Optimize Customer Behavior Through Clickstream Analysis

Project Description

For a given client website we want to recommend the most effective content sequence (web pages) to lead to a goal conversion e.g. sign up for a demo. Our clients are mostly B2B businesses and the B2B sales cycle compared to B2C has more key touch points and a time duration in months. For now, a goal is a URL and we are dealing with only website content, not other channels like emails. Because a goal is typically a generic thank you page, the visitors also need to be segmented. For a client like iron mountain, the segments would be like SMB, Healthcare, etc.. The input data is the http clickstream data (page view events). These events sometimes do have segment identifiers but we are finding they are not very useful in the current setup (we may need to run some clustering algorithm to identify “useful' segments). Other input data available includes Google Analytics and Site Catalyst. The output data we want to generate is for a given segment and goal, what is the recommended content sequence.