Search Engine Users
Part 2. What people seek with search engines
Searchers turn to search engines for both important information and trivia.
What kinds of searches are people doing? Are they turning to search engines for need-to-know information, or for trivia, or both? Most searchers, 55%, say they are as likely to look for information they consider important as for trivial information. Some 28% of searchers are more serious; they say they look mostly for information that is important to them. And 17% are more whimsical in their searching, using search engines mostly for information they don’t consider important.
About one in four internet users have searched for his own name on a search engine, just to see what comes up.
More specifically, some 44% of users say that most or all the information they search for online is critical. It is information they absolutely need to find, for instance to accomplish an important task or answer an urgent question. About one quarter of those searchers say everything they search for is that critical. On the other hand, about one quarter of all searchers, 24%, say that very little of the information they search for is that important to them.
A number of users describe their recent searches for important information12:
- “I was asked about state labor laws — timecards and the legal ramifications of paying only some of the time listed.”
- “Medical information for my son…it gave us some background information and a common ground to start asking questions about his condition.”
Conversely, 33% of searchers say they would not bother looking up most or even all of the information they search for if they lacked access to internet search engines. On the other extreme, about one quarter of searchers, 24%, say that very little of their searching is in pursuit of such trivia.
Another user describes a recent search he considers expendable:
- “I just completed a course on the Operas of Giacommo Puccini. With the internet I was able to easily obtain his biography and a complete listing of all his works, various reviews of the pieces and as much detail as I wanted. All of this without leaving my desk and to obtain this when it was convenient to me…. (without a search engine) I probably would have not bothered to obtain that information.”
Popular search terms show what is on Americans’ collective mind.
The big search engines slice and dice their search logs and compile various lists of the most popular search terms. These lists give some sense of what is on America’s collective mind, reflecting our shared culture, news, trends, events, and phenomena. Popular search terms always include a lot of seasonal references, current news, and pop culture. The search engines regularly omit terms relating to adult content from their lists.
According to AskJeeves13, the top searches for the week of Oct. 8, 2004 were online dictionary, music lyrics, games, halloween costumes, jokes, baby names, quotes, Britney spears, Paris Hilton, poems.
Or from Yahoo14: Eminem, Britney Spears, Usher, Mt. St. Helens, Nelly, Register to Vote, Halloween costumes, Jo Jo, Paris Hilton, Green Day, NASCAR, Christina Aguilera, Hilary Duff, NFL, Linkin Park, Slipknot, Drudge Report, Alicia Keys, John Kerry, My Boo
Or from the Lycos 5015, from Oct. 2, 2004: Airline Flight Tracking, Clay Aiken, Paris, Hilton, Pamela Anderson, Halloween costumes, Britney Spears, Michelle Vieth, Halloween, NFL, poker, KaZaA, Brooke Burke, beheadings in Iraq, Christmas, http://50.lycos.com/greatesthits.asp, Lindsay Lohan, Star Wars 3, The Olsen Twins, WWE. Dragonball, Baseball
All Time Hits: Searches for Britney Spears and Pamela Anderson have been on the Lycos top 50 list for 277 weeks in a row.
And other search terms are the flashes in the pan, following events or news. In the week following the first of the presidential debates, John Kerry landed at Lycos’s #36 spot, and George Bush made #40. They had not been on the list the previous week.
Top searches of the year include some timeless classics and pop culture with real staying power. These have been in the top 100 from week to week during all of 2002: the Bible, Diablo II, Neopets, Tupac Shakur, The Beatles, Sailor Moon, The Simpsons, Carmen Electra, Oprah Winfrey, World War I, New York City, Final Fantasy, World War II.
Sometimes, search terms reflect extraordinary times when the whole world seems focused on a single thing. Lycos 50 published an unedited and an edited version of top hits to their web site between noon and midnight on Sept. 11, 2001: The unedited version: CNN, News, World Trade Center, CNN News, CNN.com, MSNBC, ABC News, BBC, Breaking News, World News.
“Iraq” was Google’s #7 search of the year for 2003, after Britney Spears, Harry Potter, Matrix, Shakira, David Beckham, 50 cent.
And the edited version, which does not include news organizations: World Trade Center, Whitney Houston, Pentagon, World Trade Centre, Bin Laden
World Trade Center New York, Osama Bin Laden, American Airlines, Camp David, Nostradamus.
And Google’s Zeitgeist archive gives a sense of the organic nature of search, showing what terms are gaining or losing popularity from week to week, or the popularity of search terms as they correlate with current news events. For example, when the space shuttle Columbia was lost over Texas in February of 2003, searches for “NASA” spiked for about a week.
Beyond the popular search terms are the unique queries.
Unique queries lie outside our collective culture, in our personal interests and problems, our individual work, our eccentric curiosities, and perhaps our miscellaneous misspellings and oddities.
We get a glimpse of these unique terms, as they appear in the raw feed of search terms that users key in. These snapshots of the linguistic soup that a search engine processes are a reality check of the true hodgepodge that arrives at engines, without the benefit of neat packaging into organized lists. A sample from the MetaCrawler series16 looks like this:
A broader user base and increased web content have altered the landscape of what we search for.
