What is Marketing Mix Modeling?

What Is Marketing Mix Modeling or MMM?

Marketing Mix Modeling (MMM) relies on statistical methods, namely multivariate regression analysis. It uses as input sales and marketing time series data to estimate the impact of various marketing tactics (marketing mix) on sales and then forecast their impact on future sets of tactics. It is often used to optimise advertising mix and promotional tactics respecting the sales revenue or profit.

What is Customer-segment Mix Modeling?

CMM is based on the idea of segmenting the market to groups of customers who share similar behavior and reactions to marketing activities. Each one of the obtained segments need to be modeled to get detailed and very specific customer insight. This helps to get more accurate data-driven decisions. By focusing on the consumer segments, marketers will be able to have more targeted messages for each segment and therefore, embracing the customer first approach. Subsequently, this will lead the business to optimize future campaigns on the medium, short and long term.
CMM works at an individual consumer level, it is more actionable; more capable of measuring consumer relationships
with brands; better suited to measuring digital media; and tailored to an increasingly addressable advertising future.
In other words, CMM can measure marketing RoI at more sophisticated levels and is more actionable for the marketer because it can measure marketing RoI at an audience level and therefore direct targeting decisions.

What is the difference between Media Mix Modeling and Marketing Mix Modeling?

Media mix modelling is an analysis technique that allows marketers measuring the impact of their advertising campaigns based only on media variables to determine how various elements contribute to the evolution of the KPI.

Marketing Mix Modelling on the other hand, includes other variables such as promotion, price, product changes…

When it comes to modelling your KPI, we recommend to include as much variables impacting your KPI as possible so that will give you a holistic and rounded view on what’s really driving your performance.

Isn’t Marketing Mix Modeling just for large enterprise brands?

No its not. Regardless of their sizes, all companies will benefit from adopting MMM. Marketing Effectiveness helps your brand maximize on return and get a deeper understanding of how your business actually functions, what channel works best and what not, and from these insights one can optimize their budgeting plan, maximize their revenues while minimizing their spend. And that can bring more benefits to small entreprises to raise their brand awareness and sales, thus growing the company.

MMM is destined to any media and analytics agency or advertiser no matter the size of the company. The only condition for conducting a MMM analysis is to have enough data (Usually a minimum of 2 years historical data. The data must also be reviewed and validated to be sure of its consistency and accuracy.

How to Measure Marketing Effectiveness?

Marketing effectiveness could be measured through Marketing Mix Modeling (MMM). MMM is an advanced analytical technique generally based on regression analysis that enables marketers to:

  • Measure ROI/MROI and impact of their media and marketing efforts
  • Optimize budgets across channels and campaigns
  • Simulate the outcome of various media and marketing plans.

According to Gartner, measurement methods based on Marketing Mix Modeling Mesaurement can yield 20% to 30% improvement in the efficiency of marketing spending, primarily by optimizing media (Gartner 2016).

MMM is based on applying regression techniques to historical time series to study the correlation between different independent variables ( e.g. Media, Marketing, Sesonality, Competitiob etc.) and a chosen KPI e.g. Sales, Conversion etc.
The Marketing Mix Modeling scenery is changing fast with companies of all sizes showing interest in adopting MMM techniques to measure the true impact of their marketing activities and optimizing their budgets. This need is fueled by the rise in spending, the abundance of media channels and the good PR around the adoption of AI and Data Analytics techniques in the media industry.

What are the Marketing Mix Modeling limitations?

Marketing mix modeling can fall victim to a number of limitations. Today’s marketers need insights into a variety of elements across their marketing ecosystem, including:

  • Person-level, behavioral data
  • The impact brand authority has on marketing spend and campaign optimization
  • The impact of creative messaging across channels
  • Key times to send marketing messages

      How complex is Marketing Mix Modeling?

      Marketing Mix Modeling, as a concept, is relatively easy to understand. It is an analytical solution that help marketers to understand and simulate the effect of advertising (volume decomposition), and to optimize tactics and budget spendings.
      Even though the concept is easy, modeling and interpretation needs to have advanced analytics skills and the right tools to accomplish insights extraction and strategic path shaping. MassTer was developed in a way that will lessen the burden and make the process a lot easier; as it is extremely user friendly and does not need any programming skills. The full MMM process is integrated within MassTer to ensure a centralized and easy view on the MMM models.

          What is the minimum data required to perform a Marketing Mix Modelling project?

          You need to have at least sales data per month and if possible by week. Ideally, 3 years’ worth of data is desirable. If not possible you can do with 2 years. You also need to have data about the main campaigns run and the investments in the different channels. Marketing data (price, promotions, distribution) is also desirable to have.

              What is the CRISP-DM methodology?

              CRISP-DM stands for Cross-Industry Process for Data Mining. The CRISP-DM methodology provides a structured approach to the planning of a data mining/data science project. Below are the 6 phases of the process:

              1. Business understanding

              2. Data understanding

              3. Data preparation

              4. Modelling

              5. Evaluation

              6. Deployment