Marketing Mix Modelling
Statistical approach to attribute sales to marketing and external factors.
Marketing Mix Modelling (MMM) is a statistical technique that quantifies the incremental revenue contribution of each marketing channel and non-marketing factor (seasonality, economic conditions, pricing, brand equity) to overall business outcomes. It is the most robust attribution method available and is the approach large UK advisers and networks increasingly use to settle channel-level investment decisions.
Unlike last-click attribution or platform-reported conversions, MMM does not rely on user-level tracking. It uses aggregated weekly or monthly data across two to three years, builds a regression or Bayesian model, and isolates the marginal effect of spend in each channel after controlling for base demand, competitor activity, seasonality and macro factors.
For UK regulated advisers, MMM becomes valuable at roughly £40k per month in cross-channel spend, because below that threshold there is insufficient variance in the data to produce stable coefficients. Firms below that level should use incrementality testing (geo splits, holdout tests) rather than full MMM.
Typical outputs include channel-level return on ad spend, diminishing-returns curves (the point at which each additional pound in a channel delivers less than the previous), optimal budget allocation across channels, and a base-versus-incremental split of total client acquisition.
MMM has become more accessible for mid-market advisers through open-source tools (Meta's Robyn, Google's LightweightMMM) and specialist vendors. Platinum Prospects' benchmark work uses MMM-style incremental attribution layered on top of platform-reported data.
See also: multi-touch attribution, incrementality testing, geo experiment, diminishing returns.