The answer man

Bob Rosenschein, the man behind GuruNet and its Answers.com reference, talks about his company's complicated relationship to Google and the future of info-gathering on the Web.

Paul Festa Staff Writer, CNET News.com
Paul Festa
covers browser development and Web standards.
Paul Festa
9 min read
GuruNet CEO Bob Rosenschein will be the first to tell you he doesn't have an answer for everything.

Rosenschein is content with answers to the million topics included in Answers.com, GuruNet's 2-month-old reference site. If you ask Rosenschein how his piddling million topics can compare with the 8 billion pages indexed by Google, Rosenschein will revert to his mantra: Answers.com is not a search engine.

Instead, the company--which survived an ill-fated foray into enterprise search--sees itself as an online compendium of reference sources, collecting on a single page about 100 information sources including encyclopedias, dictionaries, almanacs and search engines on topic pages about English words and notable people, places and things.

Rosenschein insists his company isn't competing with the search engines. But even as it gets closer to Amazon.com's A9 search site and to Google itself, GuruNet is inevitably carving out a search niche of its own that could turn the company into a serious threat or a valuable acquisition target.

Q: Back in 2000, when you gave up the consumer market for the enterprise market, your then-CEO explained the decision this way: "Under our old consumer model, we were too good at getting you the information and getting out of the way to succeed at monetization. That's what people loved about us. But from an ad standpoint it didn't work." What's changed since then? And how are you making money?
Rosenschein: I think that Google has pretty much validated that you can make money from consumer ads on the Internet. The Internet has changed since then. I think he misjudged the situation. The subscription model is under more pressure now than it was then, and the general Internet consumer side is far more robust than it was several years ago.

We are not a search engine. We are the un-search engine.
Usage is up around the world; the number and length of queries is up; the amount of time spent on search engines is up; broadband connectivity is up. And it's not just faster connectivity but more pervasive connectivity. All these things are very good for the consumer Internet.

You frequently say you're not in competition with Google. But every time I Alt-click a term using your plug-in is a time I'm not Googling something.
Rosenschein: I think we're not head-to-head competitors. We are different. The search engines are different--it's built into their DNA to locate interesting and relevant Web pages, and the way they do that is by showing you lists, often long lists, of Web page links. That's what they do. Some show you advertising more prominently, some

show you sponsored links on the top, or to the side. But their basic function is to present the user a list of Web pages to navigate serially.

Our goal is to deliver concise, relevant information in one click. If you're looking for Web pages, use a search engine. If you're looking for rapid, concise explanations and definitions, then try Answers.com. It's a different value proposition. It gives me useful information in fewer clicks. It gives me the information straight. We're not the best place to compare prices of MP3 players, or to look yourself up if you're not in our database, but if you're looking for one of the tens of thousands of famous people, companies, or topics in science and technology, you can get that information from us more quickly.

Look up Intel on Answers.com. You get its history, its profit and loss, the number of employees, the address, the stock price. I defy you to find half that information in five minutes on the Intel site.

The Web, for better and for worse, is largely unedited, unfiltered, cluttered, and, if you don't mind my saying so, has some offensive or inappropriate material. That's the nature of the Web and part of its democracy, and that's a good thing. But the average user is just overloaded. Most users are not looking through all the links. They're just hoping that the first few are relevant.

We provide a more edited environment because we've licensed material from known and credited sources. So the user might get a little less overload than from the search engines and will develop a sense of trust in what's presented. This is a big deal for families, students, teachers, and kids, but also anyone who writes.

When you say the goal is to get people information more quickly, Google could say the same thing.
Rosenschein: They couldn't say the same thing. It's a fantastic product which is excellent in locating long lists of relevant Web pages for you to navigate serially. But I'm doing it in one less click. I mean them no disrespect because it's a great product, and they do what they do very well. But their whole approach is to present you a list, and sometimes you don't want that. You want to know who is that, what is that? You just want the information.

What about Microsoft? Their research library in Office 2003 even has your Alt-click system. And now MSN Search is linked with Encarta.
Rosenschein: Again, Encarta is a great product; we're big fans of theirs. But you're accessing these reference sources one at a time. We've taken a hundred dictionaries, encyclopedias and almanacs and put them together on one concise page. Try looking up Steve Jobs or Bill Clinton. We've integrated all the forms of their names. So you

could look up William Jefferson Clinton or Bill Clinton or Clinton, Bill. You don't notice that. We call it the Fred Astaire principle. He made his ever-graceful dancing look so simple, but in fact he worked very hard to make it look that simple. We put an entire reference shelf at your disposal.

We are not a search engine. We are the un-search engine. We don't scour 8 billion pages. We have a million pages, and we're growing it.

