MMM vs Attribution
Top-down vs bottom-up.
Two Lenses on the Same Question
Marketing Mix Modeling (MMM) and multi-touch attribution (MTA) both try to answer "what drove sales?" but from opposite directions.
MMM is top-down: it uses aggregate, time-series spend and sales data to estimate each channel's contribution. MTA is bottom-up: it stitches together individual user touchpoints toward a single conversion.
Advanced practitioners use both, knowing each has blind spots the other partially covers.
What MMM Actually Measures
MMM regresses sales against marketing and non-marketing drivers over time. It produces channel-level elasticities and a decomposition of sales into base and incremental contributions.
Crucially, MMM captures offline channels (TV, radio, OOH) and long-term effects that user-level tracking cannot see.
sales_t = base
+ b1*TV_t
+ b2*Search_t
+ b3*Social_t
+ b4*OOH_t
+ seasonality_t
+ error_tAll lessons in this course
- MMM vs Attribution
- Inputs and Variables
- Reading an MMM
- Budget Reallocation