3. Petri ‘Pete’ Mertanen
• Speaker, analytics coach & consultant
• BBA, Specialist Qualification in Management
• (Digital) Analytics experience since 2005
• Lecturer at Aalto University and Laurea University
of Applied Sciences
• Presentations at SlideShare
• Certifications for Analytics and Data Science:
• Elements of AI
• Cookie Consent Expert
• Statistical thinking for Data Science & Analytics
• Google Analytics Individual Qualification,
Google Tag Manager Fundamentals,
Introduction to Data Studio
5. Different kind of analysis methods
● Descriptive analysis
○ Answering mostly to question “what happened”?
○ Mostly with aggregated data, doesn’t include necessary any recommendations
○ May concentrate only on last click conversions
○ Heuristic analysis as a method - which you can always challenge!
● Mathematical / statistical modelling
○ (Data driven) attribution models and conversion optimization (testing)
○ May include predictive analysis (in GA4: visitors who are likely to buy, anomaly detection)
○ Different kind of regression analysis, usually used as basis in Marketing Mix Modelling (MMM)
○ Causal inference analysis
○ AI based analysis
6.
7. Attribution models
• Last click - so last season
• Fist click
• Linear
• Time decay
• Position based
• Data driven
9. Attribution is incomplete
because of
• Individual tracking won’t work
• No consent = no tracking
• All browsers aren’t accepting cookies
• Incognito, ad blockers, ad fraud…
• User biased
• Platform biased
• Walled gardens
11. The linear regression model explained
• Dependent variable y = we are trying to explain, e.g. ecommerce sales
• Explanatory variables, for example:
• Advertising spend (in different channels)
• Discounts or seasonality
• External factors, like weather, COVID-19 etc.
• β1 is the intercept term (constant) = it’s important to know baseline sales
• βk is the slope coefficient of variable xk (marginal effect on sales)
14. Online Marketing Mix Modelling
• Privacy friendly - no cookies!
• Data from all ad networks in one platform
• Outcomes from the back-end
• Fast and cheap to implement
• Data science in place
• Know what is your base sales!
• Find the real incremental value of advertising!
• See saturation points per channel
• Optimize / make scenarios for your ad mix
• Predict the outcome (sales / conversions)
15. How to do (online) MMM?
• With your own data scientist (or team)
• External consultant
• With BigML.com or other platform
• External platform vendor
• Our partner in crime is Sellforte
16.
17. No one can be told what the MMM is.
You have to see it for yourself.