Let’s be honest. For a while there, Marketing Mix Modelling (MMM) felt a bit like that reliable but slightly dusty sedan in a world obsessed with flashy new sports cars.
It was solid, it worked, but all the buzz was around its sleeker, more granular cousin, Multi-Touch Attribution (MTA). We were all hooked on the promise of tracking every single click, view, and touchpoint on the user’s journey.
Then, the ground started to shift. Privacy regulations got serious, browsers began waving goodbye to third-party cookies, and that perfect, granular view started to look… well, a little blurry. Suddenly, marketers everywhere are asking the same question: “Now what?”
You know what? The answer might just be hiding in plain sight. That reliable old sedan just got a brand-new engine, a high-tech navigation system, and a killer sound system. Marketing Mix Modelling is back, and frankly, it’s never looked better. This isn’t just a comeback; it’s a full-blown renaissance.
The elephant in the room: Privacy
The biggest reason for this resurgence is something we can’t ignore: the privacy-first internet. With regulations like GDPR and CCPA, and the steady disappearance of third-party cookies, the very foundation of individual-level tracking is cracking. Trying to stitch together a complete customer journey with MTA alone is becoming a game of connect-the-dots with half the dots missing.
This is where MMM strolls back onto the scene with a confident smile.
Here’s the thing: MMM has always been privacy-compliant by its very nature. It’s a top-down, statistical approach that looks at the aggregate impact of your various marketing and non-marketing drivers—like ad spend, promotions, seasonality, and even economic factors—on your sales or conversions over time. It doesn’t need to know who you are, specifically. It just needs to know that X amount of spend on YouTube ads, combined with Y amount of email campaigns during a holiday season, resulted in Z amount of revenue.
It’s no surprise that the likes of the Marketing Science Institute have noted a clear “resurgence” in MMM as companies scramble for measurement solutions that actually work in this new reality.
This isn’t the same old MMM
Okay, so it’s privacy-friendly. But what about the old complaints? Wasn’t MMM famously slow, incredibly expensive, and a bit of a “black box” that required a team of data scientists locked in a basement for six months?
Yes, it was. But that was then.
Modern MMM has been completely supercharged by two game-changers: AI and cloud computing. The complex models that used to take months to build and run can now be processed much, much faster. This means we can get insights in near-real-time, not just once a year. It’s the difference between getting a printed road atlas in the mail and using Google Maps.
And speaking of Google, the biggest names in tech are putting their weight behind this revival. Meta released its open-source MMM tool, Robyn, and Google recently launched Meridian. When giants like these offer up free, powerful tools, it’s not just a trend; it’s a major strategic signal. They are telling the market that this is the future of large-scale measurement. This move has democratised access to powerful modeling, taking it out of the exclusive domain of Fortune 500s and putting it within reach for more businesses.
Okay, but does it actually work?
All this sounds great in theory, but what about the results? Does this new-and-improved MMM actually move the needle?
The answer is a resounding yes. The proof is in the pudding, and the pudding is a big bowl of increased ROI.
Industry studies have shown that a well-implemented MMM strategy can improve marketing return on investment by a solid 10–30%. That’s not just loose change. We’re seeing real-world examples pop up everywhere. One global consumer goods brand, after running an MMM analysis, discovered that its social media ads had a much higher ROI than previously thought. They confidently increased their digital ad spend by 25% and saw a 15% lift in sales. Another major retailer used MMM to fine-tune their campaigns, leading to a 10% reduction in marketing costs while simultaneously boosting sales by 20%.
These aren’t just numbers on a slide; they are tangible business outcomes. It’s the kind of evidence that makes CFOs sit up and listen.
The power couple of measurement: Why not both?
Now, this doesn’t mean we should pack up MTA and send it off to a retirement home. That would be a mistake. The smartest marketers aren’t thinking in terms of “MMM or MTA,” but “MMM and MTA.”
Think of it like this:
-
MMM is your satellite view. It gives you the big picture. It shows you how all your channels—both online and offline—work together. It tells you how TV ads are influencing web traffic, or how a PR push is lifting in-store sales. It sees the whole weather system across the country.
-
MTA is your street-level view. It provides incredibly granular detail about the digital journey. It tells you which specific ad creative or keyword path is most effective for a particular segment. It’s the local forecast for your neighbourhood.
You need both! Google’s own Modern Effectiveness Measurement guidelines now recommend a hybrid approach that combines MMM for strategic budget allocation, data-driven attribution for tactical channel optimisation, and incrementality testing to validate it all. It’s about creating a holistic, validated view of performance. When you use insights from your MTA model as an input into your MMM, the whole system gets smarter.
Finding the right partner for the journey
So, the case for MMM is strong. But how do you actually get started?
You could try to wrangle an open-source tool like Meridian yourself, but as anyone who’s tried knows, that’s not exactly a plug-and-play experience. It requires significant data science know-how.
On the other hand, many “off-the-shelf” SaaS platforms can be too rigid, outlandishly expensive, and fail to account for the unique context and data ecosystem of your business.
This is where the idea of a partnership becomes so important. You don’t just need a tool; you need expertise. You need a team that can build a bespoke model on a powerful foundation like Google’s Meridian but customise it for your specific global needs and business questions.
This is exactly why we at XPON have invested so heavily in this space.
We’re not selling a rigid product. We offer a managed solution that combines the power of an open-source model with our deep expertise in data enablement and activation. It’s about building something that fits your business, not forcing your business to fit a tool. This approach allows for unparalleled customisation and ensures that the model isn’t beholden to a third-party product roadmap. Get in touch with our team if you’d like to learn more.
The resurgence of Marketing Mix Modelling is more than just a passing fad. It’s a necessary evolution driven by the biggest shifts our industry has faced in a decade. It’s about moving toward a more durable, privacy-safe, and genuinely holistic way of understanding marketing performance.