Is this 90s tool making a comeback?
As third-party cookies disappear, marketers are rediscovering a 90’s tool with some modern updates.
Marketing Mix Modelling (MMM) has been deemed as a ‘big brand tool’ that was complex, and slow, however MMM is having somewhat of a renaissance.
This is due to third-party cookies disappearing, multi-touch attribution becoming more limited, and marketers requiring more and more trusted , privacy-safe measurement.
Unlike cookie-based tracking, MMM doesn’t rely on user-level data. Instead, it analyses aggregated outcomes to isolate the true impact of media, pricing, and seasonality – one of the key reasons behind its recent resurgence. At the same time, advances in modelling techniques, increased computing power, a shift toward causal inference over simple statistical fit, and improved visualisations are helping to bring MMM into the marketing toolkit.
What is Marketing Mix Modelling?
Marketing Mix Modelling (MMM) is a statistical approach used to measure the impact of different marketing activities—such as TV, paid social, search, and promotions on business outcomes like sales, leads, or sign-ups. By controlling for seasonality, trends, and external factors, MMM identifies which channels are genuinely driving results, not just appearing correlated.
MMM first emerged in the 1940s, when large consumer goods companies used econometric models to understand the impact of TV, radio, and print advertising on sales. These models were often run infrequently, sometimes only once a year because they required manual data collection and complex statistical analysis. Over the decades, MMM has evolved alongside advances in computing, data availability, and marketing channels.
How it traditionally worked
Historically, MMM relied on small teams of analysts working with limited datasets, often from just a handful of channels. Data was cleaned and aggregated manually, then run through regression models to estimate the contribution of each channel. Reports were delivered months after campaigns ended, meaning insights were often backward-looking and slow to influence decision-making.
How MMM Has Evolved for Today’s Marketers
Today, MMM is faster, more flexible, and more scalable thanks to modern tools like Robyn (Meta’s open-source MMM package) and Meridian (googles MMM package). These platforms automate much of the modelling process, integrate with multiple data sources, and run frequent updates—sometimes weekly or monthly. This shift means businesses can react to changes in market conditions quickly, test different budget scenarios, and move from insight to action in days, not months.
- Bayesian methods have become more common place and bring advantages – enabling domain knowledge to be integrated into the modelling process.
- Teams are realising MMM doesn’t need to be a quarterly black box – it can be always on.
- It’s not about choosing between MMM and digital attribution. It’s about building a resilient, multi-layered measurement strategy.
Use cases
MMM can be applied across a wide range of marketing and business challenges, including:
- Budget optimisation: Reallocating spend for maximum ROI.
- Channel performance analysis: Understanding which channels drive short- vs. long-term value.
- Campaign evaluation: Measuring the impact of seasonal or promotional campaigns.
- Market expansion: Estimating how new regions or audiences might respond to marketing.
- Scenario planning: Testing “what if” budget shifts before making real-world changes.
A comparison with other attribution methods
MMM differs from multi-touch attribution (MTA), which tracks individual user interactions online. While MTA is useful for digital channels with granular tracking, MMM offers a privacy-safe, holistic view across both online and offline channels. Unlike MTA, MMM doesn’t rely on personal data, making it compliant with privacy regulations and resilient to tracking restrictions. In practice, many brands use both methods together—MTA for digital campaign optimisation, and MMM for strategic, cross-channel planning.
Advice on getting started
For businesses looking to start with MMM:
• Collect high-quality data: Include marketing spend, impressions, and key outcome metrics, ideally across at least two to three years.
• Include external factors: Weather, economic indicators, competitor activity—anything that may influence sales.
• Start with a pilot: Model a subset of markets or channels to demonstrate value before scaling.
• Align stakeholders early: Bring in marketing, finance, and data teams from the outset to ensure buy-in and trust.
• Plan for iteration: Treat MMM as an ongoing measurement tool, not a one-off project.
Find out more
Marketing Mix Modelling is a powerful tool for optimising your marketing strategy and improving ROI. At STRAT7, we have a wealth of experience and a strong view on why using an external agency to handle this analysis is beneficial.
So if you’re interested in finding out more please don’t hesitate to get in touch, or contact Steven Pesarra, Director, Data Science, directly at steven.pesarra@bonamyfinch.com.