Search Engine Shoot-Out
Whether you're searching for text, video, images, news, or local information, you have lots of options beyond Google. We tried dozens of search engines and found some worthy challengers to the king.Jeff Bertolucci is a freelance writer living in Southern California.
Jeff Bertolucci

Illustration by Steven Lyons
Which led us to wonder: Does Google deserve all that traffic, or is it living off its reputation? Are people using it because they're not aware of other, potentially better search engines? To find out, we pitted Google against its big-name competitors, Yahoo and Microsoft Live Search, as well as against smaller challengers such as AlltheWeb, AltaVista, and Ask.com--plus a couple dozen of the specialty search services, including Blogdigger, Picsearch, and TubeSurf.
Our verdict? Google is indeed the best search engine, even though two other services topped it--barely--in our text-search tests. Google's index proved to be the most accurate, comprehensive, and timely of the bunch. It also bested the majority of the specialty-search sites we tried, meaning those that focus on a category or file type, such as videos, images, news, blogs, or local info delivered on a mobile phone.
Recent enhancements to Live Search's mobile component moved that service into the lead in our test searches for local information, although you have to navigate manually to its mobile-optimized site rather than being redirected automatically when you log in from a cell phone or other handheld device (see our charts throughout this story for details).
That said, the competition is fierce--and Google had better stay on its toes. Its challengers are implementing some innovative tools and interface upgrades (Ask.com is particularly impressive in this area) that enhance the user experience and deliver more relevant information than do the standard ten blue links on a results page. We also like several useful tools that can help you go beyond the basics of search.
Chart: Text-Search Blanket Finish
In our tests, variation among the top four general Web search engines was minor, with Microsoft's Live Search coming in a distant fifth; Ask.com finished two points behind Live Search and ranked in sixth place. Click the icon below to see how the tested engines performed, including their scores in each round.
| Text-Search Blanket Finish | |||||
|---|---|---|---|---|---|
| In our tests, variation among the top four general Web search engines was minor, with Microsoft's Live Search coming in a distant fifth; Ask.com finished two points behind Live Search and ranked in sixth place. | |||||
| Service | First round | Second round | Third round | Average | Comments |
| AlltheWeb | 10 of 10 | 24 | 24 | 24 | This Yahoo-owned site blends its parent's powerful search index with a homely but easy-to-use interface. |
| AltaVista | 10 of 10 | 22 | 23 | 23 | Another Yahoo property that's a top-notch text-searcher. The spartan interface is more aesthetically appealing than AlltheWeb's. |
| Google Search | 9 of 10 | 21 | 20 | 21 | The goliath of Web search, Google still manages to be comprehensive, timely, and a breeze to use. |
| Yahoo Search | 9 of 10 | 20 | 21 | 21 | Great engine, but Yahoo's overwrought home page is the Vegas Strip of search sites. (The Yahoo engine also powers AlltheWeb and AltaVista.) |
| Microsoft Live Search | 8 of 10 | 16 | 16 | 16 | A notch below Google and Yahoo, yet a solid engine. But it isn't well integrated with other Live services the way Google and Yahoo services mesh. |
| Ask.com | 5 of 10 | 14 | - | 14 | A decent showing, but some results weren't as timely or as accurate as those from competing sites. Its experimental Act X interface shows promise. |
| LookSmart | 5 of 10 | - | - | - | - |
| Gigablast | 2 of 10 | - | - | - | - |
| Open Directory Project | 1 of 10 | - | - | - | - |
| Wikipedia | 1 of 10 | - | - | - | - |
| Chart note: The average score is composed of scores from three rounds of testing. In the first round we posed ten queries, awarding one point if a link to the target response--or the response itself--was included in the first page of results. In the second and third rounds, we awarded three points when the target response was available from the first result (or above it, in several cases), two points if the target was in the second or third result, and one point if the response was listed elsewhere on the first page of results. | |||||
Undisputed Champ?
If you use Google and are happy with it, you have no reason to switch engines. Simply put, Google is an excellent search tool. It scored at or near the top in every search category we tested, including the standard search for text information, as well as our searches for videos, images, news, blog postings, and local info. While other services performed as well or slightly better in specific areas, only Google aced virtually every test we threw at it.
Our initial round of testing included dozens of general-purpose and specialty search sites; the final two rounds focused on the top engines in each of our six categories from the initial round. See our complete test questions and results.
This isn't to say that the search sites that did poorly in our tests don't have noteworthy features. Ask.com deserves praise for its usability and its image searches, although in our text-query tests it was just so-so. Yahoo fared well in most categories, including text-info and local search, but it was a middling performer in image and video queries.
In addition to the big dogs, dozens of specialty Web sites are devoted to searching for images, news, blogs, and other specific items. For some queries, these boutique engines are a good choice. Superpages.com and WhitePages.com, for instance, excel at people searches. And TubeSurf, a metasearch engine that simultaneously scans Google, MySpace, Yahoo, and YouTube for videos, is a quick way to query multiple sites. In many categories, however, including video, image, and news search, the big engines were just as good as--and often better than--the little guys.
Maybe you've heard that search engines all spit out the same list of hits. There's some truth to this claim: For instance, when you query news-search sites about a major, breaking story, you can expect very similar results. But in our tests, the items returned varied considerably from engine to engine, particularly with image, video, and blog queries. Take images, for example. Our query for Mardi Gras photographs often produced a motley mix of hits, everything from amateur close-ups of bleary-eyed revelers--who, frankly, could have been partying anywhere--to professional-caliber shots of the Mardi Gras parade. And a search for the more-obscure "windform" produced a grab bag of thumbnails, from patio decorations to wood sculptures. (We were actually looking for an image of a musical instrument called a windform, which is a horn that measures about 20 feet long.)
