Ask.com Rediscovers Its Roots as a Question & Answer Site-Powered by People This Time

11/16/11Follow @wroush

What’s the best way to tell my mother-in-law that we’re not going to visit her at Thanksgiving?

My college friend has taken over my friendship with another guy, and now both are ignoring me. Should I ask them why?

What do you do if the girl you asked to marry you says no?

[Questions posed by Ask.com users, November 15, 2011]

Google, the Mountain View, CA-based search and advertising giant, may have a 66 percent share of the search market, but there are plenty of questions its Web crawlers and ranking algorithms just can’t answer. That includes squishy, emotion-laden questions like the ones above, as well as detailed, local queries like “what’s the best place for Mexican take-out in St. Cloud, Minnesota?” For those types of questions, there’s just no substitute for bringing real humans into the loop.

Across the bay from Google, in downtown Oakland, there’s a 15-year-old search company that’s now banking on the power of these human connections over algorithms. It’s called Ask.com, and it’s the perpetual fourth-place finisher in the search engine races, after Google, Yahoo, and Microsoft’s Bing. Last year, Ask.com gave up on its long struggle—financed by Barry Diller’s IAC/InterActiveCorp, which purchased the company for $1.85 billion in 2005—to compete with Google in traditional Web search. Under Doug Leeds, who became CEO in mid-2009 after a stint at the helm of IAC property Dictionary.com, Ask.com decided to transform itself into a question-and-answer service. And as of this fall, the transition is complete: in September the company opened the live Q&A portion of its site, previously in beta testing, to all 63 million monthly visitors.

Ask.com's iPhone app

The change marks a return, in some ways, to Ask.com’s heritage as Ask Jeeves. Early in its history, the company’s claim—symbolized by its butler mascot Jeeves—was that it could find definitive answers to natural-language questions such as, “Who was the world chess champion in 1956?” or “What’s the difference between a leopard and a cheetah?” The problem was that the claim didn’t always hold up; the computational problem was just too hard, often leaving Jeeves at a loss. (Even today, search utilities like Apple’s Siri can only answer natural-language questions reliably within restricted neighborhoods of knowledge, such as the weather.) In 2001, Ask Jeeves bought an algorithmic search engine called Teoma and started spewing out long lists of blue links just like Google’s.

The company continued to welcome questions phrased in natural English, and it tried to siphon traffic away from Google by marketing itself as a question-answering site. But “we started talking to users in a way that was different from the way they wanted to talk to us,” Leeds told me in an interview this summer. “What they really wanted was an answer at the top of the page, not a list of links.” Even the so-called “smart answers” that Ask.com offered at the top of a search result page weren’t addressing the need. “As good as we think we got at this, at least 60 percent of the questions that come in every day are things we can’t answer from the information we crawl,” says Leeds.

It turns out that people are really good at coming up with questions that have never been asked or answered before. So Ask.com decided to shift strategies again. “Toward the end of 2009 and through 2010, we built a platform not just for searching documents, but for searching people—to give them a chance to answer the questions,” says Leeds.

Today, when you arrive at the Ask.com website or fire up the Ask.com iPhone or Android apps, you see a big search box where you can type in your question. If the service can find a definitive answer somewhere in the database of 700 million question-answer pairs it has culled from the Web, it will simply show you the answer. (Question: Is a Granny Smith apple more tart than a Red Delicious? Answer: Yes.) If it can’t, it will show you a list of related Web results. And if those results don’t answer your question, you can click on the “Q&A Community” tab, where you can submit the same question for review by other Ask.com users.

There’s no guarantee anyone will answer your question, but in my limited experiments, I’ve always received at least one answer within an hour. “The biggest challenge for us is how to integrate these two services,” Leeds says—meaning the Web answers that arrive in milliseconds and the human answers that might take minutes or hours. But the social media revolution, where tweets and status updates and text messages rule, is working in the company’s favor, he says. “Users are becoming more and more accustomed to publishing in short spurts—just an answer rather than a whole page.”

Before it could rediscover itself as a Q&A service, Ask.com endured a few years of struggle, Leeds says. Former CEO Jim Lanzone left in 2008 to found Clicker, the Internet guide to TV programming. (He’s now president of CBS Interactive.) Lanzone’s replacement was Jim Safka, the former CEO of IAC property Match.com. According to Leeds, under Safka, the company invested heavily in marketing and in new ways of attracting search traffic. IAC bought Dictionary.com, the leading online dictionary, on the theory that it could be retooled to send downstream search traffic to Ask.com. Leeds, a vice president within Ask.com, was sent to head up that effort, which worked—but the dictionary site didn’t turn out to be a very strategic part of Ask.com, and it’s now run separately.

At the same time, Ask.com was blitzing the airwaves with TV commercials and became a NASCAR sponsor. There was less emphasis on innovation and product development during this period, Leeds says. Safka left in May 2009, and Leeds was asked to take over. “By the time I came back there wasn’t much left of the product organization,” he says.

Leeds decided to stop spending money on advertising and listen more to users. “It was clear to me that we are an Internet company and that we are going to compete by building what users need,” he says. “We are still number six in the U.S. in terms of the number of people coming to the site every month. Why not stop telling them what they should do and find out what they want and build that? That’s what we did.”

