Commentary

It's Time To Reacquaint Yourself With Media Mix Modeling

The following was previously published in an earlier edition of Marketing Insider.

A year ago, MMA Global released a study revealing that 89% of marketers saw quality data as a challenge to implementing multitouch attribution (MTA). Between the death of cookies and recent and impending privacy regulations, MTA looks less viable than ever for performance marketing.

One perhaps surprising contender to fill the void: media mix modeling (MMM), an analytics practice that long predates MTA and is emerging as a less clunky, more usable tool for today’s marketers in predicting which channels will contribute to a more efficient holistic strategy.

What is MMM?

If you haven’t heard much about MMM, you’re not alone. In brief, it’s an analytical system that ingests all kinds of online and offline marketing data, including campaign performance, sales or promotions, performance history and seasonality, etc., and uses that data to produce insights on how specific marketing campaigns affected KPIs in focus.

It works at scale, not on a user level, and as such doesn’t rely on cookies or unique identifiers to be accurate. Vendors range from enterprise-level (Nielsen, Neustar, etc.) to more high-growth-friendly options like Pecan.ai and Channel Mix, and more (like Adobe) are entering the fray as the idea gains momentum in the industry.

An important note about today’s MMM compared with the MMM of yesteryear: AI and machine learning tools have made the practice faster, more nimble, and much more useful for businesses of all sizes.

Why should you consider MMM?

Whether you’re on the agency side or the in-house side, you’ve likely had conversations about how to get out from under the ever-increasing engagement costs of Google and Meta (and even LinkedIn if you’re heavy in B2B).

Those conversations are important. Most brands these days are overinvesting in Google and Meta, even as their engagement costs climb and/or signal quality continues to lag in the wake of iOS14. And MMM can give you very clear signals pointing you to more efficient ways to spend your budget.

Advertisers who remain stuck in place on the analytics front, relying on walled-garden data and more rudimentary attribution models like last click, will resign themselves to a fate of inflated CPAs and investment over-indexing, which could lead to huge challenges with the parameters of future privacy regulations.

What’s the best way to test the waters of MMM?

If you’re hesitant to commit to a full pivot to MMM, good. It’s best practice to test different solutions and models to find out which works best for your business. Besides that, you need to make sure you have the right resources on hand to take advantage of MMM’s capability. That means you need:

-- at least one to two analytics resources, whether they’re in-house or sit within your agency partner, to make sure the data is flowing in and out with actionable takeaways

-- a full creative team (in-house or agency) to create and adapt multimedia assets for your evolving mix of channels.

MMM isn’t perfect. The lack of user-level data presents challenges with smaller audiences, and the discipline is still anything but plug-and-play. But I strongly recommend getting at least familiar with its capabilities. As more and more big companies develop these kinds of solutions, the space will advance quickly and become viable for brands of every size and vertical.

Next story loading loading..

Discover Our Publications