Search Engine “Cuil” Focuses on Content Analysis

An index of 120 billion web pages dwarfs the competition.

Archive notice: This article was originally published on July 29, 2008. Links and embedded videos are preserved as part of the historical record.

California start-up Cuil unveiled its search engine of the same name yesterday, Monday, challenging established search services. According to the company, Cuil covers more than 120 billion web pages, giving it by far the largest index in the search market. “Cuil also presents searchers with content-based results, not merely popular ones,” emphasizes CEO Tom Costello. With this approach, the company founded by former IBM and Google employees aims to distinguish itself from competitors and make relevant results easier to access.

Rather than link analysis and rankings, Cuil relies on closer analysis of web-page content. A page’s popularity is not the primary factor determining result order, the company says. Instead, page context is related to queries to organize search results. Cuil uses a magazine-style layout rather than a link list, with longer extracts from web pages, often accompanied by images. Especially for broad queries, results are grouped into categories and suggestions are offered for narrowing the search parameters. Cuil also sees the number of indexed pages as an advantage: 120 billion pages is said to be more than three times as many as any other search engine.

“By using our expertise in search architecture and relevance methods, we have designed a more efficient yet richer search engine from the ground up,” says Cuil President and COO Anna Patterson. Like colleague Russell Power, she has experience developing search at Google, while Costello worked in the field at IBM. At launch, their product still has minor flaws; for example, Cuil ranks “The New York Times” ahead of the city’s official website for the query “New York.” However, US search expert Danny Sullivan notes on Search Engine Land that Cuil is performing better at launch than many other search-engine start-ups.

Competition for Cuil could come from Powerset, which introduced semantic Wikipedia search as its first product in May. The concept is intended to be implemented more broadly later. Mark Johnson, Product Manager at Powerset, believes the acquisition by Microsoft announced in early July may make that easier, giving Powerset access to the resources of Microsoft’s Live Search team.