Unfamiliarity stems demand for 'soft' search

The online search scene continues to evolve as more contextual search sites sprout up in the market. However, analysts say demand for such search engines remain niche.

Unlike search engines that run on keywords, search sites such as Hunch, Answers.com, Mahalo and ChaCha, aim to deliver results tailored to specific human situations. Hunch, for example, asks users a series of five to 10 questions before churning search results.

Madan Sheina, principal analyst at Ovum, described contextual search engines as the combination of two components: hard statistical machine-driven search algorithms, and soft human contextual metadata.

People almost always run searches to help them make decisions, Sheina said in an e-mail interview. The purpose of search, he added, is to assist the users in "some kind of information [or] research--be it finding a local plumber, analyzing a company's latest 10-K form, or discovering how the investigation regarding Michael Jackson's doctor is progressing."

Contextual search is akin to bringing search to users, as opposed to bringing users to search, Sheina explained, adding that such engines provide search capabilities that understand what users are doing. "I think this is something all users would like to have, but up to now, don't know who to ask or where to look," he said.

He noted that keyword-based search results are often too broad and too irrelevant to be of immediate use. "Often, it's quite an arduous task to filter or sift through crudely ranked keyword results...to find the information for the need at hand," he added.

Sheina explained that keyword searches are generated on limited search query metadata, which are directed by user cues and prompts.

Studies have shown that using context to guide search effectively offers a more advanced Web search platform even to inexperienced users, he said.

Allowing natural context to guide online searches provides a better match to users' current needs, instead of simply relying on users to come up with what they deem to be the most appropriate keyword, he noted. For example, as more people use Hunch over time, the search site will be able to churn more refined results that are specific to each user, Sheina said.

Restrained by lack of understanding
Despites its potential, however, demand for contextual search remains niche.

Phil Harpur, senior research manager at Frost & Sullivan, said a reason for this is that people do not understand the behavioral component of contextual search engines well enough.

People are concerned that confidential information they provide to these sites can be collected by external parties and revealed to third parties for use in other purposes, Harpur told ZDNet Asia in an e-mail.

"The reality, however, is that users are only tracked by these sites from the time they enter the site or do their search till the time they complete their search and leave the site," he explained.

For now, contextual sites will have minimal competitive impact on keyword search sites, according to the analysts.

Harpur said: "Conventional search, dominated by Google, is of course the first 'port of call' for the vast majority of the Internet user community. This balance is unlikely to change much over the next few years."

Sheina added: "It's not a matter of what technology will return the best and most relevant result, but the site that has the most sticky marketshare. And, right now, that's Google/Yahoo keyword based searches."

However, he noted that Yahoo and Microsoft are starting to take contextualized search more seriously.

In 2005, Yahoo launched the Y!Q search service designed to help people find more-relevant content online. The site does this by offering contextual search that analyzes the page being read, so it can provide a list of related search results. Sheina noted that Microsoft's acquisition of Powerset's natural search technology also suggests the company is eyeing contextual search.

According to Harpur, people are just starting to understand the power of contextual search engines.

Sheina concurred, adding that the rise of the semantic Web will also spur smarter approaches for answering search queries. The semantic Web project seeks to enable the World Wide Web to intelligently interpret what people are looking for when they are searching over the Internet.

"Expect the semantic Web to greatly enhance search functionality on the Web," he said.

Advertising is also driving the development of more intelligent search models. Sheina said: "Competition in search cannot be decoupled from advertising--the real money-spinning battleground for Internet search vendors--where contextual search provides an opportunity for companies to run more targeted advertising on their sites."