Understanding the Specificity of Web Search Queries

Understanding the specificity of Web search queries can help search systems better address the underlying needs of searchers and provide them relevant content. The goal of this work is to automatically determine the specificity of web search queries. Although many factors may impact the specificity of Web search queries, we investigate two factors of specificity in this research, (1) part of speech and (2) query length. We use content analysis and prior research to develop a list of nine attributes to identify query specificity. The attributes are whether a query contains a URL, a location or place name along with additional terms, compares multiple things, contains multiple distinct ideas or topics, a question that has a clear answer, request for directions, instructions or tips, a specific date and additional terms or a name and additional terms. We then apply these attributes to classify 5,115 unique queries as narrow or general. We then analyze the differences between narrow and general queries based on part of speech and query length. Our results indicate that query length and parts-of-speech usage, by themselves, can distinguish narrow and general queries. We discuss the implications of this work for search engines, marketers and users.