Just in case you weren't paying attention, Google just turned 10 years old. It's been a long time since we first got lucky, right? This monolithic super-company is clearly winning the search engine war, with between 60-80 percent of the market. So, just a thought: Can we please retire the term "Google killer?"
The phrase is a cliché -- a tired marketing slogan. It's also a fantasy, just like there was and still is no Microsoft killer. These two companies are so entrenched in the worldwide computing experience that no single company is going to knock them out completely. Additionally, consider a few of Google's recent successes beyond search: Chrome, Gmail, Gchat, GoogleDocs, GOOG411, Google Maps, Google Earth, Picasa, AdSense and AdWords. All useful, popular applications. So, why do we want a Google killer again?
No one is going to kill Google. But that doesn't mean new companies aren't looking to step on Google's toes while they carve a space for themselves. Three young search engines -- Blinkx, Mahalo and Cuil -- all have had their own share of recent success in the search marketplace, and each is doing its part to make the internet a little less Google-centric. This article outlines the role each of these engines plays in the search marketplace, as well as some of their recent successes and continued shortcomings.
Consider Christopher Columbus; he couldn't discover everything. Amerigo Vespucci, Francis Drake and even Daniel Boone all had to do their parts. So, it's not implausible that a smart new search engine could discover some new geography online.
"Our focus is video search, and when we say video search, we mean a single jumping point for video across the web," says Suranga Chandratillake, founder and CEO of Blinkx, who believes his company has found such an untapped new space. Blinkx's goal: to index the millions of videos that are online, determine the content of those videos and make them all searchable.
Video search to Blinkx is not like how we normally think of video online. The main function of YouTube -- owned by Google -- and the dozens of other video sites is simply to host content, not search it. "It's ironic, because Google usually means search," Chandratillake says. "Google is more of a Yahoo model -- a lot of hosting, and a little bit of searching here and there, but primarily a hosting service.
"The few attempts there were before we came along for worldwide video search were based entirely on metadata, so the text, the descriptions, the tags," he adds. And metadata, as Chandratillake puts it, is "at best, incomplete; at worst, directly misleading."
Instead, Blinkx uses a video's actual content to search. Blinkx first spiders the web, following every video link it can and scraping together information from where the video resides. It then processes each video at a basic level (audio, high quality, etc.) Then the video is processed by Blinkx's audio engine, which is capable of many English accents and other popular online languages. The engine picks out what is being said on a phonetic level. Blinkx also processes the video visually, looking for a library of famous faces among the digital images and breaking up the video into different scenes -- almost like a DVD chapter search -- with previews for each.
"All of this data is combined into a single record in the Blinkx index," Chandratillake says. "And this record is pretty interesting because, on the one hand, it has all this information about the video -- this person is in it, these things are said, here is metadata, etc. -- but it also knows when some of these events happen. It knows not just what words were said, but when those words were said."
Don't you wish YouTube could do that?
Despite having 26 million hours of video indexed and handling video search for dozens of major media clients, Blinkx is still running under the radar here in the U.S. So Blinkx isn't exactly competing with Google -- yet. Will Google join the video search fray? Blinkx isn't waiting to find out.
Rather than discover a new marketplace, Mahalo is trying to rethink the search result. With Google, we get the most popular linked-to sites (approximately) related to our search. A few of the top 10 might be no good, so we scroll past them and onto the next page. We might even go a few pages before we find what we want. Mahalo wants that first page of links to be practically definitive.
In essence, Mahalo is a community-involved search engine, with the editorial team creating individual pages -- based on popular search data -- listing the best possible links, all tested and useful. Mahalo encourages users to take part in improving those search results by submitting pages and links themselves. The idea is to make every page complete and every link great, with no spam, repeats or other unrelated nonsense.
As Mahalo CEO and President Jason Calacanis explains it, "Our world view is that the future of search is equal parts search, content and social bookmarking. No one has threaded the needle on these three services yet, but whoever does figure it out will have a $10 billion business." It stands to reason he's betting his company can do it.
"We have almost 100,000 pages, and we are ramping up to the point at which we can do thousands a week (if we wanted to)," says Calacanis. "The future of search will be getting a hand-curetted result for the fat part of the long tail and machine/social results for the rest of the tail. So, it doesn't have to be all human all the time."
But he's still trusting humans to help bear the weight. By letting users online contribute pages and links, Mahalo is adopting the Wikipedia model, putting power into users' hands to improve their own experiences and the level of information they can obtain. People put value in that; it shows respect for and a willingness to work with the customer. That appeal could be a game-changer.
The main problem with Mahalo is that most search terms don't have pages yet. That will change over time, given that Mahalo is only 14 months old, and the editorial team is building new pages every day. Can Mahalo reach its goal of snagging between 1-5 percent of the search engine market in the next five years? It's unlikely, but no one saw Google coming either.
