. In this guide, we prepared all that you need to know about media mix modeling to start optimizing your campaigns.

Short on time? Here’s the table of contents:

What Is Media Mix Modeling in Marketing?

Media mix modeling (MMM) is a marketing analysis technique that determines the impact of marketing efforts on sales. An organization’s media mix is part of the marketing mix model and consists of combining an organization’s communication channels to get its brand message and marketing tactics to potential customers. 

The media mix may combine traditional advertising channels, like print, broadcast, TV, social media, and online advertising. Companies talk about the marketing mix when planning their campaign goals, which is an essential part of their marketing strategy.

What Is Media Mix Optimization?

Organizations optimize their media mix in order to gain insights into what they need to target their audience effectively. Not all companies can optimize their media mix because it is more suitable for online marketing. It requires looking into the analytics and ROI of different marketing strategies.

This is where media mix modeling comes to help.

How does Media Mix Modeling MMM work?

Media mix modeling, also called marketing mix modeling, analyzes collected and processed data from the channels that form the marketing mix. Some solutions enable marketers to factor in traditional channels, promotions, seasonality, and other variables.

The modeling collects data from disparate sources, which then applies advanced statistical analysis to, and enables to get insights into how effective the current campaign. MMM leverages metrics and variables like sales, ratings, or online analytics, allowing analysts to have a broader picture of the impact of the campaign in the marketplace in a measurable way.

MMM analyzes linear and non-linear variables. This means there are variables that a direct relationship with sales can measure. The more you increase the marketing input, the more sales grow. But other variables, like broadcasting, are more difficult to track. If a marketer would do that manually, it would be extremely difficult. MMM technology enables marketers to use artificial intelligence and advanced analytics to find out a quantifiable impact of each marketing effort, regardless of the channel.

The goal of a media mix modeling study is to measure the impact of each marketing activity on each channel. It works by quantifying the effect of advertising, pricing, PR, and sponsorships.

The term was coined in a paper by the Harvard Business Review, and the technique has been around for a few years. Thanks to the advance in statistical methods and artificial intelligence, media mix modeling can be done now in a simpler way.

The factors that may affect the marketing mix can be categorized as: 

Incremental drivers: this refers to business outcomes generated by marketing activities like print ads, digital spending, price discounts, social outreach.

Base drivers: this refers to outcomes achieved without any advertisements, usually due to brand equity. These outcomes usually don’t change unless there is an economic or environmental change.

Other drivers: related components of baseline factors, measured over a period of time.

Media mix modeling diagram

Example of a media mix modeling diagram (Image source)

How do you measure MMM?

Media mix modeling is measured by using regression analysis, specifically multi-linear regression. The model uses dependent and independent variables to identify a relation between them.

Analysts form an equation between the dependent and independent variables. Depending on the relation between the variables the equation can be linear or nonlinear. Here’s an example of an equation of multi-linear regression where each beta shows that an increase affects the total increase of sales.

Example of a sales equation

Example of a sales equation (Image source)

MMM helps marketers in optimizing future spending and maximizing the effectiveness of the marketing campaign.

Media Mix Modeling Ratio

Besides complex equations, the MMM ratio consists of three key components:

  1. What marketing channels are you using?
  2. How much money are you spending on each channel?
  3. What were previous campaign results and insights?

The answer to these three questions may determine what is the rate of effectiveness of your marketing efforts.

Marketing Mix Modeling Examples

Understanding media mix modeling can be daunting at first. Let’s explain it better with some marketing mix modeling examples: 

Case 1. 

A fashion brand that wants to find out how each media channel’s marketing contributes to sales. Let’s say the brand has both online and social media ads, and they want to know if it should focus its efforts more on social media. They’ll take the collected sales data during a specific time frame, typically two to three years. Then they use MMM to run a multivariate test, which can show how sales would look when changing the media set.

Case 2. 

