Wolfram Alpha's niche continues to elude

Stephen Wolfram's most ambitious project to date has yet to make a dent in the search market. However, the Internet search problem is far from settled.

Tom Krazit Former Staff writer, CNET News
Tom Krazit writes about the ever-expanding world of Google, as the most prominent company on the Internet defends its search juggernaut while expanding into nearly anything it thinks possible. He has previously written about Apple, the traditional PC industry, and chip companies. E-mail Tom.
Tom Krazit
5 min read

One year after its debut, the world is still not ready for Wolfram Alpha.

Few would argue that despite the success of Google, Internet search is a solved problem. The way that content is being shared across it is evolving so quickly means that better ways of discovering and presenting that content will always be welcome.

Stephen Wolfram
Wolfram Alpha founder Stephen Wolfram stephenwolfram.com

Wolfram Alpha certainly provides a different way to think about Internet search. It's heavily weighted toward computational queries, and its practice of curating its results as opposed to simply serving up whatever is available on the Web means its results can be more authoritative than a list of links.

But that strategy--useful as it might be to researchers and technical types--hasn't resonated with the general public. ComScore's assessment of unique users to wolframalpha.com over the past year shows that fewer people visited the site in April 2010 than did in May 2009. That traffic last year was undoubtedly juiced by curiosity and media attention, and usage has risen since a trough in late summer 2009, but as a search provider Wolfram Alpha doesn't even register on ComScore's radar.

Changes are coming that might boost Wolfram Alpha's profile among those without Ph.D.s. The company plans to make over its home page, and will start adding data for more pop-culture-friendly information such as sports, music, health information, and even its own take on local mapping.

"We're building something that understands the utterances that people put into it," said Stephen Wolfram, founder of Wolfram Alpha's parent company Wolfram Research, makers of the famed--and lucrative--Mathematica software. "We have to listen to all those users."

Storming the gates
Wolfram Alpha's debut was a bit hairy: the site decided to broadcast its launch event to the public in the middle of a tornado watch at its headquarters in Champaign, Ill., and networking and database problems subjected Wolfram and his engineers to much heckling from the digital peanut gallery.

But once the kinks got ironed out, Wolfram Alpha settled into a role as a curiosity within the ranks of Web searchers. The main problem was that people didn't understand how to use Wolfram Alpha. They were used to keyword-generated searches on Google, Yahoo, and Bing, which isn't necessarily the best way to produce results on Wolfram Alpha.

They also didn't understand exactly what was produced in Wolfram Alpha's results, so conditioned to expect links to a destination. Unlike other search engines, Wolfram Alpha produces its own search results by analyzing a search query and examining a database of content reviewed and approved by subject matter experts.

Wolfram Alpha failed to return a response to a given query nearly one third of the time a year ago. The company has reduced that error rate to about 10 percent, and Wolfram also points out that 50 percent of all queries that Wolfram Alpha processes return no results in traditional search engines, meaning it's answering questions that others can't.

"What we're trying to do is encapsulate expert knowledge in everything, and expert knowledge involves experts," Wolfram said. "Data you find randomly lying around the Web is data you don't want to use."

There are only about 250 people working full time on Wolfram Alpha, which means the company has to rely on an outside network of volunteers to help verify the data it presents to searchers.

Therein lay some early frustration: there simply wasn't enough expert-reviewed and approved content on the site when it launched. Over the last year the company has doubled the amount of curated information available through the service, but Wolfram acknowledged the difficult dilemma it faced last spring between launching with less information than practical and needing human input to refine search queries.

Climbing the charts
For Wolfram Alpha to grow as a service, it needs to attract people other than mathematically inclined researchers.

So with the first anniversary, Wolfram Alpha has expanded its content. Local street maps will be available on the search engine, and--perhaps a little less useful but kind of cool--weather information for outer space.

Wolfram Alpha poker
Wolfram Alpha: the poker player's ace in the hole? Screenshot by Tom Krazit/CNET

The aforementioned pop-culture data will be more comprehensive than before, perhaps allowing baseball researchers to start using the service as more of a tool than an amusement. Wolfram Alpha will also help gamblers calculate their odds, reassuring poker players that they have only a 1 in 2,548 chance of drawing poker's worst hand (7,5,4,3,2, off-suited, of course).

Wolfram Alpha has also decided to borrow a page from the traditional search engine competitors by offering an alternative query if the one you originally entered produced few or no results. They're calling it the "nearest" query, similar to Google's "Did you mean...." link.

And in hopes of attracting more mobile users, Wolfram ruefully admitted that its extremely high price point for its iPhone app--initially priced at $49.99--will not deviate from its current price of $1.99.

"There was a certain reaction to that price: one was developers saying that somebody is actually charging a substantial amount for a useful app, and the other side was people saying, 'you are crazy,'" Wolfram said. "Probably the second group was right."

Padding the coffers
One thing about Wolfram Alpha that sets it apart from other start-ups that have tried to challenge Google is that it's profitable. "Last year it made money, this year it should make money," Wolfram said, although he declined to provide specific numbers.

A content deal with Microsoft's Bing certainly helped that bottom line, but Wolfram Alpha has also signed sponsorship deals with some advertisers, as opposed to more traditional search advertising against search results, the way Google has made its fortune. Wolfram wouldn't comment on other potential licensing deals, but noted "one of the things that's been really nice about this, we've kind of had the opportunity to interact with all of the obvious players."

Profits are not what motivate Wolfram in the search market, but obviously they are nice to have. For him, the challenge is more about trying to tackle an insanely large project and actually pull it off; in this case, trying to build a system in which knowledge can be computed like mathematical equations.

However, Internet search is a relatively static market at the moment. Microsoft and Yahoo are spending millions to try to get traction against Google with little success, as most people seem content to stay with what they know.

That means it could take a very long time for Stephen Wolfram's hobby to turn into something that truly shakes up the search industry.

"I'm a guy who spends my life doing long-term projects," Wolfram said, referring to the decades of development that have gone into his Mathematica software. "Wolfram Alpha will be a similarly large project."