What is Marketing Mix Modeling and How Does it Work?

We’re going to show you how marketing mix modeling works, why it’s needed, plus the pros and cons of using it to assess your marketing efforts.

You know what it feels like to put together a report for your head of marketing, CEO, or investors.

It’s hard work.

You need precise data. You need to justify your conclusions. You need to prove why you deserve a bigger marketing budget.

With marketing mix modeling, you can combine all of your channel metrics, analyse them, and show the effect of marketing on your sales over the period.

It doesn’t just sound great. It is great. 

But before you get ahead of yourself, it helps to understand what marketing modeling (MMM) is, why it’s needed, how it works, and the pros and cons of using it to assess your marketing efforts.

And that’s exactly what we’re going to discuss.

Here’s what you’ll learn: 

• What is marketing mix modeling
• How does marketing modeling work
• Marketing mix modeling vs marketing attribution
• What are the benefits of MMM and attribution
• How to get started with MMM

Pro Tip

Chances are you’re here because you’ve considered using marketing mix modeling. If so, why not give Ruler Analytics a try? Ruler takes the best of marketing mix modeling and multi-touch attribution to give you the bigger picture and help you understand the factors that affect your sales, deals and ROI. 

Book a demo to learn more

What is marketing mix modeling?

Marketing mix modeling (MMM) is a statistical analysis technique used by companies to evaluate the impact of each marketing input on consumer behavior, sales and ROI. 

The concept “marketing mix” was first coined in the 1960s, and marketing mix modeling as a formal analytical technique was developed and popularised in the 1980s and 1990s by marketing experts.

In recent years, marketing mix modeling has made a comeback in digital due to the challenges we face around iOS 14.5 and third-party cookie tracking.

MMM breaks down your results by channel, allowing you to see which marketing activities are having the most significant impact on your desired outcomes.

It includes several variables to evaluate the impact and effectiveness of your marketing efforts. These variables typically include:

• Sales data. Historical sales data e.g. unit sales, revenue, and market share.
• Marketing spend data. The amount of marketing budget spent on various marketing activities, such as advertising, promotions, and sales support.
• Market and consumer data. Data on consumer demographics, attitudes, and behaviours, as well as data on market trends and conditions.
• Product data. Product features, pricing, and distribution.
• Economic data. Data on economic indicators, such as interest rates, inflation, and consumer confidence.
• Competitor data. Information on competitor marketing activities, such as advertising spend and promotions.

The goal of MMM is simple. It’s to provide you with insights into how changes to the marketing mix can impact sales and customer behavior and help make better predictions about the likely outcomes of future marketing decisions.   

How does marketing mix modeling work?

Marketing mix modeling uses the principle of multi-linear regression (MLR). Multiple linear regression is a statistical technique used to model the linear relationship between a dependent variable and one or more independent variables. 

With MLR, you can estimate the impact of each independent variable on the dependent variable and make reliable predictions about the impact of future changes to the marketing mix. 

If you’re new to marketing mix modeling, you’re probably wondering what the difference is between a dependent and independent variable.

A dependent variable in MMM is typically a measure of business performance. Here are a few examples:

• Volume of sales. The goal is to understand how different marketing activities influence business results and to make predictions about the impact of future changes to the marketing mix on sales and ROI.

• Revenue. Similar to sales, revenue is a common dependent variable in MMM, reflecting the total amount of money generated by sales.

• Market share. Another dependent variable used in MMM is market share. This is to understand the impact of marketing activities on market position.

So what makes an independent variable? An independent variable represents the marketing mix elements that are believed to impact the dependent variable. Here are a few examples: 

• Advertising spend. The amount of money invested in promoting a product or brand through various channels, such as TV, print, online, and outdoor advertising.

• Price. The price of a product is often used as an independent variable in MMM to understand the relationship between price changes and sales.

• Promotions. The impact of promotions on sales is also evaluated using MMM, such as discounts, coupons, or buy-one-get-one-free offers.

• Distribution. The availability of a product in different retail locations, including online and brick-and-mortar stores, is an important independent variable in MMM. Changes in distribution can impact sales, and MMM can be used to understand the relationship between changes in distribution and sales.

While MLR is a widely used method in MMM, other statistical techniques, such as time-series analysis, logistic regression, or machine learning algorithms, may also be used depending on the specific needs and goals of the analysis.

