At Media Merchants, we’ve always operated on a simple principle: every media dollar should be measurable and accountable. It’s easy to say, but in today’s advertising landscape, it’s getting harder to deliver. That’s why we’ve started building marketing mix models using Google’s Meridian – and it’s already changing the way we think about measurement for our clients.
The attribution problem
If you’re running media across multiple channels, you’ve probably felt this tension. Your Google Ads dashboard tells one story. Your Meta reports tell another. Your TV buyer has their own reach estimates. And none of them agree on what’s actually driving the phone to ring or the website to convert.
That’s the attribution problem, and it’s only getting worse. Third-party cookies are disappearing. Privacy regulation is tightening. Walled gardens are limiting what you can see. The gap between what platform dashboards report and what is genuinely driving growth has widened to a point where relying on any single source of truth is a risk.
This is exactly the challenge that marketing mix modelling is built to solve. And with the release of Google’s open-source Meridian framework, it’s now more accessible than ever.
What is Google Meridian?
Meridian is an open-source marketing mix modelling framework built by Google and released in early 2025. In simple terms, it’s a statistical model that measures the true incremental impact of your marketing across every channel — not just digital, but television, out-of-home, radio, and everything in between.
What makes it different from the attribution tools most advertisers are used to? A few things. Meridian doesn’t rely on cookies or click-based tracking, so it’s built for a privacy-first world. It uses geo-level Bayesian modelling, which means it analyses how marketing activity in different regions correlates with business outcomes, rather than trying to track individual users. And it integrates Google’s own search query data, which gives a much more accurate picture of how channels like paid search and YouTube are contributing to results.
The end result is a model that can tell you, with real statistical rigour, how much each channel in your media mix is genuinely contributing to your business outcomes – and where your next dollar would be best spent.
Why this matters for our clients
Historically, marketing mix modelling has been the domain of large holding-company networks and specialist econometric consultancies. It’s been expensive, slow, and largely out of reach for most mid-market advertisers. Meridian changes that. Because it’s open-source and built on modern data science infrastructure, we’ve been able to build and deploy an MMM in-house for one of our major clients — giving them a level of media accountability that simply wasn’t available to them before.
In practice, that means we can now move beyond siloed channel metrics and show our client exactly where their budget is working hardest, where there’s diminishing return, and where reallocation could unlock growth. It’s scenario planning, budget optimisation and cross-channel accountability rolled into one.
As our Managing Director, Michael Gee, puts it:
“The reality is that traditional attribution is broken. Cookies are disappearing, walled gardens limit what you can see, and most platform reporting only tells you part of the story. We needed a measurement approach that could sit above all of that and show our client, with genuine statistical rigour, where their budget is actually driving results. Meridian gave us the framework to build exactly that.”
