Mutinex on MMMs, MTAs And More

Published in Partnership With Mutinex

Following on our recent piece on Mutinex’s GrowthOS and the evolution of marketing mix modelling (MMM), we interviewed Henry Innis, CEO and co-foudner of Mutinex by email to explore the topic and his business further.

Henry, what’s your take on why MMM is better than MTA (Multi-Touch Attribution)? Or maybe more specifically, under what circumstances, if any, do you think MTA is superior to MMM?

I see MTA as a rule-based system that tries to document precisely which touchpoints occurred on a user journey, while MMM is fundamentally a math-based approach that aims to identify what actually caused an outcome. Think of MTA as a system that tracks every breadcrumb a customer leaves—clicks, views, interactions. It’s good at describing events but struggles to isolate how other factors (like pricing changes or economic shifts) influence those events. Because MTA looks at a narrow slice of the overall environment, it can become less valuable as privacy norms tighten and walled gardens make it harder to capture a complete user trail.

MMM, by contrast, brings a holistic lens. It pulls in everything from competitive activity to macroeconomic trends and brand spend, so it’s able to determine which elements genuinely drove results. Sure, MMM used to be slow and retrospective. But modern MMM solutions, including the platform we’ve built at Mutinex, have changed the game. We can process data quickly, iterate models more frequently, and get insights at a pace and granularity that older MMM approaches never could match. As a result, when you’re making big, strategic calls about budget allocation, MMM tends to be the more robust tool.

That said, MTA can still be powerful for tactical, user-level optimizations within a single channel. If you want to customize creative or refine ad placements based on individual user segments, MTA can show you which content is working within that limited scope. In other words, I’d let MMM drive the big-picture decisions—like how to split budget across different channels—and then rely on MTA to tweak the details inside each channel. Ultimately, I see these two as complementary rather than mutually exclusive, though in practice, privacy shifts mean MTA’s data pool is shrinking, and its limited perspective is becoming more evident.

Under what circumstances would you recommend neither gets used? Or maybe putting it differently, there are surely brands and businesses that never would have launched if they had gone through a modelling exercise. How do you think of the role of “gut” or “feel” in interpreting either of these approaches?

I’ve always viewed both MMM and MTA as optimizers. They’re designed to fine-tune or improve upon what you’re already doing, rather than provide your next big creative spark. If you’re a startup launching a brand-new venture with almost no historical data, modeling can be misleading. The danger is you might overfit early numbers or draw overly deterministic conclusions from tiny data sets and potentially shut down an idea that just needs room to grow.

That’s why “gut” or “feel” still matters—especially in new or highly creative endeavors where data simply doesn’t exist yet. There’s a point at which an entrepreneur or marketer has to rely on intuition to do something that defies existing patterns. However, once you have enough data to work with, MMM or MTA can help you scale those intuitions more effectively.

I believe the best results come when you merge the two: let your creative instinct guide you toward a bold idea, then use data-driven models to confirm and refine that idea. With a solution like our GrowthOS/MAITE AI at Mutinex, much of the heavy number crunching is automated, which means you can channel your creativity into the big decisions rather than getting lost in spreadsheets.

Let’s shift to the role of AI in your product. How do you distinguish between automating a relatively standard process and using your tools to do something non-standard or otherwise replacing “insight.” Are we there yet?

It’s easy to conflate fancy reporting with genuine insight. Most dashboards simply restate what happened—“X percent came from search,” or “Y percent came from social.” That’s not insight; it’s just data reporting. Real insight weaves together disparate data sources—like point-of-sale information, media spend, competitor activities, and economic indicators—to tell you not just what happened, but why it happened. AI’s role, as I see it, isn’t to replace human expertise but to remove the slog of filtering massive data sets so we can focus on the interpretation that drives meaningful decisions.

By pairing MMM with AI, we’re drastically reducing the time it takes to go from raw data to an actionable story. Instead of spending days or weeks setting up models, we can run them at scale and generate meaningful narratives within hours—sometimes even faster. This means marketers and executives can spend more time on strategic thinking and creative execution, rather than drowning in data prep.

So yes, we’re increasingly using AI in ways that are “non-standard” compared to the old approach of manual data cleaning and modeling. But it’s not about displacing human insight—it’s about amplifying and accelerating it. I believe that’s the real breakthrough: using AI to handle the volume and complexity, so the human mind can synthesize the results and make smarter, faster decisions that grow the business.

Thanks for that perspective. So why do you think your MMM product is better – or at least differentiated – vs. others on the market? Or maybe putting it differently, what kind of marketer or what kind of circumstances do you think Mutinex is best suited for? What kind of marketer or circumstance do you think you’re poorly suited for?

Yeah, so I'd be really careful on the word "better". There are lots of great MMM vendors out there I deeply respect - for example Analytic Partners, Transunion amongst others. They have built great, enduring businesses in the sector and clearly have built the category over many years. And they're fantastic examples for pioneering a category brilliantly and with high integrity, which has set it up to be adopted by marketers.

I think what we are trying to fundamentally improve is the customer experience of MMM. At the recent ANA Media Masters conference we ran a poll, which basically showed around ~3% of marketers felt their measurement solutions really went above and beyond. Measurement, to my mind, has really only been built for the data science department, which turns most of the amazing tools and techniques into internal consulting departments. That makes getting insight and answers at the moment you need it a painful product experience. Ultimately what do marketers want from measurement with their analytics teams? They want answers to questions in minutes, not weeks, and I think that's what gets marketers back in control of their budgets.

We need to solve for that problem. So everything in Mutinex is about getting a good, trusted answer fast with granular precision. If you're a marketer who has lots of questions and who wants to grow, then we're a great fit. I also happen to think we work really well with teams have have a degree of in-housing because of that speed philosophy. Conversely, if you're a marketer who really needs an internal transformation - bodies on the ground, deep experience and departments who are incredibly siloed then we aren't the right partner. The deep, white glove experience that other vendors will serve you incredibly well in this instance and that strong, external voice may be beneficial in navigating some of the politics. I believe faster answers solves most political problems, but it's also down to a preference of what most want and need.

For us, we're focused on constantly giving an incredible product experience around an answers product and providing support around that. For others, I think the focus is on providing an incredible service experience and then finding ways to insert some products into that experience. So the mindset is different in the companies and I think that anchors our different philosophical approach.