Blog-search results were equally unpredictable. In one test we queried "Huffington Libby" (without typing the quotation marks), our target match being an article from The Huffington Post blog on the trial verdict for former White House aide Lewis "Scooter" Libby--breaking news at the time. At several of the blog-search sites, such as IceRocket, Sphere, and Technorati, the results didn't include links to the target story, but rather pointed to articles on other blogs that mentioned The Huffington Post's Libby piece. The lesson here is that some specialty sites are worth the detour, while others are not. For the record, Google topped our blog-search tests, followed by the specialists Sphere and Blogdigger.
Our scoring awarded a search engine three points if the first link in the results led to the target answer or site (or if the answer itself appeared at the top of the results page or within the first result). A link to the correct answer as the second or third result was worth two points, and a link to the correct response elsewhere in the top ten results scored one point. (For image search engines, which typically include more than ten thumbnails per page, we awarded a point if the target photo appeared on the first page.) If the target answer or item didn't appear anywhere on the first page, we awarded no points to the search engine.
With some queries, we were seeking a particular news story or blog article. We deemed other answers accurate as long as they had the requested information--for example, multiple sites contain a photo of Rutherford B. Hayes, the nineteenth president of the United States.
Some tests forced us to make judgment calls. A mobile search site that required us to enter an exact street address (for example, "123 Main Street, Oak Park, California") as the starting point for driving directions didn't score as well as one that needed only a city and state.
Chart: Scan the Blogosphere
Among the blog-search specialists, only Sphere challenged Google for timely, accurate results. Click the icon below to see how the tested engines performed, including their scores in each round.
| Scan the Blogosphere | |||||
|---|---|---|---|---|---|
| Among the blog-search specialists, only Sphere challenged Google for timely, accurate results. | |||||
| Service | First round | Second round | Third round | Average | Comments |
| Google Blog Search | 18 | 16 | 20 | 18 | This service was more up-to-date than the blog-search specialists. Results pages show the time of blog posts (e.g., "12 hours ago"), too. |
| Sphere | 15 | 13 | 20 | 16 | Timely and well designed, this site's top-of-screen links make it easy to narrow your search to, say, the last 24 hours or seven days. |
| Blogdigger | 14 | 9 | 16 | 13 | Blogdigger has an easy-to-use interface, but the service sometimes couldn't find same-day blog posts in our tests. |
| Bloglines (Ask) | 11 | 10 | 13 | 11 | This site offers news feeds and blogging tools, but its search feature seems like an afterthought, with a window tucked in the right corner. |
| IceRocket | 12 | 8 | 11 | 10 | We like IceRocket's search results page, which displays blog titles, article summaries, and post times in an easy-to-read fashion. |
| Best of Blogs | 10 | - | - | - | - |
| Feedster | 10 | - | - | - | - |
| Technorati | 10 | - | - | - | - |
| Hubdog | 0 | - | - | - | - |
| Chart note: The average score is composed of scores from three rounds of testing. Our scoring awarded a search engine three points if the first link in the results led to the target answer or site (or if the answer itself appeared at the top of the results page or within the first result). A link to the correct answer as the second or third result was worth two points, and a link to the correct response elsewhere in the top ten results scored one point. If the target answer or item didn't appear anywhere on the first page, we awarded no points to the search engine. | |||||
What's New in Search
It's hard for many Internet users to remember the pre-Google Web, but there was a time when AltaVista ruled search, and before AltaVista, Magellan was the search service of choice. The history of the Web shows that people will switch to another engine in a heartbeat if it produces more accurate hits on their queries and is easy to use. So search vendors realize they need to be on top of their game. "The amount of competition that's out there right now is benefiting the consumer," says Microsoft Live Search senior product manager Justin Osmer. "We're all trying to come up with the next great thing--the Cool New Thing--so you're seeing a lot of innovation in search."
Much of that innovation centers on how the services display results. Few people use their favorite engine's news, image, and other specialty search tools, even if doing so would mean more accurate results. The trick for vendors, therefore, is to make their search algorithms smarter, giving them the ability to anticipate what the searcher really wants.
Search engines anticipate when you may be looking for images, and display them along with standard results.
If you enter "daffodil" in the Google search box, for example, the results page will display a few photos of daffodils above the Web hits. "What users really want is a picture of those things," says Google software engineer Matt Cutts.
When we tried out the same query in Ask.com, Microsoft Live Search, and Yahoo, none of them displayed an image of the flower in their results. To be fair, we did a similar search using "Eiffel Tower"; this time, Google, Ask.com, and Yahoo displayed at least one photo of the Parisian landmark, but Live Search did not. (Live Search did provide us access to images of the Parisian landmark through a handy 'Eiffel Tower Photos' link in the Related Searches column, though.)
The integration of news stories with Web-search results is a work in progress, with Ask.com leading the way. For instance, when we searched for "titan" and "flash memory" (two separate queries), only Ask.com posted the latest headlines on related topics, including a link to a story on Saturn's moon Titan, and another to a piece on Intel's new flash memory storage--in both cases providing the information we were searching for.
When it comes to movie-related queries, both Google and Yahoo appear to have a slight edge. We searched for "300," looking for information about the Warner Brothers film, which was in theatrical release at the time of our testing. The very first match returned by both Google and Yahoo provided links to the movie's theatrical trailer and reviews, as well as a search box for entering a city or zip code to find local show times.
Microsoft Live Search, in contrast, was incorrect when it attempted to guess our location based on the IP address of the DSL account we used for testing (it was 74 miles off), and it listed movie theaters for Irvine, California, a good 90-minute car trip in Southern California traffic. Ask.com inquired if we meant "300, the movie"; we entered a zip code, clicked the supplied movie link, and received the expected trailer, reviews, and show-time links.