A company that sells across different geographic regions. They want to know what of their marketing campaigns is affecting more on a specific region’s sales. They have to break their marketing budget by different channels, for example, paid search in Google vs. Bing. The MMM models their data and then tries to explain it without marketing input

Some tips we can take from these marketing mix modeling examples include taking at least two years of data so you can have two seasonal cycles and combine marketing and non-marketing data.

What Elements does the Media Mix Modeling Measure?

MMM Is a marketing analysis technique that measures what is the impact of a campaign and determines how each part of the marketing mix contributes (or not) to its success. The results of a media mix modeling study can give you insights that you can use to improve a campaign. Let’s summarize this with a definition: 

Media Mix Modeling is a top-down approach that uses tools and advanced analytics to evaluate how media and marketing activities, pricing, seasonality, and variable factors impact sales and ROI. It provides a measure of how activities contribute to the company’s ROI

Marketing analysts use data science techniques such as multi-linear regression to determine the effectiveness of each marketing input in terms of ROI. The goal is to identify which marketing efforts have higher ROI and are thus more impactful.

Ad Effectiveness

MMM model example (Image source

1. Sales Volume

Sales volume is one of the key elements measured in MMM. But the model considers two different aspects: base and incremental sales.

Base sales: These are the sales influenced by parameters like long-term trends, seasonality, brand awareness, brand loyalty, and pricing. We can also consider baseline sales as the revenue you could have without marketing input.

Incremental sales: They are the part of sales influenced by marketing activities. You can calculate incremental sales by applying the following formula:

Incremental sales = Total Sales – Baseline Sales

2. Media Marketing and Advertising Impact

What should you look into when measuring the impact of media and advertising on sales? For instance, comparing how effective are 15-sec vs 30-sec ads. Or comparing how ads running on different platforms perform. Or finally, comparing the outcomes of running ads during different times.

3. Pricing models

Sales are highly sensitive to pricing, but how much changing the pricing can influence your sales? You can see the impact with the help of MMM. Knowing this percentage is essential for your pricing strategy.

4. Distribution Systems

Distribution is key to preventing overlooked expenses, which can drive growth. Applying media mix modeling can help teams understand how changes in distribution can impact the sales volume across channels.

5. New Product Creations

Every product launch can impact the overall sales volume. A successful launch can pick up sales, but also increase the marketing budget. Marketing mix modeling can help determine how much of the sales volume is a result of the new product.

Common use cases for using MMM

Media mix modeling or as is also known as marketing mix modeling can be used to measure and optimize your marketing channels in terms of ROI. Here are some use cases you can apply this technique to: 

Budget setting and optimization: Large companies with geographically distributed campaigns across multiple media channels can benefit from the scalability of media mix modeling. Media mix modeling leverages automation to perform large-scale marketing effectiveness. 

Media measurement: You can measure the impact of different types of media campaigns, paid, owned, and earned. You can use media mix modeling to measure the customer journey in its path to purchase. The insights you get can be used to optimize your spending and actions across those channels. 

Measuring sales drivers: Marketing mix modeling can be used to find what are the factors driving sales, so you can invest more in the winning strategy.

How to Use Media Mix Modeling?

Media modeling gives marketers the opportunity to support their decisions with data, creating a data-driven approach that is more accurate and actually can save money and effort.

Research from a Forrester study, “The Current State of Marketing Measurement and Optimization”, shows that 71% of marketers are impaired by inefficient measurement methods and tools. Here is how to make the most of your media mix modeling:

1. Collect data at the personal level

At this moment, when third-party cookies are about to become a thing of the past, marketers everywhere are trying to collect the information they need. Personal level data allows you to have an accurate picture of how customers relate to the media mix you chose.

Person-level data means you assign data from sources to an individual consumer with the goal to answer business questions and pinpoint interactions at the user level. [CLICK TO TWEET]

In the Forester report mentioned above, 99% of marketers not currently using person-level data would like to use this approach today.  This granular approach allows you to conduct analysis at the user level instead of using the already aggregated data.

