Media Mix Modeling โ€“ A Fresh Approach

If you read our last post on Media Mix Modeling (MMM), you have a basic understanding of what the methodology is, and why agencies and private companies alike are implementing this approach. This is our second post on the topic, and we will share a real life example along with information on what is needed to implement MMM. We know that any complex modeling can be as confusing as it is beneficial, so if you need some assistance wading through the options that can best help you understand the value of PR and communication in your firm, reach out to a trusted expert on the topic.

MMM vs. Data Driven Attribution

Itโ€™s important to remember that MMM exclusively measures the impact marketing efforts have on meeting certain objectives. MMM does not focus on the consumer journey, like many attribution models. MMM doesnโ€™t account for user-level engagements, whereas data-driven attribution models will measure impressions, clicks, and other user activities.

Data-driven attribution โ€“ tracks engagements through the customer journey

MMM โ€“ provides high-level insights into specific marketing tactics over a longer period of time

Each type of modeling has its own benefits and drawbacks. Both attribution and MMM have some blind spots. Namely, attribution models have limited transparency when it comes to offline conversions, which means mainly digital channels will be measured. On the other hand, MMM can measure that information, but also lacks individual channel-level insights.

MMM allows high-level insights into campaign effectiveness

One of the earliest adopters of this model was Kraft. They used this type of analysis for their launch of Jell-O products decades ago. They began by choosing three or four television networks and magazines to promote the new products. Leveraging MMM, they were able to investigate variables like advertising at different levels, in different parts of the country, or at different times of the year. They were able to get a clearer picture of how pulling those levers would affect sales in given regions at given times. For example, they could choose to advertise Jell-O products in 10 cities over 5 weeks and see if sales increased.

Interested in implementing MMM?

Itโ€™s important to note that for MMM to be considered effective, it needs to be aggregated with further marketing measurements (like multi-touch attribution). This is the only way to provide a unified and accurate measurement.

That means that companies who hope to use MMM will need a comprehensive marketing measurement platform. This will help to distill big data into actionable insights and allow for dynamic, in-campaign changes and optimizations. A platform that allows for MMM functionality alongside analysis from other models will have the following features:

  • Visibility into in-campaign insights
  • Data integration across all marketing channels and efforts
  • Transparency into individual channel-level insights along with historical trends
  • The ability to analyze the effectiveness of branding or creative messaging elements

Wondering if your marketing measurement platform fits the bill?ย 

Comment here