Chart: TV Guides of the Internet
The search services that specialize in finding Web videos weren't able to keep pace with Google's video index. Click the icon below to see how the tested engines performed, including their scores in each round.
| TV Guides of the Internet | |||||
|---|---|---|---|---|---|
| The search services that specialize in finding Web videos weren't able to keep pace with Google's video index. | |||||
| Service | First round | Second round | Third round | Average | Comments |
| Google Video | 21 | 21 | 18 | 20 | The most reliable video-search engine we tested. Its results page provides clip thumbnails and the source URL without looking cluttered. |
| AOL Video Search | 19 | 18 | 18 | 18 | Surprise! AOL's Truveo media-search engine is comprehensive and timely. The results provide fast access to top video categories too. |
| TubeSurf | 19 | 22 | 13 | 18 | Metasearch tool scans multiple video-search engines. Very accurate, but lacks thumbnail images and a preview window. |
| YouTube (Google) | 18 | 18 | 15 | 17 | If a video's online, you'll probably find it in this inexhaustible video catalog. YouTube's Google-like results page is easy to browse, too. |
| Blinkx | 16 | 17 | 18 | 17 | Solid search tool with a few annoying interface quirks. Example: The search-results page doesn't show the length of video clips. |
| Yahoo Video | 15 | 13 | 14 | 14 | Well-designed interface and a respectable clip-finder, although we expected better results from the search giant. |
| Pixsy | 14 | - | - | - | - |
| Search for Video | 14 | - | - | - | - |
| PureVideo | 13 | - | - | - | - |
| ClipRoller | 13 | - | - | - | - |
| AltaVista Video Search (Yahoo) | 7 | - | - | - | - |
| Brightcove | 3 | - | - | - | - |
| Chart note: The average score is composed of scores from three rounds of testing. Our scoring awarded a search engine three points if the first link in the results led to the target answer or site (or if the answer itself appeared at the top of the results page or within the first result). A link to the correct answer as the second or third result was worth two points, and a link to the correct response elsewhere in the top ten results scored one point. If the target answer or item didn't appear anywhere on the first page, we awarded no points to the search engine. | |||||
Smart Interface Tricks
The best search-result pages provide links to related items, even if the pages don't include your exact query (a great Beatles photo when you query "John Lennon", for example). It also helps if they know a little about you: If your IP address indicates that you live in the United States and you query "civil war", you're probably interested in the American Civil War and are looking for a war timeline, maps of the conflict, or maybe even biographies of Generals William T. Sherman and Robert E. Lee.
You'll find links to recent news stories related to your query at the top of your search results when you use Ask.com.
Ask.com does the best job of displaying related information and links. If you search for "WWII" (World War II), the service's Narrow Your Search column to the right of the results helps fine-tune your query with such links as 'History of WWII', 'Cause of WWII', and so on. Yahoo and Live Search provide similar tools, but they're neither as easy to use nor as comprehensive as Ask.com's.
Other sites are improving the user experience in different ways. For instance, Live Search has a "Smart Scroll" feature for image results. As you scroll down a collection of image thumbnails, the page reloads itself, adding thumbnails on the fly. Since there's no pagination, you don't have to click a link to load more images. In addition, Live Search's Scratchpad is a handy visual-bookmarking tool; you can drag images from the results page and drop them in the Scratchpad column on the right. When you're scrolling through dozens or even hundreds of images, the Scratchpad makes it easier to track shots you want for a report or presentation.
Not every interface trick works, even if it seems clever at first. Ask.com's Binoculars feature, for example, allows you to preview certain site results in a pop-up window by holding your pointer over results that have a binoculars icon associated with them. But for text-page previews, the large thumbnails don't provide enough detail to help you determine the page's content.
Chart: Find the Perfect Pics
Image-search specialists such as Picsearch couldn't best the accuracy and clear interface of Google's top-scoring image-search component. Click the icon below to see how the tested engines performed, including their scores in each round.
| Find the Perfect Pics | |||||
|---|---|---|---|---|---|
| Image-search specialists such as Picsearch couldn't best the accuracy and clear interface of Google's top-scoring image-search component. | |||||
| Service | First round | Second round | Third round | Average | Comments |
| Google Image Search | 26 | 24 | 20 | 23 | Great attention to detail. Below each thumbnail is the image size in pixels, the file size, and the domain where the image is located. |
| Ask Images | 23 | 20 | 19 | 21 | Returned very accurate image results, and it has a well-designed search results page. This is a worthy alternative to Google. |
| Lycos Image Search (Ask) | 23 | 19 | 19 | 20 | This service uses the Ask.com image index. It is about as accurate as Ask.com, but we prefer the latter's less-cluttered look. |
| Live Search Images | 19 | 18 | 19 | 19 | Clever interface, particularly the "Smart Scroll" feature that autoloads new images on the search-results page (no click to the next page). |
| Picsearch | 18 | 18 | 18 | 18 | Comprehensive and quick, Picsearch proves that a little engine can compete with the big players. |
| Ixquick Metasearch | 18 | 17 | 17 | 17 | A metasearch engine that's easy to use. Good performance, but a notch below the top performers. |
| AltaVista (Yahoo) | 18 | 13 | 13 | 15 | A decent showing for Yahoo's sister site. We liked the simple interface, but accuracy could be better. |
| Yahoo | 17 | 14 | 14 | 15 | Given Yahoo's impressive text-search scores, we expected a better showing here. |
| Corbis | 18 | 5 | - | - | - |
| Ditto.com (Picsearch) | 17 | - | - | - | - |
| Pixsy | 14 | - | - | - | - |
| Flickr | 11 | - | - | - | - |
| TimePix (Getty Images) | 7 | - | - | - | - |
| American Memory (historical) | 6 | - | - | - | - |
| Tiltomo (Flickr) | 5 | - | - | - | - |
| FreeFoto.com | 3 | - | - | - | - |
| Chart note: The average score is composed of scores from three rounds of testing. Our scoring awarded a search engine three points if the first link in the results led to the target answer or site (or if the answer itself appeared at the top of the results page or within the first result). A link to the correct answer as the second or third result was worth two points, and a link to the correct response elsewhere in the top ten results scored one point. (For image search engines, which typically include more than ten thumbnails per page, we awarded a point if the target photo appeared on the first page.) If the target answer or item didn't appear anywhere on the first page, we awarded no points to the search engine. | |||||
Human Insight
Search sites are adding more human opinions to search results. "If you're looking for a hotel for your vacation, traditionally you've gotten a list of hotels," says Tim Mayer, Yahoo Search's senior director of product management. "But there's no person or authoritative expert saying, 'I went to this hotel and recommend it.'"
