What are the limitations and areas in which MMM (Marketing Mix Modeling) does not excel?
MMM (Marketing Mix Modeling) is an effective method that many companies have used to analyze the effectiveness of their advertising and marketing activities, but every method has its limitations, and MMM is no exception. In order to use MMM appropriately and effectively, it is important to understand not only the advantages of traditional MMM, but also its limitations.
So in this article, we'll give a brief overview of traditional MMM, explain its typical limitations in detail, and introduce some alternative methods that complement MMM's weak areas, so please read on.
table of contents
MMM Basics
Definition and purpose of MMM
MMM is a method for statistically analyzing the impact that various marketing activities, such as advertising, promotions, and pricing, have on business results such as sales. This method allows companies to measure the effect of each measure on their marketing investment and calculate ROI (return on investment), making it possible to simulate optimal budget allocation.
MMM's Approach
MMM simultaneously analyzes the impact of multiple marketing activities based on past data. This analysis uses performance data such as sales and sales volume, advertising costs and advertising placement data, and external factor data such as economic indicators, and employs statistical methods such as regression analysis.
Key Benefits of MMM
By using MMM, you can quantitatively evaluate the effectiveness of each marketing activity, enabling optimal allocation of investment. By understanding which measures are generating high ROI, you can aim to implement measures that are effective and efficient.
Furthermore, because MMM is a scientific approach that bases analysis on concrete data, it enables data-driven decision-making that does not rely solely on intuition or experience.
The Limitations of MMM
Although MMM is a very useful method, it is not an all-purpose method. In particular, it has the following limitations. It is important to understand these limitations and determine how to use MMM appropriately, and take measures or consider alternative methods as necessary.
Pre-implementation limitations
Large amounts of granular data are needed
MMM requires a lot of granular data to be accurate (in most cases, daily or weekly data is required, and the data spans the last 1-3 years). The quality and quantity of data determines the reliability of the model, which can be a big hurdle for companies with small marketing activities or limited data collection processes.
Key impacts:
- Lack of data makes analysis impossible or reduces the accuracy of analysis
- For some companies (who lack or cannot obtain data), it is difficult to introduce MMM.
Countermeasures/alternative methods:
- Establishing a data collection process and ensuring a sufficient collection period
- Utilizing alternative analysis methods, such as A/B testing
The implementation process is time-consuming
In implementing MMM, maintaining data integrity is essential to ensure the accuracy of the model, and this process is very time-consuming, especially in terms of data preparation and cleansing, and detailed model settings to ensure sufficient accuracy.
Key impacts:
- The workload of analysts in data collection and model building
- Difficulty in sustaining implementation in resource-constrained organizations
Countermeasures/alternative methods:
- Automating data collection and model building processes
- Simplifying the model by narrowing the scope of analysis
Depends on the expertise and knowledge of the analyst
Because MMM uses statistical methods to estimate the effectiveness of marketing activities, the reliability of the model and the interpretation of the results depend heavily on the analyst's statistical expertise and marketing knowledge. If the analyst lacks specialized knowledge in either statistics or marketing, there is a high possibility that the model design, variable selection, and interpretation of the analysis results will be incorrect. In addition, not only the analyst, but also the recipient of the analysis results may not be convinced of the results if they lack an understanding of statistics.
Key impacts:
- This could lead to incorrect model design and interpretation of analysis results.
- Doubts arise about the reliability and interpretation of analytical results, making it difficult to explain (and therefore take action) to management and other departments.
Countermeasures/alternative methods:
- Hiring data scientists with specialized knowledge and experience
- Stakeholder training and awareness-raising activities
Limitations regarding structural difficulties
Relying on past data
Because MMM utilizes past performance data, it tends to be unsuitable for unprecedented business or marketing activities, or for making future predictions in rapidly changing markets, as past data alone may not be sufficient to make predictions.
Key impacts:
- Unable to analyze new markets or new products
- Missing a trend
Countermeasures/alternative methods:
- Regular and short cycles of model updating and retraining
- Active use of external data (market research, industry benchmarks, norm values, etc.)
