Marketing mix modelling - when does it actually make sense

Marketing mix modelling - MMM - is a statistical methodology that has been used by large advertisers for decades. It models the relationship between marketing inputs and business outputs while controlling for external factors like seasonality and economic conditions. It is distinct from digital attribution, which relies on tracking individual user journeys. MMM uses aggregate data and econometric modelling to measure incremental contribution at the channel level.

Why MMM is being discussed more now

The decline of third-party cookies and the increase in privacy restrictions have reduced the reliability of user-level attribution. When you cannot track individual journeys across devices and channels, the relative contribution of different touchpoints becomes harder to measure accurately. MMM, which uses aggregate data and does not depend on individual user tracking, becomes relatively more attractive in this environment.

When MMM genuinely makes sense

MMM requires a meaningful volume of data to produce reliable models - typically at least two to three years of weekly sales and marketing spend data across multiple channels. It makes sense only when the business is running multiple marketing channels simultaneously at meaningful scale. For a business spending primarily on Google Ads with limited other channel activity, MMM does not provide much additional insight over standard attribution. The businesses for which MMM genuinely adds value are those spending significant budgets across TV, digital, print, and out-of-home where digital attribution cannot capture the full picture.

Lightweight alternatives

For businesses not yet at MMM scale, incrementality testing offers a more accessible way to measure true channel contribution. Pause a channel or reduce spend in a specific geography and measure the impact on overall conversions. The difference between test and control gives you a read on how much of attributed volume was truly incremental. This requires careful design but is considerably less complex and expensive than full MMM, and often produces actionable insight for businesses spending 50,000 to 500,000 pounds annually across channels.