2. Check the type of data

Media modeling works better if you are working with digital channels than for traditional marketing methods. It is more difficult to measure the results of a newspaper ad or a radio broadcast. Achieving the right marketing mix, with a greater investment in online marketing channels (including mobile) can give you a more accurate picture. This is also consistent with trends towards online and mobile content consumption by users. By migrating campaigns to online channels, you can measure ROI more accurately and have better insights for decision-making.

3. Choose a platform that’s right for your organization

Using analytics software gives you an advantage. You can analyze the media mix by using platforms that collect user interaction data and provide tracking reports. The best approach is to choose a platform that gives you complete visibility of all the channels you are implementing. A software that can provide accurate and timely reports is also a must. You need to know how your channels are performing individually and as a part of your marketing campaign.

4. Analyze the data

Before getting into the analysis, you need to choose what metrics you want to measure for each channel. The wrong metrics can give you a completely different picture that is not akin to reality. Choose the metric you want to measure according to the goal you want from that channel or activity. For instance, email marketing newsletters can be measured by click-through rate more effectively than by measuring opening rates.

Once you chose the metrics and got the data, it is time to analyze and understand the reports. It is important to know what the data is telling you to use it to your advantage. Following the example of the email newsletters, if you see a high CTR from them, it is a sign that you should use this strategy for the next campaign for that audience.

Try to find the “high-performers” and the “low-performers” too. Knowing where your strengths and weaknesses in the campaign will help you adjust and improve it for next time.

5. Keep in mind social sentiment and brand perception

The success of a marketing campaign is not only measured in terms of conversions or clicks. Understanding how your potential audience perceives your brand can provide context and help you interpret the data better. Factor consumer opinion in your media mix model. How do you do it?

Conduct social media and search for sentiment analysis. Take note of what people are saying about your brand, the positive and the negative. You can use that information to create a survey and prove your findings by rating your company. Specifically, asking how likely they would recommend your business to a friend and what type of marketing content would they like to see more. This will give you an idea of where to focus your marketing efforts next.

How Do You Know The Media Mix is Right for Your Brand?

How do you determine the right media mix for your campaign? Let’s look at some of the factors you should consider when choosing your media mix.

Using multiple marketing channels to promote products and engage your users is a popular approach. However, randomly choosing as many channels as possible is not only ineffective but also can lose you a lot of money. Choosing the right mix is essential to achieve a successful campaign.

How do you start? By knowing and understanding your target audience. After all, your goal is to engage them. There are two key steps in choosing the right mix for your marketing strategy:

Define your target audience

This is the foremost step because without understanding your audience you are in the dark. Start by mapping basic demographic data: Location, gender, income, age, education level. Then you can further dig into finding interests, platforms they visit. How do you do it?

  • Check at your competitors: you can gain a lot of information about your potential customers by checking your competitor’s campaigns and social media sites.
  • Search in social media groups of interest: your customer talks about products related to yours on social media and review sites. Take a look at what they are saying, where they are located, and such.

You should know who the potential customers are for your product or service. A good rule of thumb is to create buyer personas to have a detailed idea of who is your ideal customer.

Collect and use reliable data

Gather data on your target audience according to what you know about your audience. For instance, organic research, competitor audits, sentiment analysis. Check data from media viewing research sites too to have a broad picture. The right data can provide the insight you need to choose the market mix that works.

Related content: What is Media Buying and the Best Templates to Download

Why is it worth Implementing the MMM (Marketing Mix Modeling)?

Marketing mix modeling is worth implementing, because it can help your organization understand the data behind your sales, understanding exactly how much exactly your marketing tactics are working. With the MMM you can understand how different activities drive a product’s metrics. You can use the modeling to optimize your marketing budget and get the most ROI.

Media mix modeling vs Data-Driven attribution modeling

Modern marketing is based on hard data, especially digital marketing. One of the questions prevalent in marketing departments is where the marketing budget goes. Attributing where the money was spent to lead generation and marketing goals is one of the key objectives of every marketer. Despite ongoing efforts and data-driven analysis, it is a challenge to attribute accurately. Marketers diverge if attribution modeling or media mix modeling is the best measurement model to use.  Let’s examine each one.