One solution is to add more "human expertise" to the autogenerated results. Yahoo has a homegrown source of this content: The Yahoo Answers service is an online community where users ask and answer questions on a variety of topics. The company recently began adding Yahoo Answers posts to the bottom of some search-result pages. For instance, when you search for "summer vacations", you will see a 'Shared by Yahoos' section at the bottom of the results; the section includes links to several summer-vacation 'Best Answers' (meaning the most relevant), as chosen by Yahoo Answers users.
Cynics may point out that this feature benefits Yahoo by routing users to its in-house content, boosting its page views as a result. Perhaps, but we found the Yahoo Answers opinions more relevant than many of the standard, top-ten hits, a few of which were just random links to vacation rentals. In contrast, Yahoo Answers posts were more interesting: One contributor described seven travel destinations in India, while another offered a comprehensive list of family-friendly theme parks.
Opinionated results are a good start, but at present they aren't always blended smoothly with the conventional search results, usually being relegated to the bottom of the page. Few people are likely to scroll down to find them. Yahoo's Mayer acknowledges this issue: "We're just beginning to integrate this content, so we're going to be learning more about how to rank it. Long term, we will incorporate it more aggressively on the search page."
The same "summer vacations" query run on Ask.com, Google, and Live Search produced a conventional (though mostly useful) collection of travel-site hits. While Live Search did serve up a link to an MSNBC message board on seasonal travel, it was labeled as merely 'boards.live.com', which gave us no indication that it was related to our search specifically.
Yahoo's 'Also try' feature at the top of its results page presents related queries culled from users: Search for "muffins" on the site, and 'Also try' lists "muffins recipe", "banana muffins", and other alternate forms of relevant content.
Chart: Keep Up With the Times
No news-search service beats Google's timeliness, which is everything when it comes to hearing stories first. Click the icon below to see how the tested engines performed, including their scores in each round.
| Keep Up With the Times | |||||
|---|---|---|---|---|---|
| No news-search service beats Google's timeliness, which is everything when it comes to hearing stories first. | |||||
| Service | First round | Second round | Third round | Average | Comments |
| 19 | 20 | 19 | 19 | Comprehensive listings of local and national news sources; the service also provides image thumbnails for some of its news stories. | |
| AltaVista (Yahoo) | 20 | 17 | 16 | 18 | AltaVista returns accurate news results, but it provides access to fewer local news sources and image thumbnails than Google offers. |
| Yahoo | 20 | 11 | 15 | 15 | To find News Search on Yahoo, you must click the More button; we think the feature should be easier to find in this cluttered kitchen-sink layout. |
| Live Search | - | 14 | 16 | 15 | Impressive showing, but its search-results page is too sparse. Adding some image thumbnails of news events would help. |
| Ask.com | - | 14 | 11 | 13 | A decent news-search tool, but you'll get more timely hits elsewhere. And in our tests, Ask displayed fewer image thumbnails than Google does. |
| Daypop | 13 | 15 | 4 | 11 | - |
| NewsTrove | 16 | 5 | - | - | - |
| News-Is-Free | 11 | - | - | - | - |
| InfoPlease | 0 | - | - | - | - |
| Chart notes: The average score is composed of scores from three rounds of testing. Our scoring awarded a search engine three points if the first link in the results led to the target answer or site (or if the answer itself appeared at the top of the results page or within the first result). A link to the correct answer as the second or third result was worth two points, and a link to the correct response elsewhere in the top ten results scored one point. If the target answer or item didn't appear anywhere on the first page, we awarded no points to the search engine. We added Live Search and Ask.com to the second round of tests after several contenders in the first round failed. | |||||
Mobile Search
Using search on a mobile phone or other handheld device is a unique and sometimes frustrating experience. Odds are you're traveling and you want quick, local information, like directions to a nearby restaurant or gas station, the phone number of a long-lost friend, or tickets to a baseball game. But your device's 12-button keypad or tiny thumb keyboard makes entering the wrong street address or zip code far too easy. Depending on your ISP and your physical location, your mobile Internet connection may not be particularly fast or reliable. And obviously, the road is no place to be searching for videos.
All of the major search services help you retrieve local information on a mobile device. We found Microsoft Live Search for Mobile offered the most accurate local results, but only by a small margin over Google Maps, Yahoo Local, Ask City, and Whitepages.com. Live Search for Mobile worked particularly well at finding contact information for businesses and people. That said, you'll want to use at least two mobile search sites, as different engines excel in different areas. For instance, if you're searching for phone-book-style information, use Live Search for Mobile, Superpages.com, or WhitePages.com. To find an online ticket seller for baseball seats, go to Ask.com, Google, or Yahoo--but use their Web search rather than the local-search option.
What's the difference between the two? The Web search allows traditional queries. Using our baseball example, if you're in Los Angeles, you can enter "dodgers tickets" and find an online ticket site such as StubHub. In contrast, local search may help you find the name, address, and phone number of a nearby ticket dealer--the kind of information you'd get from a phone book--but what you're probably looking for at the moment is an online seller.
