What is marketing/media mix modeling (MMM)?

Marketing Mix Modeling (MMM) predicts business outcomes through a statistical analysis using multivariate regressions, with marketing tactics and spend as variables. The regressions provide contributions of each variable to outcomes, which are then used to predict what conversions and sales would be with different inputs or marketing mix.

How does marketing mix modeling work?

Marketing Mix Modeling, also called Media Mix Modeling, collects aggregated data from marketing and non-marketing sources over a multi-year historical period, also factoring in external influences such as seasonality, economic data, weather, and promotions. The data is then used to develop a demand model which quantifies the historical contribution of each marketing and non-marketing input to a business outcome, like sales or conversions.

Marketing mix modeling example

A clothing brand marketer wants to know how each media channel contributes to sales. If the brand has collected sales data and advertising spend for each channel during a several-year time frame, MMM can be used to run a multivariate test on many different points in time. The analysis will show what expected sales will be when there is a change made to media spend. While the model can be effective, especially if there is a large amount of data available, it’s based on historical data, meaning it only reveals correlation, not necessarily causation.

What are the advantages of marketing mix modeling?

If you are an established brand, data is likely readily available and MMM can glean a lot from two to three years of historical data. MMM is also able to model non-media variables such as macro-economy influences (like COVID-19), competitive influences, seasonality, promotions, and other trends. The biggest advantage of MMM is high-level analysis across the entire media portfolio - great for delivering strategic long-term planning insights into your non-addressable and addressable media - but not ideal for tactical or ongoing insights.

what are the limitations of marketing mix modeling?

MMM estimates marketing impact on historical business outcomes based on probability and can be subject to the correlation vs. causation dilemma. For forward-looking projections, MMM relies on a number of assumptions for non-marketing factors as well the assumption that channel level media mix, cost, and response does not diverge from the historic data that is the basis for the demand model.

While well-built models based on high-quality data can overcome the correlation vs. causation dilemma to provide channel lift and forecasts, the limitation on degrees of freedom and challenges with overspecified models means that they cannot be used to inform tactical decision making at the sub-channel level. Because models rely on multiple years of historical data to determine an average read for marketing inputs, they are challenged in teasing out dynamic changes to marketing channels and/or business changes in recent periods. In other words, it's not exactly agile and won't deliver the level of insight needed for day-to-day optimization.

An alternative approach to understanding each marketing mix contribution and informing media investment decisions is to run ongoing incrementality testing.

With Measured, you can easily run incrementality measurement and testing on 70+ media publisher platforms. Utilizing our API integrations with media platforms, you get a cross-channel view of your marketing mix in less than 24 hours.

Is MMM a fit for you? If you are looking for support on long-term planning decisions, use primarily non-addressable media, and have at least two years of historical data to work with, it's worth looking into! If you need access to the latest performance data for ongoing media optimization, the Measured Intelligence Suite delivers incrementality insights for informed and agile planning without years of data.

Measured vs platform reporting, Multi-Touch Attribution (MTA) & Media Mix Modeling (MMM)