Marketing mix modeling vs marketing attribution: What’s the difference?

Marketing mix modeling is often compared to marketing attribution, but they’re actually quite different. They’re both related but have distinct analytical approaches for evaluating the impact of marketing on business performance.

Multi-touch attribution

Marketing attribution is the process of assigning credit for a sale or conversion to the specific marketing touchpoints (e.g. ads, email, website) that contributed to the outcome. 

Multi-touch attribution, in particular, uses various models to determine the contribution of each touchpoint and assigns credit to the interactions that had the most impact on the final conversion.

Related: What is multi-touch attribution and how does it work?

Multi-touch attribution, in particular, uses various models to determine the contribution of each touchpoint and assigns credit to the interactions that had the most impact on the final conversion or sale. 

Here are a few examples of MTA models:

• Last-click attribution – assigns credit for a conversion to the last touchpoint or interaction that a customer had with a brand before making a purchase.
• First-click attribution – assigns credit for a conversion to the first touchpoint or interaction that a customer had with a brand.
• Linear attribution – assigns equal credit to each touchpoint or interaction that a customer had with a brand before making a purchase.
• Time-decay attribution – assigns more credit to touchpoints or interactions that occur closer in time to the conversion.
• Position-based attribution – assigns more credit to touchpoints that occur at the beginning and end of the customer journey and less credit to touchpoints in the middle of the journey.

These are just a few examples of commonly used marketing attribution models. The choice of attribution model will depend on the specific goals and objectives of the organisation and the type of data available for analysis. 

It’s also possible to use a combination of attribution models to create a custom model that combines elements of different models.

Pros of marketing attribution:

• Multi-touch attribution can be quick and easy to implement, ideal for organisations that need fast and efficient results.
• It provides immediate feedback on the performance of individual marketing channels, making it a useful tool for organisations looking to make quick decisions about marketing spend.
  

Cons of marketing attribution: 

• Multi-touch attribution tracks click path data and can’t account for ‘invisible’ contributions like ad views, offline marketing, and word of mouth.
• Multi-touch attribution is less effective due to privacy changes and regulations (e.g Apple iOS 14.5).
• Attribution models can be subject to bias, especially if they are based on incomplete or inaccurate data. 
• It often attributes too much credit to demand-capturing channels (e.g. direct and search) and undervalues demand-generation channels (e.g. dark social.)
• Attribution can only prove the value of your short-term marketing activities.

Marketing mix modeling

We’ve already covered marketing mix modeling and how it works. To recap it’s a statistical technique used in marketing to analyse the impact of various marketing inputs on sales and other business-related metrics. 

But there’s one thing we haven’t yet pointed out. 

Unlike marketing attribution, MMM uses a broader approach to marketing measurement. MMM often incorporates marketing attribution as one of the inputs in its analysis. So you can think of marketing attribution as a subset of MMM.

Marketing attribution information can be used in MMM to better understand the effectiveness of different marketing channels and how they work together to drive sales. 

This information can then be incorporated into MMM’s analysis to provide a more comprehensive view of the impact of marketing activities on business outcomes.

Pros of marketing mix modeling:

• MMM quantifies everything, so you have real numbers to back up your reports. For example, seasonal changes or changes in general user behaviour.
• Allows you to measure the impact of offline media channels like TV, radio, and traditional print advertising.
• It’s statistically reliable. After all, it’s based on an extensive data set, and you’re applying proven statistical modeling methods such as regression analysis to it.
• MMM is a privacy-safe measurement solution, making it a more effective approach in a post-cookie and privacy-legislated world. 
• It provides insights into the long-term impact of marketing activities, which can be useful for planning and forecasting future sales and marketing efforts.
  

Cons of marketing mix modeling:

• It only provides a top-down view and can’t show you an accurate picture of your individual campaign performance.
• MMM is often very complex and requires the help of specialised software or a highly skilled statistician to run them (P.S. We have a workaround for this, so stick around.)
• Marketing mix modeling requires you to collect a large amount of data for it to run successfully. 
• Traditional MMM is expensive and time-consuming to build.

How to get started with marketing mix modeling 

Marketing mix modeling has become an increasingly popular and important tool for marketers, especially with the “cookie apocalypse” approaching fast. It’s estimated that 40% of Fortune 500 companies already use marketing mix modeling technology. 