The mobile search engines we tested are optimized for handhelds. Some accomplish this more gracefully than others, however. On our first visit to Google Maps and Yahoo Local via our Nokia E62 phone, we had to click the Mobile link on their home pages to view the modified display; on follow-up visits, both sites automatically switched to the mobile interface. By comparison, Ask City, Superpages.com, and WhitePages.com immediately recognized that we were connecting through a handheld and offered their mobile-friendly interfaces automatically. Using Live Search for Mobile was somewhat confusing: The main site, www.live.com, isn't optimized for mobile devices. Rather, we had to manually browse to m.live.com to access the mobile-optimized service.
Even though the search engines are optimized for handheld browsers, most of the sites they'll point you to aren't. StubHub, for instance, is built for conventional browsing only, which makes it painful to navigate on the Nokia's miniature screen. Indeed, when you're viewing only a small portion of the page at a time, it's easier to get lost. Ask City is one of several sites that attempt to reformat a page to fit the small screen, but the results are often just as difficult to use.
Chart: Dial Into the Local Scene
Newcomer Microsoft Live Search for Mobile edges out Google Maps atop our list of services that retrieve local info; local specialists trailed despite their many noteworthy features. Click the icon below to see how the tested engines performed, including their scores in each round.
| Dial Into the Local Scene | |||||
|---|---|---|---|---|---|
| Newcomer Microsoft Live Search for Mobile edges out Google Maps atop our list of services that retrieve local info; local specialists trailed despite their many noteworthy features. | |||||
| Service | First round | Second round | Third round | Average | Comments |
| Live Search for Mobile | - | - | - | 19 | Very good at finding names and directions. But the main Live.com URL doesn't accommodate mobile devices; you have to navigate to m.live.com. |
| Google Maps | 18 | 18 | 18 | 18 | A tad better than most. The service is great for maps, directions, and restaurant searches but poor for locating long-lost friends. |
| Yahoo Local | 15 | 15 | 17 | 16 | Yahoo's local search scored well overall, though it doesn't list the distance between a business and your location (no "1.2 miles," for example). |
| WhitePages.com | 15 | 15 | 15 | 15 | The service did well in our restaurant and person searches; but it's not a good choice for general queries, such as finding an online ticket broker. |
| Ask City | 14 | 14 | 14 | 14 | To get driving directions, Ask City required that we enter a full street address (like 123 Main St.) rather than just a city name. |
| Superpages.com | 14 | 14 | 14 | 14 | Online phone book is great for local names and businesses, plus maps and directions. |
| Chart notes: The average score is composed of scores from three rounds of testing. Our scoring awarded a search engine three points if the first link in the results led to the target answer or site (or if the answer itself appeared at the top of the results page or within the first result). A link to the correct answer as the second or third result was worth two points, and a link to the correct response elsewhere in the top ten results scored one point. If the target answer or item didn't appear anywhere on the first page, we awarded no points to the search engine. A new version of Live Search for Mobile launched at the end of our test period, so we tested it only once. | |||||
Google's Secret Sauce
So how does Google maintain its edge over rivals? A comprehensive index is key, as is a fanatical devotion to excellence. "We live and breathe search. There are guys who'll come in over the weekend. They'll say, 'I was doing a query on this strange thing and I wanted Google to return this page.' And then people will spend a few hours figuring out how to tune our algorithm to return more relevant results," says Google's Matt Cutts.
Fair enough, but other sites certainly have their own search fanatics. So what else? Timeliness is crucial. Cutts says, "If we don't have an important document or breaking news within a day or two after it happens, we need to do better. It used to be that search engines would update once a month, and we've led the charge in trying to return and refresh little bits of our index every day." Within a week or two, Google has refreshed its Web index.
Many of the smaller search engines license their indices from larger players and then massage the results with their own home-brewed algorithms, or via human indexing, a technique pioneered by Ask.com but now used by several search services. The Yahoo-owned AlltheWeb and AltaVista, for instance, each use their parent company's index, while Lycos hooks into Ask.com, and AOL Search is really Google in disguise.
Why use a smaller site if it's simply mirroring another engine's results? For one thing, you might like its interface better. Google's iconic, simpler-is-better approach is the industry standard-bearer: just a search box, a fanciful graphical treatment of the brand, plenty of white space, and no ads on the home page. Both Ask.com and Live Search mimic Google's minimalist motif, although the former spices things up a bit by adding a handy Search Tools column on the right side of the screen, giving you easy access to useful search helpers that you might otherwise miss, such as the service's encyclopedia, dictionary, and maps.
AlltheWeb's home page is unadorned as well (some might call it homely), consisting mostly of a query box and a few tabs for accessing specialty-search sections (news, pictures, and so on). On the other hand, its parent site is the Times Square of Search: Yahoo's home page is so jam-packed with news headlines, ads, and links to other features that the search box at the top of the screen is easy to overlook entirely. Meanwhile, if you enter a query in AOL's search field, you'll get much the same results as you would by doing the same search at Google.
Specialty engines vary in their presentation. This is illustrated by our two favorite blog-search sites, Blogdigger and Bloglines. Blogdigger adheres to Google's less-is-more rule, while Bloglines uses its home page to pitch its RSS newsreader and blog-publishing tools.
The Future of Search
Search sites continually test new looks, some of which may go mainstream at some point. For instance, Ask.com's Act X interface divides the results into three regions: The middle displays standard text and links, the left has links for expanding/narrowing a search, and the right serves up related dictionary and encyclopedia entries, as well as shopping links. Google lets the public test its experimental interfaces at www.searchmash.com.
One company that is taking a different approach to search is Snap, which lists search results in the left half of the browser and a preview of the first result in the right half. Each result has a Preview button you can click to see the page. Or click the Next Result button to scroll through previews in sequence.
