Last Updated Aug 5, 2010 9:26 PM EDT
So why is Ask.com (IACI) trying to put a human being between me and my precious information?
The venerable search engine reinvented itself last week: no longer a purely algorithmic search, it has morphed into a human-powered question-and-answer service, joining a handful of like-minded competitors like ChaCha, Squidoo and Mahalo.
The reinvention comes at a time when the smartphone has (at least, in theory) rendered a lot of searching unneccessary. If you have apps shortcut straight to your mobile banking, your Google (GOOG) maps, and your hyperlocal information (like, say, Yelp reviews), then what's the use of asking a question?
According to Doug Leeds, Ask's President, all this information -- location, buying habits, friends -- doesn't actually make searching any easier. It only provides information that can narrow results -- something that isn't currently being done well by other engines.
"I think asking questions is the best way to get information. If search engines didn't exist -- and for most of human history they didn't -- the way you got information was asking for it," he says.
Google, he believes, has trained us to find information in three steps. "You just give Google the 2.1 worlds to look for, and it'll find you lots of content with those words. Google is about finding information that can answer your question; they shorten answer between question and answer, and deliver you to the right part of the web." Then you're left to do the filtering. (Squidoo and Mahalo work on a simplified but similar model.)
That, Leeds says, is two steps too many. He believes we're accustomed to searching for the "right part of the Web," when we should be (gasp!) hoping to zero in on the exact information we want. Searching works well enough, but it's an inexact science. (Of course, Google's founders admitted long ago they have capital incentive to make it inexact. But still.)
Leeds believes that the tools we now use (Google Maps, product search, and so on) are too manual. "These [tools] are simply contexts to content. What sandals should I buy? That question is bounded by price," he says. "With location search, you're bounded by area. What we've found is that the best starting place is still asking a question."
Once the Ask system generates answers, he says, it can layer on the context or have the parameters implied. "A phone should know know you're in lower Manhattan, and only look for stores in lower Manhattan," he says. Although this kind of functionality isn't working on Ask.com just yet, Leeds says it's in the pipe. "We are definitely building in location awareness." (Leeds also says Ask will go mobile "very soon.")
The unspoken disruptor here is the phone, which has made traditional search a pain. On a desktop, Googling is easy: you search, look at results, and search again. But on a phone, mistakes feel more costly. Connections are slower; typing is more laborious. Ask is hoping that for mobile users, accuracy will drive a sense of primacy.
Others have ventured here. ChaCha works most similarly to the way Ask has rebuilt itself, with a network of human operators all over the country who answer questions in areas of their "expertise." In this case, expertise doesn't require extensive knowledge so much as an extensive facility for finding the right knowledge. "We're asking our subject-matter experts simple questions," says Leeds. "The idea is to not overwhelm them. We ask: what are your hobbies? Where have you lived? What industry do you work in? What do you spend your weekends doing? A little down the road, we may begin inferring that stuff from their social Web presence, too."
Right now, Ask is human-answering a big portion of its queries, but not all. It has also seeded its "subject matter expert" community with some paid members to start, although later, this will revert to a Wikipedia-like free contributor network. ChaCha, by contrast, pays its experts per query, and focuses on text-message based questions.
Unlike ChaCha, which uses a detailed taxonomy to route questions, Leeds says Ask's network will be more holistic. "We're not using any kind of forced taxonomy with respect to sorting by topic. We use what we call a topic cloud, or 'bag of words.' Using search history, we can see patterns; if you're doing a search on earthquakes the next search you may do is survival. We use those search patterns to create topic clouds." Experts are grouped with a similar method. "We also create clouds for interests of people answering questions, which helps determine what [questions] we should send them."
The main difference is the construction of the "community" that provides the answers. ChaCha, while effective, hasn't exploded in popularity. Ask may have a better shot because of its brand recognition and its legacy in search. The way it uses "members" -- who have to sign up to ask and answer questions -- might also make a difference. If members are answering questions, not paid employees, then they become increasingly involved in the community. When new members arrive to ask questions, they have the opportunity to participate. That network effect could keep users coming back.
Like so many other Web tools, half of Ask's potential lies in how it can be used to leverage other services. Leeds believes that, like maps and shopping, other tools will fit nicely under the purview of Ask. Twitter and other real-time tools are one example. "Twitter is great for finding out the moment something happens," he says. "The earthquake is a great example. Once you know there's an earthquake, you might say, 'what should I do?' That question is very hard to answer with Twitter, which is just a firehouse of information. Nor is it a great question for Facebook. But if you ask it on Ask, you can route it to people who know about earthquake response."
If natural disasters don't get people interested in faster search, well, then perhaps Google has won once and for all.