July 5, 2024
Marketing Mix Modeling and Artificial Intelligence: A Risky Alliance?
  • Data management and audiences

The world of Marketing Mix Modeling (MMM) is undergoing significant changes. Since the beginning of the year, several players (agencies and digital platforms) have positioned themselves with AI-based marketing mix modeling solutions. The new miracle solution to understand the impact of your media activities on your business results!

A solution that could cost you up to 40% of the revenue generated by your marketing activities, according to the founder of MeasureMonks, a firm specializing in the field.

Why? Because an AI model doesn’t know what it doesn’t know. Artificial intelligence cannot "see" what is not in the data (internal and external contexts). This skews the results. Every time a variable or an important context is missing, you get a distorted view of your marketing performance. And you make bad decisions based on these falsehoods.

Seventeen marketing experts signed an open letter published by the World Advertising Research Center (WARC), including Les Binet and Dr. Kite Grace, authorities in marketing and statistical modeling. They warn against using artificial intelligence as a solution for MMM. The human factor remains essential, if only to know how to use artificial intelligence. Thus, you always need a statistical expert to manage the modeling and ensure that the model is sound, and a marketing expert to correctly interpret the modeling results and make actionable recommendations.

Google Meridian is an "open source" platform. The quality of the models will depend on the expertise of your own resources, as Google offers no support for its AI modeling tool. This will cost you an expert in advanced statistics. I have also read analyses conducted by firms with modeling specialists (PWC, Nielsen). They produce incomprehensible reports for marketers, filled with various coefficients, without any marketing insights. Be aware that Omnicom Media Group supports you on both fronts with experienced statisticians and marketers, as evidenced by several dozen models developed from our offices in Montreal and Toronto over the past few years, whose effectiveness has always been demonstrated.

The new miracle solution may be tempting, but I would like to quote Nancy Smith, CEO and President of Analytic Partners: “Don’t be a guinea pig to inexperienced vendors.” Now it's my turn to offer you two options:

- On one hand, the choice of knowledge, a path that may be longer but will reflect your business reality.

- On the other hand, the comfort of a reality simulated by artificial intelligence, but which actually keeps you in ignorance.



Marketing Measurement | Marketing Effectiveness| Media.Monks

An open letter to marketing departments from experts on evaluating advertising | WARC

Marketing evaluation experts warn of dangers of AI (marketingweek.com)




We are steadfast in our belief that marketing and media creativity needs to be constantly redefined. Because we are emboldened by the ever-more meaningful, memorable and impactful connections this creates between a brand and its audiences. And most importantly, because it drives business success.