In all these experimental interfaces, the search box is still the sole means of data entry. Don't expect this to change anytime soon, although work continues on voice-driven search. Google recently took the wraps off its experimental Voice Local Search service: Dial 800/GOOG-411 (800/466-4411) to try it out yourself. (The company warns that the service is currently available only in the United States, for businesses in this country.)
New Services Put Search in Your Hands
From personalized search engines to deep-Web infiltrators to throwback human-powered Web indexers, these are the new search resources that we've come to rely on.
Rollyo: The big search engines are great for everyday queries, but what if the information you're looking for is a bit esoteric? Rollyo is the place to go. Here you can create your own "Searchroll," a customized search engine that provides results from up to 25 sites of your choosing.
Getting started is easy: Enter a Searchroll name and some sites (Rollyo says 5 to 15 work best), and then pick a category (such as 'News & Media'). Add a few tags or keywords if you wish to help others find your creation. The real fun is in exploring other people's Searchrolls. The site has lists of popular and recently added Searchrolls--who knew that Diane Von Furstenberg had one?--and you can search by keyword or URL for topics that interest you. So if Google isn't delivering on your "Antarctic beach volleyball" query, Rollyo is the place to go.
Congoo: Ever hear of the "Dark Web"? It sounds ominous, but don't be afraid. The term refers to the huge chunk of Web content that search engines can't index, including paid-subscription sites (such as that of the Wall Street Journal) and password-protected corporate and government databases. In addition, some pages can't be indexed because there are no links to their content, or because their very long and complex URLs are too difficult for search engines to crawl.
Certainly, much of this content is hidden for a reason. A corporation, for instance, doesn't want the product development and marketing plans on its intranet made public. But you do have a way to search many paid-content sites. Even better, you can access that content--some of it, anyway--without paying a penny. Congoo NetPass, a free utility that's available at their website, allows you to search and read a limited number of stories from 35 popular paid-content sites, including the online homes of Billboard, Encyclopaedia Britannica, New Republic, and the WSJ. You can make between 4 and 15 visits per month, depending on the site. When you exceed the monthly allotment of visits to a particular site, you're prompted to subscribe to the service, since you're apparently finding it useful.
Download the free Congoo NetPass toolbar, which attaches to your Internet Explorer or Firefox browser. Use it as you would any conventional toolbar: Simply type a query in the search box and press <Enter>. The service searches Yahoo for the standard Web results, but it also uses its own index to search for any related paid content, such as a full company profile from the MorningStar financial service. If you already use another search engine's toolbar, you might find that installing Congoo's shrinks your browser's viewable area, but just right-click your toolbar area and uncheck the options you're not using to enlarge your browser window.
PreFound: Another mashup of social networking and search, PreFound lets you express your expertise in a subject by creating your own Groups, which are composed of links to the best Web resources, whether news, images, video, or other information. The service offers to share its revenue from the AdBrite online-ad network with "Finders" who devise more than 150 Groups. You could also become one of the site's Featured Finders, who include university professors, outdoors enthusiasts, and even professional baseball players. Along with the PreFound search results for your query, you'll find the standard Google listings.
How We Tested the Search Engines
Our goal was to replicate the real-world use of search engines. Most people aren't search gurus or reference librarians, and they typically don't use or understand Boolean commands (such as AND, OR, NOT); the average Web user simply goes to a search site, enters a word or phrase, and presses <Enter>. True, there are many ways to fine-tune a query--see our list of our favorite search tricks--but most folks usually can't be bothered. That's why we didn't place quotation marks (e.g., "hiking boots") around our test queries, even though doing so would almost certainly have produced more-accurate hits.
We tested a total of 55 services in six categories: general (text info), video, mobile/local, news, images, and blogs. We conducted our tests over three weeks. We used ten terms in each category. For general text-information searches, for instance, our categories included technology, pop culture, research/academic topics, books, and travel/shopping. In each category we posed what we considered an easy query and a difficult one: Just about every search engine could find the 2007 Academy Award winners, for example, but only a few could locate a portrait of Italian poet-philosopher Giambattista Vico. We tested the engines at different times of the day--morning, afternoon, and night--on all seven days of the week. (However, the time of day and the day of the week had no impact on the results.) We occasionally changed our queries, particularly in the news and blog categories, to keep them timely. For instance, our easy pop-culture query changed from "Academy Award winners" to "2007 NCAA basketball tournament teams" later in testing.
We performed all of our testing on the same Windows system, a 2-GHz Celeron desktop with 512MB of memory and a 1.5-megabits-per-second (download speed) AT&T/Yahoo DSL Internet connection. We didn't test for query speed, but this typically varied by only fractions of a second. Among the most-popular search sites, we saw little difference in between the time we pressed <Enter> and the appearance of the search results. However, we did detect significant differences among the blog-search engines, with some smaller sites--such as Best of the Blogs--taking several seconds to post the results. For the local-search tests, we used a Nokia E62 smart phone.
Chart: Our Test Queries and Target Responses
We asked ten queries in each of six categories: text information, video, images, local information, news, and blogs. Within each category we sought both an easy and a difficult target in five different topics or subjects. In some instances our target was a specific Web page; but unless noted in the chart, any site that provided the target response was considered to have given a correct answer. For topical searches we changed the queries in each of the three rounds of testing. Click the icon below to see our complete list of test queries and the results we wanted for each.
