First impression on unpacking the Q702 test unit was the solid feel and clean, minimalist styling.
Ask.com bets on semantic search, targeting special audiences
- — 03 July, 2009 04:31
In the past eight months, Ask.com has unfurled a set of changes to its search engine that the IAC unit calls a success, although its share of U.S. search queries has actually shrunk during that time period.
In October 2008, Ask.com announced it had sharpened the relevance of its search results, made the engine faster, simplified the site's layout and boosted its ability to handle queries in natural language through semantic search technology.
In January of this year, it struck a broad marketing and technology deal with NASCAR (National Association for Stock Car Auto Racing) intended to attract its more than 75 million fans to Ask.com.
However, Ask.com's share of U.S. search queries dropped from 4.5 percent in May 2008 to 3.9 percent in May of this year, while market leader Google grew its share from 61.8 percent to 65 percent, according to comScore.
Things looked brighter for Ask.com several years ago. For example, in November 2005, Google handled almost 40 percent of all U.S. queries, while Ask.com placed fifth with 6.5 percent, according to comScore.
Scott Garrell, president of Ask Networks, views things from a different perspective, pointing out that Ask.com's queries are growing. The search engine handled 486 million U.S. queries in May 2008 and 555 million in May of this year, according to comScore. "In a very tough and competitive market, we're holding our own," he said in an interview.
Garrell also points out that Ask.com and the other sites that make up the Ask Network, like Dictionary.com, are collectively the sixth-largest Web property in the U.S., ahead of powerhouses like eBay, Facebook, Wikipedia and Amazon, according to comScore.
Garrell is particularly encouraged by Ask.com's advances in semantic search and in its attempts to attract specific audiences like NASCAR fans to the search engine.
Semantic search is a problematic term because different search providers define it in various terms. In general, it refers to search technology that lets an engine understand the meaning of text, whether in a query or in a search result, so that users can phrase queries in natural language instead of keywords.
"People don't talk in keywords," Garrell said.
Back in the mid-to-late 1990s, when Ask.com was known as Ask Jeeves and the world hadn't heard of Google, Ask.com was a leading search engine, known for encouraging people to type in queries in the form of regular questions.
According to Garrell, that perception persists, although after the dot-com bubble burst, Ask Jeeves abandoned the consumer search market for several years to focus on enterprise search, before reversing course in 2003.
"We get more queries in the form of a question than the industry average, and we get queries that are longer," he said.
Consequently, Ask.com considers it essential to be the best search engine at semantic search, which is why it has invested much time and effort in its question-and-answer search engine. Introduced in October, this Q&A engine uses semantic search technology to interpret questions and return relevant answers found on the Web. Last month, Ask.com announced that the engine has a database of 300 million question-and-answer pairs.
In addition, Ask.com has reaped very good results from its collaboration with NASCAR, which included the selection of Ask.com as NASCAR.com's official search engine, the launch of a NASCAR-branded browser toolbar and the sponsorship by Ask.com of a racing car, he said.
Garrell thinks Ask.com can pursue this "audience-centric" strategy with eight to 10 vertical markets per year, having seen that it's an effective and interesting approach to promoting and growing usage of the search engine.
It remains to be seen whether these and other initiatives will allow Ask.com to grow its share of queries in a market so vastly dominated by Google.