Marketing Mix Model on the rise again (1)
Today, marketing has become a very important part of every company, and its growth has never slowed down. There are all kinds of products, all kinds of marketing tools, and all kinds of questions about how to effectively stand out and engage with customers with limited tools.
But the questions are not new. While the actual behavior has changed with the evolution of marketing tools and competing products, the basic question has always been the same. It’s about “optimizing resources” and “maximizing performance” — getting the most out of a given set of resources.
To this end, various methods have been devised to measure the performance of each of the various budgetary expenditures in marketing, optimize them, and maximize the performance of marketing within a given budget. The Marketing Mix Model (MMM) is an application of statistical methodology for these methods.
What is an MMM?
The MMM is an econometric approach created to evaluate the effectiveness of marketing in terms of return on investment (ROI) and utilize it as a way to understand how many business results (e.g., sales) are generated by marketing activities.
MMM starts by collecting data. This can be collectible time series data such as product sales or advertising budgets, seasonality or macroeconomic factors, or data gathered through surveys of consumer responses to promotions or ads. After segmenting and aggregating this data to separate the contributions of marketing strategies and promotional activities from other uncontrollable factors, time series modeling reveals how much each activity contributed to business results. The time series models used in traditional MMM are primarily multivariate regression models.
This allows you to understand the impact of each of the 4 P’s (product, price, place, and promotion) and the more granular marketing elements on each of them, which can then be simulated to optimize marketing resources. The results are often produced in the form of a revenue decomposition for each of effectiveness, efficiency, and ROI.
How the traditional MMM flows
The term “marketing mix” was first used in 1949 by an advertising professor named Neil Borden. In a journal article, Borden explained the concept of the marketing mix by stating, “When building a marketing program to fit the needs of his firm, the marketing manager has to weigh the behavioral forces and then juggle marketing elements in his mix with a keen eye on the resources with which he has to work”.
MMM, the statistical modeling implementation of this concept, began to gain attention in academia in the 1980s and was adopted by advertising agencies as a statistical analysis method for forecasting sales and informing marketing investments. However, like many statistical modeling approaches adopted by companies, MMM has its limitations in that it requires processing and analyzing large amounts of data about marketing, sales, and pricing, and because variable creation and modeling is a different area of expertise than traditional marketing, it was implemented and operated as a large-scale project by a few large companies with sufficient capital.
Pros and cons of traditional MMM
MMM enables organizations to more truly understand the impact of marketing on ROI, which in turn provides data-driven insights for effective budget allocation. MMM can also be used to forecast and simulate product sales and utilization trends.
However, MMM has not been widely used because it is only feasible to implement and operate on a large scale. The cost of communication between marketers and statisticians to utilize MMM is high, and the process of collecting and processing data is fraught with problems. The models do not perform well because they use problematic data, and they are based on long-term time series data, which is not ideal for fast-moving and changing products.
The mobile age and the rise of data-driven attribution
At a time when numerical and objective marketing analytics were necessary, but MMM was heavy and cumbersome in many ways, the mobile era, with its increasing use of smartphones, marked a major turning point in marketing. It was possible to categorize users, collect individual data, and track them. With the ability to track user activity in near real-time and market based on it, data-driven attribution, such as Last Touch Attribution (LTA), has emerged as a new marketing methodology. Marketers can now quickly see ad impressions, clicks, and purchases made by users on mobile and easily understand marketing performance and optimize budgets. In this environment, MMM became increasingly obsolete, and as data analysis became easier, MMM was not used as a framework at all, but marketing analysis using time series analysis, etc. were used, but they also seemed to be somewhat outdated.
The Age of Privacy
The ubiquitous use of personal information today has led to the introduction of data privacy laws, such as the European General Data Protection Regulation (GDPR), and the strengthening of existing privacy laws. This has led to the introduction of app tracking transparency policies and stricter browser cookie policies, making it increasingly difficult to collect personalized data. These changes have led to a need to reduce reliance on individual user-level data tracking and analytics, which have been popular in recent years, and have led to a renewed interest in MMM after a brief hiatus.
It’s also easier than ever to collect and utilize data today, making it easier than ever for organizations of all sizes to design, implement, and operate an MMM that works for them. There are also a number of methodologies that have emerged that fill in the gaps in MMM.
(Continued to Part 2)
