Client: A fashion retailer

Project: Create visualizations of web paths

Challenge: Keep development costs reasonable while solving a combinatorial problem.

Achievement: The online marketing department wanted to see common paths (and statistics) taken by shoppers on a website. Most web traffic reports counted the number of customers and the number of times a page was viewed, but did not show the interaction between customers and pages.

The solution consisted of 2 separate strategies. A sequencing algorithm was configured to calculate all explicit paths. However, the cardinality of the number of pages on the site was so high that it quickly became overwhelmed. The alternative was to ask the business stakeholders how they would try to interpret the information and learn that a starting point of the path had to be an anchor for the analysis. Given the starting point, all sessions were processed and the explicit paths deduced. Shoppers with identical paths were aggregated and the results were transformed into the required schema. That schema was converted into an XML feed that a JAVA applet using tools from SAS Graph.

The effort was reduced to several days of work and provided the visualizations on time and under budget.