What is the Attribution Model?

Attribution Modeling is a bottom-up approach used for measuring marketing efficacy. This method analyses and identifies the value of each marketing initiative by looking at the actions users take before converting. 

Attribution modeling focuses on the outcomes of the marketing efforts as measurements, online sales, advertising, and similar conversion efforts.

There are five types of attribution models: 

  • Last interaction

Last interaction

This involves attributing the credit of the conversion to the last lead a user interacted with. This method is used by default in many marketing teams. For example, if a user finds your site by a Google Ad, but makes the purchase finally from a Twitter ad, the ad gets 100% credit for that sale. 

  • First Interaction

First Interaction

This involves assigning credit to the first introduction of the user to the business. In the above example, the Google ad would get credit instead of the Twitter ad. 

  • Last non-direct click

This model also attributes all credit to a single interaction. The basis of this approach is that the last action is triggered by the last non-direct click because it is when the user is exposed to your marketing efforts. 

  • Linear attribution

This model divides the attribution equally among all user interactions before conversion. That means ⅓ would go to the Google ad, ⅓ to your website, and ⅓ to the Twitter ad. The problem with this model is that it doesn’t account for the level of influence of each interaction. 

  • Time decay attribution

An evolution of the linear attribution model takes into account when each interaction takes place and gives more importance to the interactions that happen close to the time of purchase. This would give the Twitter ad more value than the other interactions. 

  • Position-based attribution

This model also splits the difference when allocating conversion credit. It gives 40% to the first interaction, 40% to the last, and 20% to divide among all the other interactions.

The difference of Media Mix Modeling

Media mix modeling uses a regression analysis evaluating the impact of multiple variables on a single variable like sales figures. It calculates the relationship between the independent variables and the dependent variable.

Attribution modeling could have worked in the past for simple marketing strategies with few channels. However, this proves difficult for the complex and distributed strategies of today’s marketing. Media Mix Modeling can account for a wide range of data from diverse sources.

Measured vs. Platform Reporting, Multi-touch-Attribution (MTA), and Media Mix Modeling (MMM) Comparison Table:

Neutral and Independent
Causal incremental
Scale testing
Granular insights
Cross channel
Walled Garden
Strategic Planning
Timely Insights
Data Management
Built for marketing analytics
Data Quality
Data Management
Other Measurement
Data Management
Other Measurement
Data Management
Other Measurement
Data Management
Measured Advantage
Trusted measurement;
Productized Experiments
Identify saturation curves
Future proof
Depth of Measurement
Daily and Weekly Insights
Bottom Up Forecasting
On Time, reliable
Analytics Ready
Source of Truth Platforms

Pros and Cons of conducting a MMM

When should you use MMM? Implementing a media mix modeling can be more effective:

There is enough data to estimate the parameters in the model.
There is a range of variability in the advertising levels and control variables.
The model inputs vary independently
The model accounts for the drivers that may impact ROI.
The model captures the relationship between variables

There are challenges from issues that may affect the reliability of results from MMM. 

So, what are the pros and cons of using marketing mix modeling? Here is a summary:

MMM Pros
It is a great starting point for planning a marketing budget at a high level.
It provides a holistic approach to marketing trends.
It highlights trends and insights on what is working on the campaigns.
MMM Cons
It doesn’t give insights at a granular level.
It is difficult to measure digital and traditional marketing ROI.
It doesn’t account for cross-channel impact.
Doesn’t account for different times to send marketing messages

Limitations of MMM Marketing Mix Modeling

When using marketing mix models, marketers need to take into account several elements across their ecosystem, which may include:

  • Person-level, behavioral data
  • The impact brand authority has on the marketing spend
  • What are the key times to send marketing messages
  • What’s the proper attribution on individual media effectiveness

Taking into account all these metrics may have caused issues of the reliability of the media mix modeling. Marketing mix models enable marketers to unify the measurement. 

