Google Designs an Engine for Eggheads
After raking in billions of dollars helping the masses search the Web, Google Inc. is targeting academics like Daniel Branton.
The Harvard University biology professor has used Google for years to locate sensors and other laboratory equipment. But it wasn’t very good at finding peer-reviewed research papers and other academic publications.
That changed with the recent release of Google Scholar. The specialized search engine uncovers papers better than Google.com, Branton said, and is simpler to use than the expensive databases in Harvard’s vaunted libraries.
“I found that I could find, in a much more rapid fashion, what I was looking for by just putting in a few keywords,” he said. “Even if it had a bit of advertising, I wouldn’t mind.”
That’s exactly what Google was hoping the professor would say. Google Scholar doesn’t feature ads, but one day it just might.
Along with rivals Yahoo Inc. and Amazon.com Inc., Google is in hot pursuit of the egghead crowd as a new source of traffic -- and revenue. They all see opportunity in scholarly information, a $12-billion business long dominated by publishers such as Reed Elsevier and Thomson Corp.
The Web companies are aiming to snag scholars before they head to subscription services to do research. To do that, the companies have to equip their search engines to scour some of the billions of pages locked away in subscription sites, hidden corners of the Web and books.
“The search engine market is very competitive,” said Ed Pentz, executive director of CrossRef, a nonprofit organization in Lynnfield, Mass., that works to make scholarly papers available to the public. “This is the next stage, and it’s prestigious, high-quality content. It expands the range of what they can offer users.”
It also gives search engine companies more pages on which to place ads. Google generates nearly $3 billion a year serving up targeted ads tied to keywords, and the Mountain View, Calif., company says it could sell even more if it could just find relevant pages to put them on.
If it does, advertisers will probably pay extra to deliver their highly targeted messages directly to potential customers such as Branton, said Chuck Richard, a vice president with market research firm Outsell Inc.
When the Harvard professor uses Google Scholar to find papers on his specialty, nanotube biology, companies that sell the $6,000 humidity sensors he favors might be willing to pay a pretty penny to have their ads appear on his computer screen.
“They’re not trying to get the average person,” Richard said. “One of the strengths is it calls the right people to it.”
Google Scholar relies on the same index of websites as Google’s mainstream search engine but ranks those websites differently.
The regular search engine uses a system called PageRank, which turns the Internet into a huge popularity contest. When a user enters a query into Google, Web pages are ranked in large part according to the number of times they are linked to by other websites.
That approach presents problems when it comes to specialized research, because the most relevant papers may not be the most popular. Studies that would be useful to scientists may live in obscurity on the Web, buried beneath hundreds of sites that command more links.
So Google Scholar doesn’t rely on how many websites link to a particular paper or book. Instead, it examines who wrote it, who published it and how many other scholarly works cited it. Based on those sorts of criteria, dissertations, peer-reviewed papers and other scholarly literature can float to the top of search results.
The newest Google search engine has received mixed reviews. Branton and others complain that, unlike nearly all other scholarly databases, Google won’t disclose what sources it draws from, so academics can’t know what they’re missing.
It also appears to be incomplete. In an online review, Peter Jacso of the University of Hawaii at Manoa’s library and information science program wrote that a search for all records from the medical research database PubMed returned only 879,000. The database actually contains 15 million records.
“Like its popular counterpart, searching Google Scholar is easy,” he wrote. “Finding the gems is difficult.”
Anurag Acharya, a former computer science professor at UC Santa Barbara who created Google Scholar, said the 1-month-old program would improve as Google got feedback from users.
Google is working to collect more scholarly information by expanding its Google Print program to digitize some or all of the books in five research libraries, including those at Harvard. The task will take years. But when it’s finished, Google says, millions of books will be available for academics -- and others -- to search.
That mirrors efforts at other big Internet companies.
Amazon has scanned more than 200,000 books in the last year as part of its Search Inside the Book program. Last month the Seattle-based online retailer capitalized on that investment by quietly introducing Amazon.com Citations to help drive sales of scholarly texts.
Now, when Amazon provides descriptive information about a book, it includes a list of other books that are mentioned in the text, footnotes or bibliography. If the book is cited in other texts, those are mentioned as well.
“It’s yet another way for people to find related books,” said Bill Carr, Amazon’s director of digital media. Carr declined to say whether sales have grown as a result.
As for Yahoo, it’s enlarging its searchable index to include more Web pages from sources such as the National Science Foundation and the Smithsonian Institution. Among other items, the Sunnyvale, Calif., company’s 9-month-old Content Acquisition Program has netted nearly 100,000 digital documents about ancient Babylon from UCLA’s Cuneiform Digital Library Initiative.
“There’s no question that we are getting much wider publicity as a result of it,” said Bob Englund, a professor of Assyriology and director of the cuneiform library.
In addition, Yahoo has recalibrated its system for gauging the importance of a website to raise the visibility of academic papers and documents. Because they tend to deal with esoteric subjects, few other sites link to them -- a situation that would normally relegate an online document to obscurity. Instead, Yahoo relies on publishers like Englund to identify the keywords and other descriptors that will boost their obscure material in search results.
Google, analysts say, has been the most aggressive in its efforts to tap into the academic market. “It’s very shrewd,” said John Tinker, an analyst with ThinkEquity Partners. “It keeps differentiating them from Microsoft, and it keeps them intellectually ahead of Yahoo.”
Both Google and Yahoo were founded by computer science graduate students from Stanford University. While Yahoo has evolved into an online media company -- its chief executive, Terry Semel, ran the Warner Bros. studio before going to Silicon Valley -- Google has stuck close to its roots. Its fevered courtship of engineers, scientists and mathematicians has made for a workforce attuned to the needs of the academic market.
“This is one way we give back to the community,” said Acharya, the professor-turned-Google engineer.
The scholarly search engine could pave the way for similar services targeting other research-heavy professions, analysts say. Logical candidates include Google Doctor and Google Lawyer.
“I see it as an opening that can be further developed,” said David Garrity, an analyst with investment firm Caris & Co. “Google can certainly hold itself out as being the site where people can go to gain not only breadth but depth of information.”