All of the company’s focus-group research indicated that brand loyalists still associated Ask.com with Ask Jeeves-style question answering—and that they chose Ask.com over Google whenever they needed specific answers to well-defined question. “Their intentions may have been the same, but their expectations were very different,” says Leeds. “A user of Google wants all the information you have—it’s like going into a library and asking for the card catalog and going back and forth to the stacks. A user of Ask.com wants to go in, ask the librarian a question, and get the answer.”

From that point on, Ask.com’s goal was “to have an answer for every question that comes in,” says Leeds. He hired a new team of engineers to rebuild the company’s technology. By combing the Web for existing question-answer pairs, “We went from about 20 percent of questions having an answer at the top of the page to about 60 percent,” he says.

But there was still that pesky 40 percent left over. “It wasn’t that no one knew the answer; it was that no one had thought to publish a Web page or a blog about it,” Leeds says. “At the time, social media and Facebook were taking off, and users were coming around to the idea that they could get what they were looking for from other users. That led us to believe that we had to do something to combine these things.”

And that’s what Ask.com has done with its two-tab system for Web and community searches. Of course, it’s hardly alone in the question-answering market: there are old incumbents like Answers.com and Yahoo Answers, upstarts like Quora.com, conversational sites like Formspring, and specialty search engines like ChaCha. Of all of these, ChaCha is probably the most similar to the new Ask.com. Its answers are supplied, or at least located, by paid human guides. The new Ask.com “isn’t that different from what ChaCha is doing,” Leeds acknowledges.

But Ask.com isn’t paying anyone (although it does have a large subsection of users raising money for breast cancer research at the Susan G. Komen For The Cure Foundation; it donates 25 cents to the foundation every time one of these members answers a question). Even without paying, “It’s not a problem getting people to answer questions, as long as you can give them a relevant question to answer,” Leeds says.

Building the smartphone apps has been another huge boon to participation, Leeds adds. For reasons that aren’t completely clear, people who connect from their iPhones or Android phones answer four times as many questions as those connecting from a desktop browser.

The company still earns money the same way it always did: by showing a combination of banner ads and text ads. “It’s 99 percent keyword [ads] and 1 percent display,” says Leeds. Most of the keyword-based ads arrive via Google’s Adsense program—which makes Ask.com one of the many Bay Area companies that probably wouldn’t exist if they didn’t have Google’s wide shoulders to stand on. But as long as Ask.com can provide a service that Google doesn’t, it will have a place in the search market, Leeds says. “To this day, with all Google has done, they still don’t understand questions very well,” he says. “It’s too hard to do what Google wants you to do: run a search, check a page, click back to search, iterate. It’s much more important to get an answer, right when you’re asking it. All of these things are coming together to get us the best growth we’ve seen in a long time, and real traction with users.”

But why, in the end, do people spend time answering questions on Ask.com? The site doesn’t hand out any kind of payment or reward, other than a thank you. There are no points or badges or other attempts at gamification. Leeds thinks part of the answer is that people simply like to be helpful while they’re killing time. It’s more productive than playing Angry Birds, after all. “It’s become a thing that people like to do if they’re waiting for a meeting,” he says. “You can browse for questions, see if there’s something you know how to answer, and answer it. There’s not much other incentive, except for the ability to show you know what you’re talking about.”

But to find the definitive answer, I went to the users of Ask.com themselves. I posted the question “Why do people answer questions on Ask.com?” and within 15 minutes I’d received 14 responses. Within an hour, there were 25. I’ll let these selected Ask.com users have the last word. (I’ve rounded up more of the answers in a separate post today.)

GuardGirl: I guess we’re just curious to see what other people are thinking and i personally like to help people:)

Daisy46: It’s fun! and i like to give people accurate information that will actually help them. It makes me feel like I am making a tiny positive difference in that persons life.

nadarkins: I don’t really know why but it’s kind of addicting. I guess it’s a fun thing to do if you have a little bit of down time

FormerGlory: It’s therapeutic to help people, make friends, and make people laugh.

Dozy: Lots of reasons…I enjoy writing, so many of my answers push that 1200 character limit. I’m not sure how many people actually read the stuff, but I have fun writing it. Then, I like to be creative so I often pull a question out of context and answer it from a different point of …But mostly it’s for the brilliance of the comments threads. A lot of us are pretty average but some of the people on this site are brilliant, and I enjoy their input.

VirginiaL: I came late last December, and was intrigued by this amazing, interactive software that is so immediately responsive…you can post your question, and within minutes or hours responses come in from all over the world. It is a wonderful sense of very high-quality social connection with people everywhere, and you never need to leave home. The questions, answers, and discussions really help you develop your power of reason and insight. Also, there have been times when this website has very possibly saved a life…someone comes on in deep despair, asking for suggestions, opinions or resources…and a dozen people respond in support…There have been (and still are) problems, mostly as the software continues to be refined, but those of us who hang in there have definitely felt it is worth it.

Wade Roush is a contributing editor at Xconomy. Follow @wroush

By posting a comment, you agree to our terms and conditions.