Another problem with marketing mix models is that they usually require a lot of data to give accurate insights. Typically, two years of historical data. Many organizations don’t have the data infrastructure in place needed to collect and process these massive amounts of data.

Advantages of Marketing Mix Modeling

While MMM cannot identify individual opportunities to optimize their campaign optimization. It gives the starting point for high-level marketing budget planning, providing a holistic approach to general market trends, giving marketers a full-circle view into their prospective markets.

Common myths about MMM

Like many analytic solutions has become very popular, but does it live up to all the hype? Here are a few misconceptions people have around media mix modeling:

  • Media mix models are obscure: since there are datasets and advanced analytics involved in media mix modeling, these methods are considered lacking in transparency. This raises the question of how do you know if the model is accurate if you cannot see it all? The right approach is implementing a transparent approach, determining deliverables, outlines, milestones, and reports.
  • MMM doesn’t provide real-time data: the truth is that MMM is based on historical data. However, modern media mix models can provide near real-time marketing insights, that can evaluate new campaigns, and assess the effectiveness of a  running campaign.
  • Is biased to offline/online channels: media mix strategies may focus more on offline channels. But modern media mix models consider all channels, digital and offline. Media marketing models are adapted to take into account each channel and its importance as a factor.

The History of MMM

Marketers started to use Media (or Marketing) Mix Modeling in the golden age of advertising, around 1960-1970 when marketing was much simpler than today. One of the early users of media modeling was Kraft Foods when they launched Jell-O. 

Traditional MMM allowed Kraft to see how sales would be affected according to different levels of advertisement and geographical location. For instance, how would sales increase by running campaigns in 10 cities instead of four?. 

Nowadays, with the application of artificial intelligence data analysis to media mix modeling, analysts can get insights practically in real-time as campaigns are running.

What to look for in MMM tools

To implement an effective media mix modeling you need marketing performance tools that give you the insights you need. Here’s what you need to know when looking for a solution:

  • Balancing long and short-term growth: most of your efforts should be focused on long-term growth but don’t overlook short-term goals. The Institute of Practitioners in Advertising suggests a ratio of 60/40 long and short-term marketing activities. Your marketing performance tool needs to analyze how both campaigns will grow your business.
  • Collects and measures data from disparate sources: this is one of the basic features of a marketing performance tool. To be effective at media mix modeling you need a tool that can collect, process, and analyze data from digital and traditional media. Since most of these data sources have their own analytics, you need an orchestration platform that can ingest the data from these sources and give you the insights you need.
  • Takes into account external variables: political, economical, and social changes can affect marketing efforts. A good tool needs to recognize disrupting variables and assess how they would impact your long-term campaigns.  
  • Consider the customer journey: a media mix modeling needs to account for the interactions along a customer journey. Your tools should be able to tell you what is the impact of each step, considering customer purchase patterns and predicting consumer trends.

FAQs About MMM

How do you do a market mix model?

Base variables or incremental variables are taken into account, quantifying them and breaking down business metrics to find out how marketing and promotion activities contribute to ROI. 

What type of Modelling method is critical for marketing mix evaluation?

Marketing Mix analysis is typically done using linear regression. Other effects as non-linear and lagged are included to have a broader approach. 

What is market mix Modelling?

Market Mix Modeling is a technique that helps to quantify several marketing inputs on sales or market share. The goal of marketing modeling is to understand how much each marketing technique is affecting sales in terms of ROI. The idea is to detect which marketing inputs have a higher ROI and which ones need adjustments.

How CodeFuel optimizes media buying and management

Media buying is an essential part of media mix modeling. After all, the channels you use will determine how they will impact your marketing campaign. CodeFuel simplifies media buying by serving intent-based search and shopping ads. By delivering the right ads to the right buyers at the right time, your campaign is more effective. Start optimizing your marketing campaigns effectively with CodeFuel today. Sign up.