In January 2025, Google announced it was making Meridian — its open-source Marketing Mix Model (MMM) — available to everyone. Like all MMM platforms, Meridian uses statistical analysis to assess the impact of your marketing channels and marketing spend against your results. It then uses this data to model potential future scenarios, helping you allocate your budget effectively to achieve your advertising KPIs and maximize revenue.

Measuring the effectiveness of your marketing efforts is a longstanding challenge for marketers, so this is exciting news. An accessible MMM could lead to greater transparency and control over your budgets at a time when platforms like Performance Max are becoming more restrictive. 

Marketing Mix Models have traditionally been quite complex, limiting their use to brands who can afford to hire experienced data scientists. But Meridian is designed to simplify the MMM process, making its data more accessible to marketers. 

In the next five-or-so minutes, you’ll learn everything you need to know about Meridian, including how, when, and why to use it—plus its broader implications for performance marketers.

What is Meridian?

Meridian is Google’s open-source Marketing Mix Model platform. Unlike traditional MMMs, which have trouble incorporating digital marketing results into their models, Meridian gives you a holistic view of your marketing efforts. It combines several different elements — from channel type and ad spend to seasonality and current economics — to analyze your marketing success and help you make better budget decisions.

Importantly, Meridian doesn’t use cookies to track users or measure performance. Instead, Meridian aggregates anonymous data to offer accurate information while respecting user privacy. 

Similar to Meta’s Robyn, Meridian is open-source and customizable, making it easier to use than other third-party MMM tools. (It’s also free—you can download the code directly from GitHub.)

Meridian’s modelling features (and how they can help marketers)

Here’s a breakdown of Meridian’s key features, how they work, and why they’re important for marketers:

• Bayesian modelling — Meridian uses the Bayesian causal inference methodology, which combines prior knowledge of your business with learned data signals to estimate impact and quantify uncertainty. Knowing the level of certainty associated with your results can help you make better decisions, especially if you’re working with limited data. 
• Geo-level modelling — With Meridian, you can analyze the impact of your marketing efforts at a local or regional level and improve certainty for locations with limited data or prior knowledge.
• Understanding query volume in search — It’s not always easy to infer whether a change in results is due to higher ad spend or higher consumer demand. Meridian accounts for search query volumes as a confounder when analyzing search ads. 
• Includes return on investment (ROI) priors — You can set the ROI of each marketing channel as a model parameter in Meridian. This enables you to incorporate incrementality experiments, industry benchmarks, and other domain knowledge into your modelling.
• Non-impression based reach and frequency — Meridian can model any media channel's impact based on reach and frequency data, rather than simply using impressions (or any other single metric). This offers more precision in your impact estimates.
• Data dashboards — The results of your Meridian modelling are displayed in a dashboard, making the data easy to interpret and understand. Here are some examples from Google’s demo data:

‍

‍

Why should marketers use Meridian?

Meridian can help drive better business outcomes, from higher revenue to a better return on ad spend. But that’s not the only reason to use Meridian. Google’s MMM can also:

• Combine historical data with current information to create informed, accurate performance analysis.
• Measure cross-channel impact, giving you a clearer picture of attribution for all your different marketing channels.
• Help you adapt to a cookieless future, thanks to its privacy-first data aggregation systems.
• Restore control and transparency. Because it’s an open-source program, Meridian allows you to use it however you need to, rather than using black-box third-party solutions.
• Bring statistical analysis in-house, helping you reduce costs associated with hiring analytics companies or consultants.

Who is Meridian best for?

While Meridian is now open to everyone, it will work better for some advertisers than others. Consider using Meridian if:

• You have a high marketing budget — MMM works best when analyzing trends across significant spend levels.
• You invest in both online and offline media — MMM accounts for traditional media and digital marketing, so you can see the impact across all relevant channels.
• You have a good set of historical data — The more data you have, the more meaningful your MMM insights will be.
• You already use attribution models and platform analytics — A good grounding in analytics will help you make good use of your MMM insights. Meridian can complement and affirm insights from other systems and offer a greater range of comparison data.

While there’s nothing to stop you using Meridian if you have a smaller budget or you advertise exclusively online, your insights may not be as accurate or helpful.

Key use cases: when to use Meridian

Here are four potential use cases for when marketers can use Meridian effectively.

1. Allocating budget across channels

Meridian’s primary goal is to help marketers better allocate their budget. And with ROI as its top priority, Meridian could be an excellent tool for determining the most cost-effective mix of search, social, display, TV, and offline marketing efforts.

When you know which channels are most likely to be profitable, you can adjust budgets accordingly. It’s good practice to put your budget behind your best performing campaigns.

