A foray into Understanding the Next Billion Search Users
Naman Gupta, Carolyn Penstein Rosé · 2009
This paper presents an investigation into the search behavior of low literacy users in the developing world, like the participants from rural areas in India for our study. The field of personalization research has been hindered by a paucity of appropriate data for inducing effective user models that target the real problems in information access for needy populations, such as low literacy users. Our goal is to address these limitations using a data-driven, user centered methodology. We present results from an experimental study that demonstrates that some important assumptions underlying current probabilistic models of information seeking on the web that govern the behavior of popular search services such as Google and Yahoo are not valid for our target user population. We present an analysis that offers specific suggestions for better supporting the information seeking practices of such users.