To get started with MMM, you need to gather and organise your data, decide which models to use, and set up the system. 

These steps typically include: 

• Gathering data. To start with MMM, you need to collect data from your marketing and sales tools. This includes data on your marketing activities (e.g. media spending, ad campaigns), as well as business outcomes (e.g. sales, ROI, etc.)

• Setting up a MMM system. Once you have the data, you’ll need to set up the modeling system. This involves choosing the models you want to use, deciding on the metrics to track, and setting up the software.

• Analyse the data. This generally involves looking at the relationships between your marketing and business data and discovering insights about how your marketing activities impact your business outcomes.

• Adjust and optimise. Finally, use the insights from your analysis to adjust and optimise your marketing mix and maximise your return on investment (ROI).

These steps sound simple, but in practice, it’s quite difficult to set up successfully, especially if you’re new to the principle of marketing mix modeling. 

Another option is to invest in an attribution and MMM solution. It’s less complicated and often more cost-effective than having to hire a statistician to build a marketing mix modeling system from scratch. 

Take Ruler Analytics, for example. 

Ruler takes the best of multi-touch attribution and marketing mix modeling to help you better understand your marketing performance and identify areas to prioritise.

It lets you see how all of your channels (online and offline) are performing and how well they can be attributed to conversions and sales.

How does Ruler’s attribution and marketing mix modeling work?

At Ruler, we’re on a mission to blend marketing attribution and marketing mix modeling to create a hybrid approach to marketing measurement. 

This allows marketers to: 

• better understand the impact of clicks and ad views on revenue outcomes
• overcome the walled garden and cookie limitations
• track how offline events, TV and radio affect the business and ROI

But how does Ruler work? Let’s dive deeper into its technology.

Ruler starts by tracking the entire customer journey at the visitor level on a first-party basis. 

It captures the marketing source from each session, page views, UTM variables, Click IDs, and Cookie IDs which are then matched to a lead driven from a form, phone call or live chat.

Ruler passes the marketing source data you’ve captured on your leads to your CRM and other marketing tools. This allows you to enrich your leads and opportunities with attribution data so you can see exactly how your marketing impacts pipeline generation.

When a lead is marked as closed as won, the revenue data is sent back to Ruler. With revenue and opportunity data in Ruler, you can easily measure and validate the impact of your marketing sources, campaigns and keywords. 

But there’s more yet. 

Ruler isn’t just your ordinary attribution platform. It uses marketing mix modeling alongside its multi-touch attribution technology to give you the bigger picture and help you understand the factors that affect your sales and ROI. 

Here are the three core features of Ruler’s MMM: 

1. Impression attribution modeling

Ruler’s impression attribution gives you more insight into your invisible touchpoints e.g. ad views, tv and radio impressions.

A good example would be when a user is served an ad, and then instead of clicking, they go directly to your website to convert into a lead or customer. With Ruler’s impression modeling, you can better understand when, where, and how people are seeing your ads and offline initiatives.

2. Budget allocation recommendations

Due to the way impression impact is calculated using diminishing return curves, Ruler is able to model your ROI over time and project how much headroom is still left in your advertising channels.

From there, Ruler is able to recommend an optimal allocation to different channels depending on performance, allowing you to get the best bang for your buck.

3. Conversion modeling

Since iOS 14.5, retargeting options on Facebook have been limited. With Ruler, you can send MQLs, opportunities, and closed revenue back to Google Ads and Meta as conversions for reporting and algorithmic optimisation purposes (e.g. smart bidding).

By combining these two approaches, you can gain a complete picture of the relationship between your marketing and customer behavior and the overall impact of your investments on the business.

Book a demo of Ruler to see it in action for yourself here. 

Need help getting started with marketing mix modeling?

With marketing mix modeling set up successfully, you’ll have data linking your sales to your marketing efforts.

You’ll be able to see a clear picture of your marketing wins (and losses), so you can improve and optimise your marketing strategy for maximum results.

If you’re looking for a marketing reporting system that unifies data across online and offline channels and links your conversions directly to revenue, then Ruler is for you.

It allows you to access revenue data and link it directly to your marketing activities without any of the headache of doing it yourself.