Marketing Mix Modelling

Hot Topics


Measuring granular media in marketing mix modelling

How do you manage advertising effectively in the digital world? Media fragmentation and new data protection rules are probably the biggest challenges facing advertisers.

A new study by our Managing Partner Prof. Dr. Alexander Preuss with Antonio Prada from Snap Inc. shows how we can overcome these hurdles. By combining traditional statistical methods with a newer technique called ridge regression, we have developed an innovative two-stage approach:

First stage: using the traditional ordinary least squares (OLS) method for stable results controlled by statistical significance. Digital channels are considered as a single unit.

Second stage: Use of ridge regression to precisely analyse individual digital channels. This technique helps to deliver stable results despite many variables.

Third stage: The results are calculated and the total variable for the digital channels is split using the coefficients of the ridge regression.

📊 Results and advantages:

Stable results: The model remains stable, even with many variables.

Precise measurements: The impact of individual granular media channels and their return on advertising spend (ROAS) can be measured.

Adaptability: Universal use for different products, brands and target KPIs such as sales or brand parameters.

Our study offers valuable insights and practical solutions for companies looking to optimise their marketing strategies in an ever-changing advertising landscape.

Curious to find out more? Then click here for the full study in our whitepaper!

Cookie Consent with Real Cookie Banner