Using Attention Models to Predict Web Page Conversion Rates -- Thomas Levi

Speaker: Thomas Levi, Director of Data Science @ Unbounce

Title: Using Attention Models to Predict Web Page Conversion Rates

Abstract: Unbounce is constantly striving to give our customers the best information we can so they can make data driven decisions and improve their landing pages. The research team has embarked on a program to understand how each piece of a landing page ((e.g. copy, design) contributes to its conversion rate. In this talk, I will discuss how we used self-attention models (specifically a modified BERT model and some related variants), as well as metadata to understand how page copy contributes to conversion rate.

Doors will open at 6 and the talk will start at 6:30.