In the earliest days of the internet and search engines, when the population of users was dominated by young men, two of the most popular search topics were sex and technology. Now, with the huge demographic expansion of the internet user population, their more diverse interests, and the vast growth of online content, the search landscape has changed. A recent study examining search trends finds the proportion of searches for especially sex and pornography has declined dramatically since 1997, while searches for the tamer topics of commerce and information have grown.17
The study took search terms from logs of three popular search engines, AltaVista, AlltheWeb.com, and Excite, spanning 1997 – 2002, and organized them into categories. The following represents the categories from AltaVista in 2002:
- People, places or things
- Commerce, travel, employment, or economy
- Computers or internet or technology items
- Health or sciences
- Education or humanities
- Entertainment or recreation
- Sex or pornography
- Society, culture, ethnicity or religion
- Government (or military)
- Performing or fine arts
Categories generated from the search terms from all three engines look very similar. The search logs, from the search engines Alta Vista in 2002, Excite in 2001, and AlltheWeb is 2002, pointed to considerable common ground among searchers.18 The first three categories included the greatest number of queries. The most popular category, “People, places and things,” captured 49% of search queries for Alta Vista, 41.5% for AllTheWeb, and 20% for Excite. Topic two, “Commerce, travel, employment and economics”, captured 12.5% of search queries for Alta Vista, 12.7% for AllTheWeb, and 24.7% for Excite. Sex and porn captured 3% of search queries for Alta Vista, 8.5% of Excite and 4.5% of AllTheWeb. Computers and technology got 12% for AltaVista, 16% for AlltheWeb, and 9.6% for Excite. All remaining topics captured single digits of the search market.
One category of particular interest is commerce. There is no exact way to measure the number of commercial searches, but at least one statistic gives us an idea of the scope of commercial searching: comScore Networks reports that 40% – 45% of search queries now include sponsored results.19 While it may not be the case that all these search queries were launched with commerce in mind, the search terms used were close enough to warrant a commercial opportunity in the returns. Further, Google’s vice president of engineering, Urs Holzle, in commenting on the large number of user complaints about increasing appearance of commercial search results, says that is the direction the Web is moving. “Even three years ago,” he said, “the Web had much more of a grass roots feeling to it.”20
When Americans don’t use search engines and what they don’t search for.
People do not always turn to search engines to find answers or information they are seeking online. Sometimes they take different paths: going straight to favorite, familiar specialized portals; using URLs someone has recommended, or following links they come across.
For example, when asking those who look for health information online about the last search did for health information, most searchers, 81%, reported that they started at a search engine. But the remaining 19% said they went right to favorite health sites, or took recommendations from friends or family, or followed links on other familiar websites.
Another example comes from the 29% of internet users who have looked for religious or spiritual information online. When users were asked about their most recent search for religious purposes, 51% said they began at a search engines, the rest at a familiar religious Web site or religious portal.21
In a similar measure, some 66% of internet users have gone online to use a government website. When asked in a 2003 survey how they got to their most recent government website visit, some 37% got there thru search engines, 19% heard about the website from friends, relatives , or advertisements, 17% went there on a repeat visit, 14% found them through government publications, and 8% through general government websites, like firstgov.gov.22
And in the study looking at changing trends in what people search for on search engines, Spink and Jansen suggest that one reason the number of searches for sex-related topics have declined over the years may be that users are bypassing the search engines and going directly to familiar Web sites.23
Sometimes, users will not search online because they think the information they are looking for won’t be on the web or that search engines won’t be able to locate it. One user writes,
- “Most of the time I am looking for journal articles for schoolwork, and (search engines) don’t provide any information I need.”
In the rapidly changing world of search engines, Google’s new scholar search may change his mind. Although another points to a further problem:
- “(I don’t search for) scientific articles, because I would have to pay for a subscription to the journals in order to download the information. So I use the library resources or go to library.”
- “I don’t use the web when my questions are very local in nature. It is quicker to call a shop in the area of my home or office. I have used the internet to get phone numbers of local shops in order to call them. For example, when my lawn mower recently broke, I needed to find a local shop to repair it. I searched the online smart pages to find all shops in my area, then I called them to discuss the problem. Local, small businesses do not have a sophisticated web presence.”
Again, this objection may soon be in the past, as engines are beginning to add local searching capabilities.
“I can’t think of anything I don’t search for,” one search enthusiast reported to us.
Finally, there are billions of web pages that users of the popular search engines have no access to. In fact, despite what feels like a veritable and ever-growing glut of information, only a small fraction of existing web pages are actually indexed by the popular engines. The rest are hidden on the “invisible web” for a number of reasons, among them: they are too much trouble technically to locate, too expensive to keep up to date in indexing, or are large numbers of web page “spam,” the bogus or deceptive pages out there that engines try to avoid indexing.24
- Comments from users come from an online survey conducted by the Pew Internet & American Life Project in May – June, 2003. ↩
- http://www.ask.com/ ↩
- http://buzz.yahoo.com/weekly/ ↩
- http://50.lycos.com/ ↩
- http://www.metaspy.com/info.metac.spy/metaspy/unfiltered.htm ↩
- Spink, Amanda and Bernard J. Jansen. Web Search: Public Searching of the Web. Springer Publishers, 2004. ↩
- Categorizing search terms, like categorizing anything from languages to plantlife, is an exercise that is partly art and partly science. The resulting taxonomy of search terms, although subjective, can give us some kind of map of the world of searching. Other directories or taxonomies of search terms could dissect search terms differently, and give us different perspectives on what we search for. ↩
- comScore qSearch: Q2 2004. ↩
- http://insight.zdnet.co.uk/hardware/servers/0,39020445,39175560,00.htm ↩
- http://www.pewinternet.org/PPF/r/45/press_release.asp ↩
- http://www.pewinternet.org/Reports/2001/CyberFaith-How-Americans-Pursue-Religion-Online.aspx ↩
- http://www.pewinternet.org/Reports/2004/How-Americans-Get-in-Touch-With-Government.aspx ↩
- Spink, Amanda and Bernard J. Jansen. Web Search: Public Searching of the Web. Springer Publishers, 2004. ↩