DENVER-The search engine market is on the precipice of some great innovation, but no one knows exactly what next-generation search engines will look like, four search experts concluded during a panel on the future of search at the Defrag conference here Nov. 5.
The conversation, which included Yahoo’s Tom Chi, Siderian Software’s Bradley Allen, Thomson Reuters’ Tom Tague (for the OpenCalais Web service) and Isys Search Software’s Derek Murphy started with the basic agreement that current search engine results from Google, Yahoo and Microsoft are flooded with too much noise.
This is noise in search: Allen said while running he heard two women talking about doing Google searches and not knowing what to do after receiving millions of results. “It’s wonderful technology and when it works it’s great, but there is no guidance” as to what to do if users can’t find what they need within the first 10 blue links, he said.
Click here to see scenes from the Defrag keynote.
Chi, whose company is the second-leading search engine behind Google and has drastically overhauled its process to change that, was less kind about the current state of search. “The only problem with search is that it’s awful, but other than that it works pretty well,” he said.
That blatant oxymoron aside, Chi said search is generally stupid. When he checks search logs, he said, he notices that people are still doing navigational searches for Ebay.com, which is “ridiculous.”
But rather than take steps to make search more intelligent on behalf of users, Chi said, major players-Google, Yahoo and Microsoft-are not highly motivated to do this because if a user spells eBay wrong and does a spelling correction to get other results, search engines still profit from those mistakes thanks to online advertising clicks.
Yahoo is positioning itself to change this by attempting an ambitious broad stroke. Yahoo earlier in 2008 opened up its search platform, which it calls SearchMonkey, to let objective outside programmers build applications that augment search for users.
Future Search Will Include Semantics, Social Graph
It is still too early to know how this is faring, but Chi noted that SearchMonkey is Yahoo’s stab at investing in an ecosystem to attract folks interested in indexing structured data. Opening the platform has paved the way for MySpace, LinkedIn, Yelp and others to influence search.
Ultimately, Yahoo wants to create world in which people with “distributed intelligence” cooperate to improve search because even the thousands of programmers at Google, Yahoo and Microsoft “won’t come up with everything” regarding search, he said.
One imagines Allen, Tague and Murphy breathed an internal collective sigh of relief. These men and their companies are all working on providing vertical search beyond the 10 blue links the general search vendors currently offer.
Tague said there will be a place for vertical search engines in the future and these won’t necessarily cannibalize top-line search, and he is optimistic that these vertical searches will reap some ad revenue. He also said companies will have to resolve how the social graph interoperates with search in an age in which walled gardens keep data tethered.
Allen said Siderean wants to help assemble search and discovery tools on demand in a business context, which is something Google and co. won’t do at this point. He said he anticipates that the paradigm of search being about getting a Web page is evolving to help users find people, places and things, so that results pages will reflect an aggregated analysis of what you want to get back.
In short, through semantic technologies that divine the meaning behind queries and the application of the social graph to search technologies, search will get smarter.
One audience member asked the panel whether and when we will be able to “skip the search results page” and receive geographically dispersed data in some other way, thanks to “folksonomies” and other relevant groupings of behavior.
Such a solution seems like some hybrid of search, social networking and wikis. Wikia and others seem to be moving in that direction, but the search results page remains.
In conclusion, no one knows what the next-generation search engine will look like, although the panelists agreed that semantics and social collaboration will play a role. Do you have any ideas?