Glossary · MeasurementTOFU
Marketing mix modelling (MMM)
The short answer
Marketing mix modelling (MMM) uses statistical analysis of historical data to estimate how much each marketing channel — and other factors like price, seasonality and distribution — contributed to sales.
Unlike click-based attribution, MMM doesn't need user-level tracking. It works on aggregated weekly or daily data, which makes it resilient to privacy changes and able to include offline channels like TV, outdoor and retail.
When MMM makes sense
- Multi-channel budgets at meaningful scale
- Offline and online media that need to be compared on one basis
- Two or more years of reasonably consistent data
- Budget allocation decisions worth a significant investment in analysis
Limitations
- Needs substantial historical data and variation in spend
- Results are estimates with uncertainty ranges, not precise answers
- Slow to reflect sudden changes
- Can be misled by channels that always move together
MMM inputs and outputs
| Inputs | Outputs |
|---|---|
| Weekly spend by channel | Estimated contribution by channel |
| Sales or conversions | Diminishing-returns curves |
| Seasonality, pricing, promotions | Recommended budget allocation |
| External factors (holidays, macro) | Scenario forecasts |
Open-source tools such as Google's Meridian and Meta's Robyn have made MMM more accessible. See attribution modelling.
MMM in practice
A brand uses two years of weekly sales, spend by channel, prices and seasonality to estimate each channel's contribution, then reallocates budget toward channels with higher modelled returns.
Common MMM mistakes
- Too little historical data
- Ignoring seasonality and promotions
- Treating results as precise rather than directional
Frequently asked questions
Is MMM better than attribution?
They answer different questions. Many teams use MMM for strategic allocation, attribution for day-to-day steering, and experiments to calibrate both.
Are there open-source MMM tools?
Yes — several major open-source MMM frameworks exist, lowering the barrier to entry, though expertise is still needed to use them well.
How often should MMM be updated?
Quarterly or after major changes in spend mix, with calibration from incrementality tests.
Is MMM useful for small businesses?
Usually not, because it needs enough data variation across channels and time.
Related terms
Full glossaryTalk to a strategist
Thirty minutes. Your numbers. A straight answer.
Book a strategy call with the people who'd actually run your account. We'll tell you what we'd do — or that you don't need us yet.