MMM is a tool to predict results. It uses regression models to forecast sales, volume, and ROI.

The dependent variables (the ones we want to predict) are clear: sales volume and value. The independent variables, not so much. They are not clear because it is hard to figure it out which marketing variables are impacting sales.

We have the ability to collect and the machine power to process massive amounts of data. The statistical models have been with us for decades. What’s challenging it is not the math part, is the art part of knowing what is influencing the buyers the most.

Because this is such a complex analysis, there aren’t many tools available in the market. Neustar Customer Analytics and Nielsen MMM are the main ones.

They won’t cover the “art” part of the analysis. If you give the tools business results data, how much you’ve spent per channel (offline and online), in a time-series format and at an aggregation that makes sense, the tools will find correlations. They can tell you to spend less on Google and spend more on Facebook for example.

The creative analysis, the consumer cohort analysis, the “whys,” that’s an entirely different problem you need to address separately.

leocelis

Hi! My name is Leo Celis. I’m an entrepreneur and Python developer specialized in Ad Tech and MarTech.

read more