2. Evaluating campaign incrementality

Incrementality refers to the extra revenue generated by a marketing campaign above and beyond normal sales or baseline performance. Meridian can model the impact of specific channels and factors, giving you a clearer picture of campaign incrementality.

If you can confirm campaign incrementality with a reasonable degree of certainty, you can replicate or even improve on these results.

3. Analyzing trends, seasonality and competitor strategies

Because Meridian accounts for external factors as well as your marketing data, you can use its insights to analyze the impact of seasonal changes, the competitive landscape, and macroeconomic factors.

Understanding how these external issues have affected revenue previously can help you position your product effectively, prepare for similar changes in future and capitalize on lucrative seasonal events. 

4. Optimizing ad spend strategies

With Marketing Mix Modelling, you can simulate different spend amounts, bid strategies, and campaign budgets to see the potential impact on revenue.

Unlike A/B testing, which tends to assess the impact of a single campaign variable, Meridian can give you the bigger picture. You’ll get a broad, long-term view of your ad spend strategies so you can optimize your marketing mix spend based on the forecasting insights provided.

How will Meridian impact marketing measurement?

Meridian could have a significant effect on marketing measurement, especially for businesses with large marketing budgets. But there are some challenges to consider, too. Let’s take a look at the pros and potential pitfalls of Meridian.

Positive impacts of Meridian

• Marketing measurement is now available to everyone, levelling the playing field for brands of all sizes. Previously, MMMs were mainly accessible to enterprise businesses with significant budgets. Higher ad spend is still a big advantage, but mid-sized brands can now leverage many of Meridian’s insights.
• Meridian is cookieless, so it encourages a shift towards the use of first-party data and privacy-first analytics.
• Customizable modelling allows marketers to adjust the model to reflect their unique business needs, while avoiding vendor lock-in.

Challenges and considerations

• While Meridian is designed to make MMM more accessible, you need a good grasp of statistical analysis to make the most of it. Data science skills are essential for implementing Meridian effectively, which aren’t always available in marketing teams.
• You’ll need a lot of high quality data to get the best from Meridian. Poor and/or inconsistent data will lead to unreliable results.
• You’ll get limited short-term insights, as MMMs focus on long-term trends over real-time performance. That’s (partly) why platform-specific analytics still have an essential role to play alongside MMMs like Meridian.

What do marketers think about Meridian so far?

The discourse on Meridian so far is cautiously optimistic. Most marketers believe that while Meridian does make MMM more accessible, it’s still far from straightforward. In a LinkedIn post, May Insights founder Tad Slaff says:

“Does this make Marketing Mix Modeling easy? The short answer is no. You still need to have someone with a strong statistical background to implement the model and interpret the results.

However, Meridian does make MMM significantly more accessible. Instead of building a model from scratch or relying on a third-party black box that you don’t fully understand, marketers now have a transparent, open-source solution that delivers better insights with less effort.”

Other marketers agree that Meridian may be less accessible than Google suggests:

‍

‍

Other marketers suggest that while MMMs should present insights impartially, there’s bound to be an inherent bias depending on which MMM you use:

But despite some doubts, there are reasons to be excited about Meridian. The fact that Meridian is tied to Google search data makes this a unique development:

Some early adopters suggest that open-source MMM capabilities are as good as vendor-based MMMs. So if you have the in-house expertise to operate Meridian effectively, it could be a really powerful tool:

Overall, Meridian is an exciting step towards making marketing mix modelling accessible for everyone. But while the code is readily available, you need to ensure you have the knowledge and expertise within your team to implement Meridian and interpret insights accurately.

What marketers should do next

Interested in using Meridian to model your future marketing results? Here’s what to do:

1. Assess your readiness — Review your data, staff knowledge, and ad spend to ensure you have the resources to get the most out of Meridian. If you don’t, weigh up the benefits and risks of investing in the resources you need to use Meridian effectively.
2. Take a hybrid approach — Different measurement methods work best in different contexts. Combining Meridian with your existing attribution models helps you understand real-time results and see the impact of external factors, giving you a better view of your overall performance.
3. Explore open-source community contributions — As Meridian is open-source, marketers and data scientists can make improvements to it over time. As the tool evolves, it may become more useful, so keep an eye on these developments.
4. Experiment with Meridian — If you’re unfamiliar with MMMs, start small. Validate any insights Meridian generates and refine your modelling over time. 

Using an MMM platform like Meridian can give you a significant advantage, but you need the right resources to use it. 

Fortunately, there are much easier ways to optimize your advertising spend in the short-to-medium term. Learn how Lunio removes invalid traffic from your PPC campaigns to reduce wasted ad spend.