How big do you expect to grow?
Rosenschein: At this stage I don't see that we're interested in 10 to 20 million topics. A million covers an enormous amount of territory, all the words in the English language, 15,000 public companies, 2,900 U.S. cities, 500 international ones. It's not just encyclopedia terms. We are actively adding new and useful content sources.

How do you determine what terms get covered?
Rosenschein: We choose a content source, like an encyclopedia, and then we unify it against the other content sources, matching appropriate terms and cross-linking. Sometimes it's more automatic than others. Which sources we use and which we don't is, of course, an editorial decision. We're growing both the straight reference sources but also some more interesting and dynamic things.

Like what?
Rosenschein: Technorati is one. It's one of the leading blog search engines. We thought it might be useful to give our users a way to search blogs, not just search the Web. It's an example of the kind of dynamic information we'll put in. We have currency conversions, weather for over 3,000 cities.

Are you an acquisition target?
Rosenschein: Of course we're committed to do whatever's in the interest of the shareholders. But we are aiming toward building a useful and profitable company, and we're happy to be partners with any of these larger companies. Acquisition is not our goal right now.

Do you have any patents?
Rosenschein: We have three patents and one pending. One of our most interesting patents involves our one-click tech and the ability to deliver promotional commercial information based on the neighboring words. If you use our one-click system to click on the word "Ford," you will get a different answer if it's followed by "Motor Company" than if it's preceded by "Henry" or "Harrison."

Microsoft has a one-click query system in Office. Are they infringing on your patent?
Rosenschein: Not to the best of my knowledge at this point. Remember, their system works within Office. Ours works across all applications.

GuruNet had a relationship with IBM. How did it work out?
Rosenschein: That was in a completely different lifetime. The entire company has shifted to the consumer space with the advertising

model. We survived long enough to go back to consumer.

How much do you think the foray into enterprise search hurt the company?
Rosenschein: In the past, it hurt a lot. In the long run, we hope we're perfectly positioned to go to the next step. We're extremely excited with user response to our product. People seem to love Answers.com because it provides different experience from the search engines. I don't see it as competition--it supplements and complements what they do. I'd love to partner with more of them.

One of your sources is Wikipedia. Why did you choose it, and what are your thoughts on it as an encyclopedia?
Rosenschein: We love it. We also use Columbia University Press and Houghton Mifflin, but Wikipedia has a currency and a breadth that's very impressive. It has a high average quality, but obviously because it's user input, not every one of the entries is going to be high editorial quality. We accept that, and we have a disclaimer to that effect. But it's a fast-growing and lively data source. We're proud that it's part of our offering.

How did you score the deal to provide Google's definition links?
Rosenschein: They evaluated our service and decided to point that traffic to it. We're very excited and hope it continues to drive traffic and grows, especially as we add more languages and things like that. It has helped in very positive ways, but it still provides well under half of our queries. In terms of the deal, there's no magic there, they just liked it. I think it underscores that we're a reference product, in their view, and there (are) many things they could do in that arena.

Like what?

You can only imagine that people with smart phones will be more interested in the kinds of rapid information we offer.
Rosenschein: Right now it's a definition link. I hope they get more interested in reference too. I'd love to license our Alt-click technology to them. There (are) lots of things they could do. They have a toolbar. We have many things we'd like to offer Google, and others.

What's next for GuruNet?
Rosenschein: Don't be surprised if you see us expanding to the wireless area. You can only imagine that people with smart phones will be more interested in the kinds of rapid information we offer. And don't be surprised in the future if we expand internationally.

Meaning sites for different languages?
Rosenschein: Right. We're very strong in English, but we get traffic from many dozens of countries every day. Most importantly, expect us to enhance our content offering, so we put more and more useful information in one place. All we do is try to get easy information to you. Look up Intel on Answers.com. You get its history, its profit and loss, the number of employees, the address, the stock price. I defy you to find half that information in five minutes on the Intel site. Or you could go to a search engine, which will show you Intel's site as the first result. I'm not knocking it. But Answers.com is just fewer steps to reliable information.

And what's ahead for the search industry?
Rosenschein: I think the whole area of search user interface is at an early stage. I think there are enormous innovations that are going to come in the next few years in the user experience, and they're almost as important as spidering 8 billion Web pages for relevance. I'm not knocking the search engines. But there's more to life than search engines.

What do you predict with respect to search user interface?
Rosenschein: Let's put it this way. You remember DOS. Tell me the difference between the average search engine interface and the DOS command line. You type in something, you get back results. Maybe people in 1982 couldn't imagine a Mac or Windows UI and they were perfectly happy with DOS. I think there's innovation ahead in the way people look for and retrieve and navigate their information.