Still, Mahalo is under no misapprehension about its current place in the market. "We are not competitive with Google; we are partners," insists Calacanis. "Google is the operating system of the internet, and we are an application that sits on top of Google's platform." He sees Mahalo as complementary, not competitive. "Most folks use two to three search engines a month, so we only have to take 1 or 2 percent market share to have a huge, multibillion-dollar business. Of course, that will take years to do. It's a mistake to think that you have to get people to 'switch.'"
User involvement could change the way we search. With Wikipedia, user involvement certainly changed the way we looked for knowledge. With YouTube, it changed the way we watched videos online. With MySpace and the like, it changed how we kept in contact with friends. People want to be involved, and Google is ultimately a closed search system. Will Mahalo become a friendly place we can call home, at least now and then?
In Cuil, we have a search engine that competes with Google even though the company says it doesn't. But Cuil is going directly after what made Google so popular: a huge index of websites providing relevant search results. But Cuil is trying to go bigger and better.
"Pure search, that's our focus," says Vince Sollitto, vice president of communications at Cuil. "Here is a search engine doing its own crawl of the web, trying to crawl the whole web, and is basically offering web-wide, consumer-oriented automated search. I don't know of any other search engine startup that has really tried to offer that in the recent past. Because of what we're attempting to do and what we're seeking to offer, we immediately get lumped into the category of Google, Yahoo, Ask, Microsoft, etc., because that was what we're offering. But obviously, we're two months in versus 10 years."
Point taken. While the visual differences of Cuil -- a black background, a magazine-style layout, longer snippets on individual results and corresponding images next to each result -- are evident, search engines live and die by their results. Here, Cuil is aiming high and claims to index 120 billion websites-- three times more than any other search engine. The engine also claims to provide more accurate, content-based search results.
On the surface, 120 billion pages looks like a big number, and more is certainly better when searching. But Google announced in July that it maps more than 1 trillion URLs. That's not necessarily all unique pages, or even pages, that Google indexes, but still -- a trillion with a "T." We ultimately don't know how many pages Google indexes, so although Cuil claims advantage here, it's difficult to determine a winner.
Even if there's more to crawl, you still have to find the good results. Cuil approaches this differently than Google, preferring to rely on page content rather than popularity. Sollitto says Cuil uses "very intensive data mining and algorithms that learn based on patterns and relationships between data sets, to determine what's the content on the page, what it's about." Cuil's process puts pages into various categories, and when people search, the engine delivers results from various related categories. Users can then "drill deeper" by browsing through a category.
While the idea of working from page content is great -- rather than relying on the mass appeal of any site -- the results aren't always that spectacular. Sites you don't expect may pop up, exposing you to new resources. But plenty of sites that you would expect don't turn up. That said, Cuil's search engine is only a few months old, so we'll cut it some slack.
Directly taking on Google, whether Cuil admits it or not, takes some courage. For that reason alone, it's worth keeping Cuil on your radar. If the content-based search improves over the next year or so, Cuil could start grabbing a small percentage of the market. The other dilemma is this: You can Yahoo, you can Google -- but can you Cuil?
Each of these three search engines has one other big advantage over Google: privacy. Since Google is our default search engine, it has access to mountains of search information, tied to IP addresses and kept on record for years. Privacy advocates continue to be furious about the amount of information Google keeps on its users. But what if we didn't have to worry about privacy from the get-go?
"Our approach is we don't retain any information about you at all," says Chandratillake of Blinkx's approach. "We just don't want to open that can of worms. We just don't think we need it to provide great relevance, so we'd rather not do it."
Cuil is much the same. "Because we provide results based on content analysis rather than popularity, we don't need to retain and analyze our users and their information and their IP addresses and their search history," says Sollitto. "We're not really interested in what is most popular; we're interested in what's most relevant."
Calacanis of Mahalo notes that, while privacy is a concern, it's just not that big of a deal to Americans. "The sad fact is that users in the U.S. don't currently care about privacy," he says. "The European market will be the one to keep U.S. internet companies in check with regard to privacy policies, it seems." While Mahalo does track all searches in its system, it doesn't tie the searches to IP addresses and throws out any IP data after 90 days.
It's a changing era, and for privacy nuts, this is good news; these search engines are disarming this thorny issue before it becomes a problem. Google should take note.
David and Goliath fought on a level battlefield, but Google doesn't. You can't knock out Google with a rock. (Well, maybe with a lot of rocks aimed at Google's server farm.) But regardless, Cuil, Mahalo and Blinkx are not rocks, nor are they Davids, so it's a good thing knocking out Google isn't necessary.
These new search engines know that Google has basic search down cold, so all three have decided to tackle something more difficult. Blinkx wants to do for video what Google did for text. Mahalo wants to involve the community in its searches, ensuring every link is a great one. And Cuil wants to index the entire net. It's simply what startups do: attempt that which hasn't been done yet.
Blaise Nutter is a freelance writer.