| Search-Engine Shoot-Out: Our Test Queries and Target Responses | ||
|---|---|---|
| We asked ten queries in each of six categories: text information, video, images, local information, news, and blogs. Within each category we sought both an easy and a difficult target in five different topics or subjects. In some instances our target was a specific Web page; but unless noted below, any site that provided the target response was considered to have given a correct answer. For topical searches we changed the queries in each of the three rounds of testing. | ||
| Text-information search | Query | Target response |
| Technical (easy) | WiMax speed and range | 40 mbps per channel in a typical cell radius of 3km to 10km, as stated on the WiMax Forum |
| Technical (difficult) | Linksys WPA2 | Yes, the Linksys WRT54G wireless router does support WPA2 |
| Pop culture (easy) | First round: 2007 Academy Award nominees Second round: 2007 Academy Award winners Third round: 2007 NCAA basketball tournament teams |
First round: The award nominees in each category Second round: The award winners in each category Third round: The NCAA basketball tournament field of 64 teams |
| Pop culture (difficult) | Tigers may 1 | The Detroit Tigers play the Baltimore Orioles at Comerica Park on May 1, 2007 |
| Research/academic (easy) | Capital of Moldova | Chisinau (Kishinev) |
| Research/academic (difficult) | weight gnu yak | Male yaks weigh 1100 to 2645 pounds (500 to 1200 kilograms), and females weigh 660 to 770 pounds (300 to 350 kilograms); a gnu (a blue one) weighs 500 to 600 pounds (235 to 275 kilograms) |
| Book excerpt (easy) | excerpt Doctorow The March | An excerpt from the E.L. Doctorow book |
| Book excerpt (difficult) | No Exit | An excerpt from the Jean-Paul Sartre play |
| Travel (easy) | 2007 All Star Game | San Francisco's AT&T Park on July 7, 2007 |
| Travel (difficult) | Sanibel Beach | The location of Bowman (or Bowman's) Beach on Sanibel Island in Florida |
| Video search | Query | Target response |
| Internet video (easy) | Colbert speech 2006 White House Correspondents Dinner | The full 24-minute, 10-second video of Stephen Colbert's roast of the president |
| Internet video (difficult) | Food for Thought DeLuca | Food for Thought; short |
| Television clip (easy) | Samurai Delicatessen | Saturday Night Live skit starring John Belushi and Buck Henry |
| Television clip (difficult) | Gilda Radner | Saturday Night Live skit starring Gilda Radner and Steve Martin dancing |
| How-to video (easy) | how to carve a turkey | Several potential correct answers, one of which is Martha Stewart's instructions |
| How-to video (difficult) | upgrade notebook drive | Any video showing step-by-step instructions for installing a new hard drive in a notebook PC |
| News (easy) | First round: tornado in Enterprise, Alabama Second round: Australian cyclone Third round: Valerie Plame congressional hearing |
First round: Tornado in Enterprise, Alabama Second round: Australian cyclone Third round: Valerie Plame at congressional hearing |
| News (difficult) | First round: Dodgers home opener Second round: Chatsworth Third round: big rig 101 |
First round: "LA Dodgers Sell Out Home-Opener Tickets" Second round: "HazMat Team Called to Chatsworth" Third round: "Big Rig Crashes Through Center Divider on 101" |
| Animation (easy) | Gopher Broke | Nominee for 2005 Academy Award for animated short subject |
| Animation (difficult) | Maestro | Hungarian artist Geza M. Toth's Academy Award-nominated short |
| Image search | Query | Target response |
| Portrait (easy) | Rutherford B. Hayes | The nineteenth president of the United States |
| Portrait (difficult) | Vico | Italian poet-philosopher Giambattista Vico, 1668-1744 |
| Building (easy) | Pantheon | The Pantheon in Rome |
| Building (difficult) | Lincoln Hall Kentucky | Lincoln Hall, Berea College, Kentucky |
| Thing (easy) | platypus | The marsupial |
| Thing (difficult) | windform | A 20-foot horn |
| Place (easy) | Times Square | The New York City landmark |
| Place (difficult) | Asilomar | Asilomar State Beach, Pacific Grove, California |
| Event (easy) | Mardi Gras | The New Orleans Mardi Gras celebration, 2007 |
| Event (difficult) | Keeneland | Keeneland Racetrack (the building, not the horses or the track itself) |
| Local search | Query | Target response |
| Restaurant (easy) | Spago | The address and telephone number of the popular restaurant in Beverly Hills, California |
| Restaurant (difficult) | Chili My Soul | Information for the obscure restaurant in Encino, California |
| Gas station (easy) | gas station 91301 | The nearest station to the test location of Agoura Hills, California |
| Gas station (difficult) | cheap gas 91301 | The lowest-priced gas near the test location of Agoura Hills, California |
| Person's telephone number (easy) | [Author's own number] | The author's telephone number |
| Person's telephone number (difficult) | telephone Bill Smith Thousand Oaks | The telephone number of Bill Smith of Thousand Oaks, California |
| Driving directions (easy) | Agoura Hills to Spago in Beverly Hills | Directions for a short trip using mostly freeways from Agoura Hills to Spago in Beverly Hills |
| Driving directions (difficult) | Agoura Hills to Zuma Beach in Malibu | Directions for a longer trip using mostly surface streets and/or exurban roads from Agoura Hills to Zuma Beach in Malibu |
| Event/show/news (easy) | L.A. Lakers tickets | Any reputable ticket vendor |
| Event/show/news (difficult) | Calabasas election | March city election results for Calabasas, California |
| News search | Query | Target response |
| Technical (easy) | First round: RINBOT Second round: EU iTunes Third round: Microsoft antivirus |
First round: RINBOT virus outbreak Second round: European Union asks Apple to change iTunes copy protection Third round: Microsoft reports that its OneCare antivirus software needs work |
| Technical (difficult) | First round: 8800 Second round: Woosh Third round: March Vista update |
First round: nVidia driver update for Vista Second round: Woosh Wireless Broadband now compatible with Vista Third round: March 2007 Windows Vista Application Compatibility Update |
| International (easy) | First round: Airbus strike Second round: Morocco Third round: Russian crash |
First round: Workers strike over Airbus cuts Second round: Morocco cafe bombing on March 11, 2007 Third round: Plane crash in Samara, Russia |
| International (difficult) | First round: Zimbabwe Second round: Nigeria Third round: Ugandan troops |