Lack of real-time and agile capabilities
MMM builds predictive models of business outcomes based on past data and assumptions, which makes it difficult to immediately adapt to rapidly changing market trends and changes in consumer behavior, and it is not adept at speedy analysis such as real-time data analysis.
Key impacts:
- Delayed response to the latest trends and unexpected events
- Difficulty in using the system when quick decision-making is required
Countermeasures/alternative methods:
- Regular and short cycles of model updating and retraining
- Use with real-time data analysis tools
Overemphasis on short-term effects
MMM often uses business results such as sales and sales volume as indicators, which can make it difficult to capture the impact of marketing activities that take more time, such as the formation of long-term brand value and customer loyalty. However, in recent years, MMMs with analysis functions that include brand equity (XICA's MMM service "MAGELLAN") have emerged, and they have evolved to enable analysis that complements not only short-term effects but also long-term effects.
Key impacts:
- This could lead to decisions that are biased towards short-term business results.
- The importance of long-term brand strategy may be underestimated
Countermeasures/alternative methods:
- Long-term indicators (such as customer lifetime value) are also evaluated.
- Consider qualitative data such as brand tracking surveys
Difficulties in integrated cross-channel and omni-channel analysis
Traditional MMM excels at analyzing the individual effects of multiple channels and is good at measuring the direct impact of each channel. However, it is difficult to fully consider the synergistic and indirect effects between channels, so it is difficult to capture the overall impact caused by the collaboration of multiple channels and the complexity of consumer behavior in cross- and omni-channel strategies. This makes it difficult to correctly capture the effect of upstream and awareness-related measures on business results.
Key impacts:
- Difficult to understand the overall effectiveness of a cross- or omni-channel strategy
- Unable to accurately analyze the behavioral patterns of customers using multiple channels
Countermeasures/alternative methods:
- Use in conjunction with other analysis methods, such as multi-touch attribution (MTA) analysis
- Adoption of more sophisticated statistical models (Hierarchical multiple regression analysis,Path Analysis,Bayesian Networks, etc.)
Difficulty in conducting detailed analysis of digital initiatives
In digital marketing, where there are a wide variety of channels, target segments, creative types, banner sizes, etc., the amount of data and variables to handle increases, making MMM more likely to experience problems due to high correlations between variables such as multi-correlations.
Key impacts:
- The analysis becomes more complex, making it difficult to accurately assess the individual impact of each variable.
- Difficult to fine-tune and optimize digital initiatives
Countermeasures/alternative methods:
- Aligning analysis policies, such as aiming for overall optimization rather than individual optimization
- Use in conjunction with analysis methods specialized for digital marketing, such as multi-touch attribution (MTA)
Other limitations
Unable to explain the "why" of results
Understanding the reasons for the success or failure of marketing activities is crucial for developing future strategies, but while MMM can quantify the effectiveness of marketing activities, it cannot provide a detailed explanation of why each initiative or campaign was successful or unsuccessful.
Key impacts:
- Difficult to come up with specific improvement measures
- Limited creative marketing strategies
Countermeasures/alternative methods:
- Use in conjunction with qualitative research (e.g., customer interviews)
- Use with CMM (Consumer Mix Modeling)
No analysis of customer experience
MMMs do not typically incorporate data on customer experience, which is an increasingly important factor in measuring the effectiveness of marketing efforts. Not taking into account customer feedback and engagement data can leave the overall evaluation lacking.
Key impacts:
- It is difficult to consider qualitative factors such as customer satisfaction and loyalty.
- Difficult to measure impact across the entire customer journey
Countermeasures/alternative methods:
- Use in conjunction with qualitative research (e.g., customer interviews)
- Use with customer experience (CX) analytics tools
Alternative and integrated approaches to complement the limitations of MMM
Below are some methods that can be combined with the MMM mentioned above to enable more comprehensive marketing analysis.
