Personalizing the checkout Samare Jarf Klarna about me
Personalizing the checkout Samare Jarf, Klarna
about me da. Ta scien. Tis. T a. T Klarna Model and understand customer behaviour Statistical models for personalized customer experience MSc. in Engineering Mathematics from LTH in Lund Studied financial statistics
how Klarna has optimized the checkout through personalization
klarna checkout
learning your product is crucial measure performance discover opportunities
average order amount high amount low amount share part payment has used part pay never use part pay
how can we know who wants to pay with what?
is this good enough?
no
Machine Learning !? “The field of study that gives computers the ability to learn without being explicitly programmed”
1. determine the problem
classification discrete/categorical variable regression real number/continuous
classification discrete/categorical variable regression real number/continuous
2. build features
“…some machine learning projects succeed and some fail. What makes the difference? Easily the most important factor is the features used. ” - Pedro Domingos A few useful things to know about machine learning. Pedro Domingos.
purchase customer historical
purchase amount estore weather. . . customer historical
purchase customer amount estore weather. . . age shoe size. . . historical
purchase customer historical amount estore weather. . . age shoe size. . . pay. Method_last_purchase. . .
3. model selection
Random Forest
Dataset Sample Random Forest
Dataset Sample Random Forest Sample . . .
Purchase Random Forest . . . Invoice Card Invoice
Purchase Random Forest . . . Invoice Card Invoice Combined prediction
optimizing conversion for each transaction
how do we achieve that?
Purchase Scoring service Card
Purchase Scoring service Pay in end of March
with data driven machine learning Klarna dynamically adapts the checkout to your preferences to optimize customer satisfaction and conversion
thank you
- Slides: 34