More Than Data

We Know Marketing Mix Modelling Isn’t Perfect—But At Least We Can Improve It

More Than Data
Mar 10, 2025By More Than Data

If you've ever browsed Reddit, you might have come across posts questioning whether Marketing Mix Modelling (MMM) is flawed. Take this discussion, for example: Reddit Thread. To be honest? We couldn’t agree more. Yes, MMM has issues. It’s not perfect. But here’s the real question—what’s the alternative?

If Not MMM, Then What?

Let’s explore the other options marketers have for measuring effectiveness:

1️⃣ Multi-Touch Attribution (MTA): Once hailed as the gold standard for digital attribution, MTA allowed marketers to track the full online customer journey across multiple digital touchpoints. But here’s the problem—third-party cookies are disappearing, making cross-platform tracking nearly impossible. Yes, you can still use authenticated user IDs, but that only works within your own digital properties (your website and apps). Another major limitation? MTA only works for digital channels. What about offline media—TV, radio, out-of-home (OOH), cinema, print? MTA won’t help you there.

2️⃣ Geo Lift Studies: This method, which is like A/B testing on a geographic level, can provide insights by comparing test and control markets. But let’s be real—finding a true control market (one without any ads) is nearly impossible. If you can’t find one, you’ll have to simulate it, and that comes with its own set of challenges. In short, Geo Lift can be useful but is limited to single-channel experiments, making it hard to scale across a full marketing mix.

MMM Is Imperfect—But It’s Still Valuable

So, where does that leave us? Despite its flaws, MMM remains one of the most useful tools for marketers. Why? Because having a structured, data-driven approach to evaluating media performance is always better than making decisions based on gut feeling.

MMM helps brands and media agencies understand how their past campaigns performed, which channels contributed the most, and how they should allocate budgets in the future. No, it’s not flawless. Yes, it makes assumptions. But would you rather have some insight or no insight at all?

The Biggest Challenge With MMM

One of MMM’s biggest weaknesses is its foundational assumption: that every marketing, media, product, pricing, promotion, distribution, and external factor directly impacts the business KPI (e.g., sales). But in reality, many factors have indirect effects.

Take MTA as an example—each user’s path to purchase is unique. Some users need multiple touchpoints before converting, while others make a purchase immediately. Some paths lead to conversions, others don’t. But just because a touchpoint didn’t directly lead to a sale doesn’t mean it had no influence.

MMM, in its traditional form, flattens all these interactions into a single regression model, missing the complex web of cause and effect in real-world marketing.

What’s the Solution? Unifying MMM and MTA

Instead of choosing between MMM and MTA, why not combine their strengths?

At More Than Data, we’re taking MMM beyond traditional regression models. We’ve developed a causal inference modelling approach that integrates both direct and indirect impacts across online and offline media channels.

Think of it this way:

  • MTA focuses on individual digital journeys, mapping how users interact with multiple online touchpoints before converting.
  • MMM provides a high-level view of how marketing investments impact business KPIs.
  • Our causal inference model connects these dots, recognizing that media influences are not linear—they work together in a complex network of interactions.

Rather than treating digital and offline channels separately, we measure them within a unified framework. This approach provides a more realistic, holistic view of media effectiveness.

Pushing the Boundaries of Marketing Measurement

MMM isn’t perfect, but instead of dismissing it, we should improve it. That’s exactly what we’re doing at More Than Data. By integrating MMM and MTA into a single, more advanced framework, we’re redefining how marketers measure success.

Marketing is evolving. Measurement should evolve too.