Analyzing Website Choice Using Clickstream Data

Title: Analyzing Website Choice Using Clickstream Data

Research Question: How can clickstream data be used to understand and predict user choices in advertising-supported Internet markets?

Methodology: The study uses clickstream data from Plurimus Corp., a company that collects data on user activity on various websites. The data includes the websites visited by a panel of users, the order in which they arrived at the sites, and other demographic characteristics. The study applies a methodology similar to Guadagni and Little's (1983) multinomial logit model, which has been used to analyze grocery scanner data. This methodology allows for the estimation of switching behavior and the impact of changing variables on market share.

Results: The study finds that the methodology has reasonable out-of-sample predictive ability, meaning that it can accurately estimate user choices in the Internet portal market. The results also show that informative simulations can be conducted to estimate the impact of increasing advertising by one dollar on the number of visits to a website.

Implications: The study suggests that clickstream data can be a valuable tool for understanding user choices in advertising-supported Internet markets. The ability to accurately estimate user choices and the impact of changing variables on market share has important implications for market researchers, advertisers, and website developers. Future research should focus on alleviating the bias in the estimates by applying more recent developments in the econometric analysis of panel data.

Link to Article: https://arxiv.org/abs/0110008v1 Authors: arXiv ID: 0110008v1