Lecture on Multi-Touch Attribution (MTA) vs Marketing Mix Modeling (MMM)

Presenters

• Matthias from Windsor.ai
• Enes from Windsor.ai

Introduction

• Customers frequently ask about the difference between MTA and MMM.
• Goal: Provide a simple explanation of the two methods.

Multi-Touch Attribution (MTA)

What is MTA?

• Measures different touchpoints customers have before a purchase/form fill on a website.
• Example touchpoints: display ad, retargeting, brand search, organic search.

Attribution Models

• First Touch Attribution: Credit goes to the first touchpoint.
• Last Touch Attribution: Credit goes to the last touchpoint.
  • Historically used in GA UA (Google Analytics: Universal Analytics).
  • Disadvantage: Ignores upper funnel channels like display or paid social.

Data-Driven Attribution with GA4

• Switch from last touch attribution to data-driven attribution.
• Uses a Markov model to better allocate credit to each touchpoint.
• Simulates removing a channel from historical data to see impact on conversions.

Challenges with MTA

• Increasing difficulty due to wall gardens (e.g., Facebook and Google ads not allowing impression tracking).
• Reliance on click data only.
• Cookie consent and GDPR issues reducing trackable data.
• Example: In Germany, up to 40% of transactions might not be trackable.

Marketing Mix Modeling (MMM)

Introduction to MMM

• MMM: Correlates spend to revenue instead of touchpoints to revenue.
• Measures overall performance of marketing models.

How MMM Works

• Can test different parameters to see impact on advertising results.
• Example: Increasing spend in a specific city or platform and measuring impact.

Benefits of MMM

• Considers seasonality and external factors (e.g., weather).
  • Example: Ice cream sales peaking in summer.
  • Example: Umbrella sales increasing on rainy days.

Comparison: When to Use MTA vs MMM

• MTA: Best for granular data (impressions, likes, clicks).
  • Limitation: Can't track offline ads (TV, radio).
• MMM: Best for setting company goals and making future predictions.
  • Limitation: Doesn't provide detailed behavior of individual users.

Unified Marketing Measurement

• Combines both MTA and MMM methods.
• Gains insights from both strategies but adds complexity to the analysis.

Conclusion

• Understanding when to use MTA and MMM can enhance marketing strategies.
• Questions and feedback are welcome.

Call to Action

• Subscribe and comment for further questions.