Media Mix Modeling: The Secret Recipe
By: Pau Mitjà Gilabert, Insights Analyst at Validators.
Marketing budgets are growing. So are the number of ways you can spend that budget. Whereas in the past you mainly chose between TV, radio, and perhaps a little print, you now have to deal with:
- 7 social media platforms
- 4 video formats
- Search (branded vs. non-branded)
- Display and programmatic
- Influencers
- Retail media
- Owned channels such as CRM and email
And now the question is: What really works?
Why more channels make the problem worse
More channels = more opportunities, right? Absolutely. But in practice, it mainly makes one thing harder: keeping track of everything.
Effects overlap. Someone sees your ad on TikTok, then on TV, Googles your brand, clicks on a search ad, and makes a purchase three days later via a retargeting banner. Congratulations. But… which channel gets the credit? Search closed the deal. But social started this story. And TV… it was there too. Probably important. I think. So how do you figure this out?
This is where Marketing Mix Modeling (MMM) comes into play
MMM helps you see through this complexity. Instead of analyzing individual channels, MMM looks at the big picture:What impact do all my marketing activities have on my KPIs when taken together?What happens to your sales, orders, and traffic when you invest in various marketing activities over time, across channels, and across the board? But what does “looking at the big picture” actually mean in practice?
Think of your media mix as a pie
As my colleague Marco Vaalburg, a data scientist at Validators, puts it so well: Your marketing mix is like a pie.
- The ingredients = your channels (TV, social media, search, etc.)
- The amounts = your budget allocation
- The appeal = the impact on your business (sales, traffic, conversions)
Essentially, MMM works like this:It tests (statistically) a wide variety of “recipes” to discover which combination works best.So: what happens if you shift your budget from social media to TV? Or invest more in search at the expense of display? Until you arrive at the best possible cake.
Why Everyone Is Suddenly a Fan of MMM
Interest in MMM is growing rapidly. This is no coincidence. Tracking is becoming more difficult because you’re missing out on more and more data due to cookie restrictions. Customer journeys are becoming more complex (e.g., someone sees you on social media, searches for you later on Google, and then converts there). Attribution models treat marketing as if it were a relay race, even though everyone is running at the same time. MMM offers a solution here because it:
✓ Does not rely on individual user data
✓ Operates at the aggregate level
✓ Provides insight into long-term effects
For many brands, this makes it one of the few ways left to truly gauge marketing impact. But… is MMM the holy grail? Not quite. Traditional MMM also has a few clear drawbacks:
x Slow (analyses take months)
x Complex (requires a lot of data and expertise)
x Difficult to implement in practice
And perhaps even more importantly: It often gets stuck at the insight stage, without any concrete decisions. You know you could make a better cake. But you still don’t know exactlyhow.
A more pragmatic view of MMM
Traditionally, MMM focuses primarily on media spend. Simply put: how much reach you have and how much you invest.
Or, to put it even more simply:Impact = Reach × Cost
It’s a simple formula. The more people you reach and the more you invest, the greater the impact. But in practice, something striking becomes apparent: two campaigns with the same budget and the same reach can yield completely different results. Why?
Because content makes all the difference
Not every message is equally impactful. Some campaigns stick with you. Others are skipped after two seconds. Some are convincing. Others get lost in the noise. This proves that strong creative is a crucial factor—one that is often overlooked in traditional MMM.
That’s why we’re adding a third variable: content. By incorporating content into your model, you can simultaneously optimize both where your budget goes and what you show. And that’s exactly where the biggest gains lie.
Where:
- Reach = how many people you reach
- Costs = what you invest
- Content = how effective your message is
This is important. Without content, you’re mainly optimizingwhereyour budget goes. With content, you’re also optimizingwhatyou show.
Image: Reach x Cost x Content = Effect
What are the concrete benefits of MMM?
By using MMM in this way, you shift from insight to action:
- Smarter budget allocation: Which channels are truly effective? Where are we over- or underinvesting?
- Understanding the long term vs. the short term: What drives conversions, and what builds brand equity?
- Understanding what content really does: Which themes or messages work?
- More robust evidenceNo more gut feelings—only data-driven decisions.
The core
More channels mean more complexity. But they also mean more potential—if you approach it the right way. MMM helps you cut through the chaos, identify the right connections, and make better decisions. So you’re no longer just throwing ingredients together at random…
But we're working hard to create the best cake possible.
Curious about Media Mix Modeling at Validators can do for your media mix? We’d be happy to walk you through it.
