The subject was maps, user experiences and whether size really matters — in fact, anything except advertising– at a roundtable discussion among representatives of the five leading search engines yesterday at the Search Engine Strategies 2005 conference in San Jose.
The size question arose because Yahoo! Search announced on Tuesday that its engine now indexes more than 20 billion Web items, including documents, images and multimedia files. That seems to put Yahoo! in the lead for broadest Web reach, ahead of Google and its 11 billion item index.
Participants agreed that the number of pages indexed is merely a proxy for other more intangible search attributes such as the relevance and freshness of results produced. “Ultimately, it’s not a numbers game,” said Bradley Horowitz, director of technical development at Yahoo!. “I think it took a lot of people by surprise that Yahoo! was up to that level of world-class infrastructure.”
“We’re trying not to grow our index too much,” said Jim Lanzone, senior vice president for search properties at Ask Jeeves. ” We don’t want to to have 150 copies of Wikipedia in our index and every misspelling of Lipitor. A lot of the junk that’s in there, we spend a ton of time trying to remove.”
Asked whether the recent wave of new search features — maps, audio search, and personalized and community-based search — has gone beyond useful innovation into empty competitiveness, Google technology director Craig Silverstein said his company and its rivals were doing necessary technical background work to push search forward. “We’re playing around with a lot of different ways of thinking about search, trying to find something that works better than what we have today,” he said. “The real challenge is to make sure that adding new features doesn’t add complexity. We think a lot about how to make these features invisible, so you can get the benefit without having to learn anything new.”
Ken Moss, general manager of MSN Search, pointed to the new mapping products launched by his company and Google as signs that new, more usable ways of organizing and viewing data were in the offing. “It’s the beginning of what I think will be some incredibly cool views,” he said. “We can combine community data with the map to innovate further in the display of search results.” Silverstein added that Google has taken the complaints that its Google Earth map feature is a “huge time-waster” as evidence that people find looking for local information plotted on maps as a very convenient, enjoyable interface. “Looking at a list is not a natural way to find information, and the only reason it persists on the web is that nothing better has come along yet,” he said. “tying geographical information to maps is a start at finding a way of sarching that gets beyond lists.”
Another hot topic is social search, in which users get to define what Web pages are important to them and to share those pages — sometimes with annotations — with their social networks. Several speakers pointed to Yahoo!’s online photo-sharing application Flickr as an example of the power of online communities to enhance Web function: Members store their personal photos together with descriptive tags and share them with friends or make the available to all. As non-textual content grows in importance in search, such tag data can be particularly useful to the search engines, since computers find it much harder to “read” a photo or multimedia file than they do a page of text.
But Gerry Campbell, vice president and general manager of AOL Search and Navigation, pointed out that sorting the useful aspects of social search from the gratuitous will be an important task in coming months and years. “Asking questions, getting results and managing results is a linear experience,” he said. “But you kick into a completely different plane when you begin to share and publish those results The challenge is to pull these things together into experiences that are naturally integrated into the user’s online activity.” In the past, personalization of search has produced a lot of red herring projects that had little to do with the tasks most users want to perform, he said.
“It’s not about who’s the first mover,” he said. “It’s about who can deliver the thing that really hits people in the heart.”
Network
