What is marketing mix modeling?
The idea is to explain past sales in terms of what drove them. MMM looks at how sales moved as spend in each channel changed over time, and separates the effect of marketing from everything else.
MMM is also called media mix modeling. It's most useful for connecting past spend to past sales, and for informing how a budget should be split across channels in the next planning cycle.
How does marketing mix modeling work?
An MMM fits a statistical model to historical data, usually two or more years of weekly sales and spend by channel, plus outside factors. The model estimates each channel's effect, including how effects build up over time and where more spend starts to produce less.
- Collect the data. Sales, spend by channel, and outside factors like seasonality, price, and promotions.
- Fit the model. Estimate how much each factor moved sales, accounting for delayed and accumulated effects.
- Read the response curves. See how sales respond to spend in each channel, and where returns diminish.
- Plan the next budget. Use the curves to test how a different split across channels would likely perform.
What is an example of marketing mix modeling?
Rebalancing a channel mix
A retailer spends across search, social, video, and TV. An MMM shows that search is near its saturation point, while video still has room to grow before returns diminish.
The team models a scenario that moves part of the search budget into video and sees a higher projected revenue for the same total spend, before committing a dollar.
Marketing mix modeling vs. attribution vs. brand tracking
Many companies use some mix of these methods. Each answers part of the question.
| Last-click attribution | Marketing mix modeling | Brand tracking studies | |
|---|---|---|---|
| What it measures | The last touch before a sale | Each channel's contribution to sales | Awareness and perception |
| Data it uses | User-level tracking | Aggregate spend and sales over time | Survey responses |
| How current | Daily, but narrow | Periodic | Quarterly or annual |
| Brand impact on revenue | Not measured | Usually limited | Measured, but not tied to revenue |
| Best for | Day-to-day optimization | Budget allocation across channels | Tracking how a brand is seen |
What are the limits of traditional marketing mix modeling?
MMM is valuable, but traditional approaches have known gaps.
- Slow and periodic. Results often arrive as a report after the budget has already been set.
- Sensitive to noisy data. Estimates can swing as new data lands.
- Brand treated as noise. Most models measure the moment of sale and miss the brand conditions that made it more likely.
- Backward-looking. A report explains the past, but teams also need to test scenarios for what comes next.
How does Alli do this?
Alli by PMG's Scenario Planner pairs media response modeling with signals from Alli Brand Equity. One model estimates the direct revenue response to media spend by channel. The other captures the longer-term value of brand equity. Two-stage estimation separates the direct effect of marketing from the accumulated value of brand.
Teams set a forecast period, budget, and channels, see predicted revenue, review recommended spend by channel, and compare scenarios side by side before any money is committed. Forecasts are modeled estimates, not guarantees.
Explore Alli Scenario PlannerMarketing mix modeling FAQ
What is MMM in marketing?
MMM stands for marketing mix modeling, a statistical method that estimates how much each marketing channel and outside factor contributed to sales, using aggregate historical data.
What is the difference between MMM and attribution?
Attribution assigns credit to the touchpoints a tracked user saw before converting. MMM estimates each channel's contribution from aggregate spend and sales over time, so it doesn't rely on tracking individuals and can include offline channels.
How much data does marketing mix modeling need?
MMM typically needs a long history of spend and sales by channel, often two or more years of weekly data, plus outside factors like seasonality, pricing, and promotions.
Does MMM measure brand impact?
Traditional MMM often treats brand as background noise. Newer approaches pair media response modeling with brand equity signals, so teams can see how brand strength supports revenue over time.
Last updated ‹ Back to the glossary