A / B test
A/B testing is a method of comparing two or more ideas to determine the most effective one.
merit:
- Low cost of operation
- Allows for rapid experimentation and learning
How to integrate with MMM:
- Conduct A/B testing on important channels and elements identified by MMM
- Use the results of A/B testing to adjust the model
Related articles:Analysis method for visualizing marketing effects: A/B testing'
Multi-Touch Attribution (MTA)
MTA is a methodology that considers multiple touchpoints a customer has before making a digital purchase and evaluates the contribution of each touchpoint.
merit:
- Enables detailed measurement of digital marketing effectiveness
- Track individual customer behavior
How to integrate with MMM:
- Understand the overall effect with MMM and analyze the effect of detailed digital initiatives with MTA
- Compare the results of both and check for consistency
Related articles:Comparison of the mechanisms of the representative analytical methods MMM and MTA'
Qualitative research
Qualitative research such as customer interviews and focus groups is a method of deriving suggestions and insights that cannot be visualized in numbers.
merit:
- Understand customers' true feelings and potential needs
- Get answers to your "why" questions
How to integrate with MMM:
- Reviewing the MMM analysis results from a qualitative perspective
- Verifying hypotheses obtained through qualitative research using MMM
Related articles:The science of marketing: How to use research design and statistical analysis'
Customer Lifetime Value (LTV) Analysis
LTV analysis is a method for predicting and evaluating the value a customer brings over the long term.
merit:
- It is possible to formulate strategies that take into account long-term customer value
- You can make investment decisions by comparing customer acquisition costs
How to integrate with MMM:
- Evaluate the short-term effect of MMM and the long-term value of LTV in combination
- Incorporate LTV information by customer segment into the model
Consumer Mix Modeling (CMM)
CMM is an analytical method that uses consumer awareness data to clarify the consumer attributes that should be targeted and the marketing activities that should be prioritized (4P, customer experience, etc.). By clarifying the factors that have a major impact on purchasing behavior, it is possible to prioritize investments to create results.
merit:
- Enables detailed analysis based on consumer awareness data, purchasing behavior and preferences
- Can provide insights into long-term customer acquisition and retention
How to integrate with MMM:
- Combine overall marketing effectiveness quantified by MMM with detailed insights based on consumer behavior from CMM
- Reviewing the MMM analysis results from a qualitative perspective
Related articles:What is CMM (Consumer Mix Modeling)?'
Summary
This article provides a detailed overview of traditional MMM, as well as its limitations and areas in which it is not strong. MMM is an effective method for statistically analyzing the effectiveness of marketing activities, and is an essential method for many companies. However, understanding its limitations and challenges and taking appropriate measures is the key to maximizing its value.
Specifically, it is effective to combine and integrate other analytical methods such as A/B testing, multi-touch attribution (MTA), qualitative research, customer lifetime value (LTV) analysis, consumer mix modeling (CMM), etc. By utilizing these methods, it will be possible to optimize marketing strategies and make more effective decisions.
XICA has over 10 years of experience in providing services in MMM and has supported over 250 companies, mainly domestic enterprise companies. Our analysts and consultants with extensive expertise in a wide range of industries can help you overcome the limitations of MMM and use data science to help you make better decisions. If you have any questions or require more information, please feel free to contact us.Contact us.
Related services:
- "XICA's MMM (Marketing Mix Modeling) Service: MAGELLAN'
- "XICA's CMM (Consumer Mix Modeling) Service: COMPASS'
↓ Click here for a list of carefully selected articles to deepen your understanding of MMM
- What is MMM (Marketing Mix Modeling)? Explaining its features, procedures, examples, etc.
- Why you should use Marketing Mix Modeling (MMM) to measure the effectiveness of your advertising
- 10 Common Misconceptions About Marketing Mix Modeling (MMM)
- What are the limitations and areas in which MMM (Marketing Mix Modeling) does not excel?
- Latest Trends and Future Outlook of MMM (Marketing Mix Modeling) (2024 Edition))
- “Modern MMM (Marketing Mix Modeling)” Practical Guide for Marketers
Recommended articles for those who read this article
- Column
Latest Trends and Future Outlook of MMM (Marketing Mix Modeling) (2024 Edition)
- Column
10 Common Misconceptions About Marketing Mix Modeling (MMM)
- Column
What is CMM (Consumer Mix Modeling)? Explains the characteristics, implementation process, and use cases of this scientific approach to clarifying consumer behavior.