Skip to Content
Uncategorized

A New Perspective on Search Results

Google is experimenting with different ways to serve up search results. But will any of them stick?
February 6, 2008

Amid the flurry of news over Microsoft’s bid for Yahoo and Google’s rebuttal, a research announcement by Google went largely unnoticed. Last week, the search giant began a public experiment in which users can make their search results look a little different from the rest of the world’s. Those who sign up are able to switch between different views, so instead of simply getting a list of links (and sometimes pictures and YouTube videos, a relatively recent addition to the Google results), they can choose to see their results mapped, put on a timeline, or narrowed down by informational filters. Dan Crow, product manager at Google, says that the results of the experiment could eventually help the company improve everyone’s search experience.

Results may vary: Google has announced an experiment in which users can sign up to receive alternate views of their search results. Instead of simply seeing the results in a standard list view, participants can view the results on a map, on a timeline (above), or by using filters that narrow down the results based on information such as dates, measurements, locations, and images.

Google’s experiment highlights the slow but steady push of engineers and designers to improve the Web search experience for the masses. While search algorithms are constantly improving, the interface has remained static for more than a decade: people submit keyword queries, and the engine spits back a list of 10 hyperlinked results. “If you compare Google search-result listings today to the Infoseek results in 1997, they’re almost indistinguishable [in terms of presentation], except for the ads,” says Marti Hearst, a professor in the School of Information at the University of California, Berkeley.

Hearst says that there continue to be attempts from non-Google engines to offer alternatives to the standard search interface. Ask.com, for instance, lets a user see a thumbnail view of each Web page before she clicks through to the link. And Clusty.com extracts words that are found on the search-results pages, letting a user drill down to a more specific search. For instance, a search for “MIT” can be specified to include references to “laboratory,” “Massachusetts Institute of Technology,” “project,” and other words or word combinations.

But these slight alterations in search have been slow to catch on, as is evident from Google’s dominance in the field and its relatively conservative approach to its user interface. Hearst thinks that many people tend to use Google and other simple interfaces for a couple of reasons. One is that search engines must accommodate a wide range of users, from the novice to the savvy. Less experienced users tend to get distracted when more information is presented on the screen, she says: people don’t respond well to being overloaded with information, especially when they want a simple answer to a query. But perhaps more important, she says, is the fact that people are familiar with a decade-old interface and, as creatures of habit, they are reluctant to try something new.

Google’s Crow says that people are generally happy with the interface as it exists today. “The basic format hasn’t changed much because it’s been successful … It works well for most of the users most of the time,” he says. “But that doesn’t mean we couldn’t do something beyond search today.”

When a person signs up for the “alternate views” experiment on the Google Labs page, he essentially adds three search filters to his results page: “Info view,” “Timeline view,” and “Map view.” (See below image). By default, a search for “Grateful Dead” serves up results in the “List view,” which is essentially the standard results page. If a user selects Map view, he could see a map indicating where the group originated (San Francisco), and where it performed its last show (New York City). Clicking on Timeline view provides a bar graph of dates associated with the group–important concerts, for instance–over the years. And Info view lets the user filter the search by dates, measurements (in this case, Google offers units of years and, oddly, tons), locations, and images.

Crow says that when a person signs up for the experiment, Google collects the same information about his searches that the company would otherwise. This includes noting the search terms and result links that are selected, as well as logging the amount of time a user stays on the page of the selected link. Crow notes that all of the information collected is stripped of any identifying information. This data, in addition to market research collected from participants who visit Google’s offices and participants who allow Google to come into their homes to track their search habits, will be used to determine the most helpful features of the experiment, and how best to sprinkle those features into search results without upsetting users, he says.

Improving the search interface isn’t easy, but it’s a crucial part of the technology, says Oren Etzioni, a professor of computer science and engineering at the University of Washington. There are billions of Web pages, and the results page only reveals 10 pages at a time. “Search is the process of drinking from the fire hose,” Etzioni says. “This means that getting the user interface right … is incredibly important.” He doesn’t see anything revolutionary about Google’s experimental views in particular, but “throwing things out there and letting people react is very smart.” He believes that in the next couple of years, search will evolve to provide more interface options for people, and not everyone will be using the same interface.

Search will change, but it will be a gradual process, since there’s a fine line between providing helpful information and overwhelming the user with text and links. “One thing to remember is that it’s still the early days,” Google’s Crow says. “People think that search is a solved problem. I think we’re still in the early days of making search work on a universal global scale. We know we can do better.”

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.