Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

Optimizing ad spend with AI-powered marketing mix modeling

AI-powered marketing mix modeling is revolutionizing ad spend optimization, offering real-time insights and boosting efficiency like never before. Here's what marketers need to know.

Optimizing ad spend with AI-powered marketing mix modeling

[goodluz / Adobe Stock]

Profile photo of Michael True

Few technologies have merged into the everyday lives of people around the world as rapidly as artificial intelligence (AI) has in the past few years. It’s like we all just woke up one morning to find AI on every device, in every news story, and dominating any conversation remotely related to technology. “Is that AI?” has become a common phrase for people of all ages, and AI enhancements within apps are as common as prepopulated search suggestions.

While AI has worked its way into every segment of business operations, perhaps none have felt its disruption more than marketing. And it’s no wonder—marketing is where most content is developed, and although AI is a very broad field of science, generative AI (GenAI) has so far been the star of the show. It creates stuff, and even if they don’t fully understand the science behind it, it is easy for people to grasp. Provide a prompt and the result magically appears.

There is another aspect of AI in marketing that has been overshadowed by its more popular GenAI sibling, yet could have a far greater impact on business performance. They may not generate the same buzz as more visible AI tech, but advancements in machine learning (ML) and predictive analytics could truly change the world. 

Paid media often consumes the largest portion of marketing budgets, especially for consumer brands, making it the area that faces the most scrutiny. Are ad campaigns working? Which parts are working best? How can we make them work better? These are questions we’ve been asking for decades, but getting clear answers has never been particularly easy.

Sure, measurement technologies have improved some over time. The massive amounts of data that became available when people started living their lives connected to the internet opened up a whole new universe of potential insights. But, as marketing channels continue to multiply and tech companies continue to tweak pixels and privacy policies, complexity has become the enemy of efficiency. Mapping customer journeys and accurately tracing revenue back to its origins in this vast digital landscape is a herculean task.

Marketers now have access to information from innumerable data sources, but few can confidently claim they understand what it says. While piecing together data from various different platforms and tools might offer some valuable insight, getting it right is incredibly difficult. Reconciling reporting discrepancies and factoring in all the potential variables requires time, resources, and a grasp of statistical concepts that most marketers simply don’t have. In my opinion, solving this spend optimization challenge is where AI will ultimately have the most profound impact on the future of marketing.

Marketing mix modeling (MMM) is not a new concept. It’s having a resurgence of sorts—for reasons I’m about to go into—but the first MMMs were in use as far back as the 1960s. MMMs have always been decent at measuring high-level performance over longer periods of time, but they were not ideal for more granular measurement and ongoing spend optimization. MMMs of the past were also generally expensive, cumbersome implementations, making them a tool mostly reserved for big businesses with deep pockets.  

Today, progress in computing power and AI technology has unlocked a new generation of MMMs that are more suited to the needs of modern marketers. MMMs anchored on advanced machine learning models can generate results faster and at much more granular levels, meaning marketers can use the insights to optimize spend between channels, tactics, and even campaigns on a daily basis. The increases in ROAS (return on ad spend), efficiency, and revenue generated by brands that are using AI-powered MMMs suggest a promising future for the modernized technology. 

Imagine a scenario where a tool automatically assigns every dollar of ad spend to the combination of campaigns that will generate the most revenue, then continuously identifies how to optimize that spend allocation in real time. It’s not that far off. The expertise and knowledge that marketers bring to the table can never be replaced, but I’ve also never met a marketer who wouldn’t benefit from gains in efficiency. Data-driven marketers are starting to realize the significant impact next-generation MMMs can have on performance. In fact, a recent study of U.S. marketers published by Snap and Emarketer shows that improving MMM is now their number one priority for upgrading marketing measurement. 

It’s that time of year when everyone is making their annual predictions about what’s in store for the next 12 months. If the first half of the 2020s has taught us anything, it is that life is extremely unpredictable. But, I’m willing to go out on this limb, even without a predictive model to back me up. AI momentum is bringing new meaning to marketing mix modeling, so watch this space. How brands optimize ad spend will look very different at this time next year.