First round: Australia cricket team defeats Zimbabwe in World Cup warm-up match Second round: Presidential candidate Umaru Yar'Adua returns to Lagos Third round: Ugandan ex-troops win $2 billion payout |
| Business (easy) | First round: Volt Second round: gas prices Third round: China interest rates |
First round: GM announces 2010 launch date for its Volt electric car Second round: Average gas prices rise 20 cents per gallon Third round: China central bank announces on March 17, 2007 (the day of testing) that it is raising interest rates to curb investment |
| Business (difficult) | First round: 7921G Second round: Alfa Third round: patriot blend troops |
First round: Cisco introduces the Unified Wireless IP Phone 7921G Second round: Russia's Alfa Group conglomerate raises its voting stake in Pelcom Third round: Newhall Coffee Roasting donates 12,000 pounds of its Patriot Blend to troops in Iraq and Afganistan |
| Popular/breaking (easy) | First round: Libby Second round: California fires Third round: war protest |
First round: Scooter Libby found guilty in CIA leak case Second round: Brushfires in Southern California Third round: Antiwar march in Washington, D.C. on March 17, 2007 (the day of testing) |
| Popular/breaking (difficult) | First round: Agoura water polo Second round: Oaks Christian Third round: Los Altos closure |
First round: Agoura High School's girls water polo team beats Laguna Beach in CIP Division II championship Second round: The Westlake Village, California, Oaks Christian High School's girls soccer team wins CIF title Third round: School board in Camarillo, California, votes to close Los Altos Middle School |
| Weather (easy) | LA weather | Weather forecast for Los Angeles |
| Weather (difficult) | Leadville weather | Weather forecast for Leadville, Colorado |
| Blog search | Query | Target response |
| Technical information (easy) | First round: RINBOT Second round: EU iTunes Third round: Microsoft antivirus |
First round: Blog discussing the RINBOT virus Second round: Blog discussing the European Union's beef with iTunes Third round: Blog discussing Microsoft's comment that its OneCare antivirus software needs work |
| Technical information (difficult) | First round: 8800 Second round: Woosh Third round: March Vista update |
First round: Blog discussing the nVidia driver upgrade for Vista (see news for details) Second round: Blog discussing Woosh's new Vista compatibility Third round: Blog discussing March 2007 Windows Vista Application Compatibility Update |
| Breaking news (easy) | First round: Libby Second round: Fred Thompson Third round: war protest |
First round: Blog discussing the Scooter Libby verdict Second round: Blog discussing Fred Thompson's possible 2008 presidential run Third round: Blog discussing the antiwar march in Washington, D.C., on March 17, 2007 (the day of testing) |
| Breaking news (difficult) | First round: Ballymena Second round: Nigeria Third round: Derbyshire cricket |
First round: Blog discussing the Irish Cup match between Linfield and Ballymena United Second round: Blog discussing the breaking news story, "Nigeria's Yar'Adua Blames Opposition for Death Rumour" Third round: BBC's Derbyshire Cricket Blog |
| Popular commentary (easy) | First round: Huffington Libby Second round: Huffington Hagel Third round: Malkin protest |
First round: Arianna Huffington opining on the Scooter Libby verdict Second round: Blog on Senator Chuck Hagel's presidential bid Third round: Michelle Malkin's blog of March 17, 2007, on that day's Washington D.C. antiwar protest and counterdemonstrators |
| Popular commentary (difficult) | Santa Paula | City of Santa Paula, California, blog written by its city manager |
| Product review (easy) | First and second rounds: iTunes Third round: Google Talk gadget |
First and second rounds: The iPod Blog Third round: Preview of Web-based module for Google Talk users |
| Product review (difficult) | First and second rounds: D40x Third round: EaglePicher |
First and second round: Preview of Nikon D40x digital camera (or similar) Third round: Preview of supersmall battery |
| Company profile (easy) | First and second rounds: Google Third round: Vonage and Covad |
First and second round: Inside Google Third round: Profile contrasting two popular VoIP providers |
| Company profile (difficult) | First and second rounds: Wetpaint Third round: Tello |
First and second round: Post on the wiki service Third round: Post on the demise of the pioneer VoIP service |
Top Search Tricks
You say you're no search wiz? Follow these simple steps to find the best answers without wading through pages of useless links.
1. Put quotation marks around your query: For example, type "air conditioning" rather than air conditioning. "This creates that query as one unit, so all the results you get back will have that set of words," says Microsoft Live Search senior program manager Jeff Osmer. Without the quotes, you'll end up with a lot of unrelated pages on air quality, hair conditioner, and the like.
2. Use the minus (-) sign: This tells the engine, "I don't want this word." If you enter Shakespeare -William, for instance, you're instructing it to display anything that matches "Shakespeare" and doesn't include "William."
3. Get definitions: Use the modifier 'define' as a handy way to make a search engine into a dictionary. Just type define word in the query window for a quick definition. (Google users must add a colon, as in define: word.)
4. Ask in plain English: Today's search engines are constantly being fine-tuned to respond to real-world queries. You'd be amazed at what they can do with even the most cryptic query. "[For] anything you throw into the search box, we'll try to find an answer," says Google software engineer Matt Cutts. For instance, if you enter "yellow polka dot bikini" 45 rpm excellent condition, that lets the engine know you're looking for a copy of Brian Hyland's 1960 novelty record without scratches.
And here's a trick that works with all four major search engines: Plug in a telephone area code as your query. Suppose that you see a number on your phone's caller ID and wonder where the caller is located. Enter the number's area code in the search window to view the city or cities for that number. Caveat: If the caller is using a VoIP line, he or she may not necessarily be located in the area code's geographical region.
5. See quick weather previews: If you're flying to, say, San Francisco, enter San Francisco weather in the search box for a fast forecast.
Skip the Advanced Classes

Illustration by Steven Lyons





