Course contentsShow
← Data Science
Lesson 1737 of 2,14541. Marketing and Growth AnalyticsFree lesson

Aggregate-Level Data in MMM

How MMM uses aggregate spend and performance data rather than individual user-level tracking.

Aggregate-Level Data in MMM

You'll learn how Marketing Mix Modeling operates on summed spend and performance metrics instead of tracking individual user actions.

Why Aggregate Data?

Unlike attribution models that trace each user's journey through touchpoints, MMM works with aggregate-level data—totals and averages rolled up by time period (usually weekly or daily). Instead of knowing "User 12345 saw ad X, clicked, then converted," MMM sees "We spent $50,000 on TV this week and had 2,000 conversions."

Think of it like weather forecasting: meteorologists don't track every water molecule. They measure temperature, pressure, and humidity at aggregate levels to predict rain. Similarly, MMM analyzes total marketing spend across channels and correlates it with total business outcomes.

What Gets Aggregated?

Inputs:

  • Total spend per channel per time period (e.g., "$100K on Facebook last week")
  • Impressions, reach, or GRPs (Gross Rating Points) delivered
  • External factors like seasonality, holidays, competitor activity

Outputs:

  • Total sales, revenue, or conversions for that period
  • Website traffic, brand search volume
  • Other KPIs aggregated to the same time grain

Why This Matters

Aggregate data makes MMM privacy-friendly (no cookies or user IDs needed), historically robust (works with years of past data), and capable of measuring channels like TV and radio that don't offer click-tracking. However, you lose granularity—you can't see which customers responded to which ad, only that increased spend correlated with increased outcomes.

Key Takeaway: MMM uses aggregate-level spend and performance data summed by time period, trading individual-level precision for broad, privacy-safe insights into overall channel effectiveness.