In an effort to accelerate the rendering of Web pages in users' computers, Google has developed a search feature that predictively preloads pages before users click on result links.
The feature, called Instant Pages, will be rolled out as a beta test for Chrome browser users in the coming week, said Amit Singhal, a Google fellow, in the company's Inside Search event on Tuesday, which was held in San Francisco and webcast.
Along with the existing Google Instant feature, which lets Google predict search results before users complete a query term, Instant Pages shaves off between four and 10 seconds from the search process, he said.
"The time it saves us is amazing," Singhal said.
Even increasing the speed of the search process by milliseconds results in more frequent search usage and higher user satisfaction across the board, he said.
On average, a Web page takes five seconds to load after a user clicks on its Google search result, but Instant Pages cuts that down often to fractions of a second, officials said.
Instant Pages preloads in Google back-end systems the results Google determines a user is most likely to click on. Webmasters tracking usage of their websites will see the difference between a page that Google preloaded from their site and wasn't clicked on, and pages that are actually visited by users, officials said.
In its initial implementation, Instant Pages will preload the page that ranks first on the search results list. Instant Pages will become available with the next beta release of Chrome, which includes a pre-rendering capability.
Instant Pages will also be offered for mobile devices in the coming weeks.
Google also announced that voice search and search by image, two features available for mobile devices, will now be available for desktop browsers as well.
Voice search, in which users speak their query terms, is being rolled out to Chrome desktop browser users, who will see a microphone icon in Google's search box.
The ability to search using an image, instead of a text query, was first introduced to mobile devices in 2009 and called Google Goggles. For desktop browsers, it's being called Search by Image and will be identified by a camera icon next to Google's search box. When the icon is clicked, users will be able to enter a photo URL into the search box, or upload an image into it. The Search by Image feature is designed to work for images of well-known locales, and not for identifying individuals' houses or faces, officials said.
In addition, Google Instant is also being extended to Image Search.
"Google is taking a big step toward expanding the repertoire of search by bringing voice and photos and images into the same interaction paradigm as text," said IDC analyst Hadley Reynolds via e-mail.
Google is also leveraging its search traffic, which is much larger than its competitors', to improve the machine-learning process of its engine in the areas of voice search and search by image, Reynolds said.
Google also announced improvements to its mobile search interface for general Web search and local and maps search.
Specifically, Google's mobile home page now will feature icons of local businesses that are commonly searched for, like restaurants, coffee shops and bars, to serve as search shortcuts.
Also, the map image with markers for local results now remains fixed at the top of the mobile results page, and it gets refreshed as users scroll down the list of results.
These announcements highlight Google's growing focus on mobile search, Reynolds said. "All the features -- the camera linkage, the use of voice recognition and the Instant Page capability -- play directly to the limited screen real estate and hamstrung keyboard of the smart devices; and they work fine on the PC as well," Reynolds said.
Mobile search traffic on Google has increased by a factor of five in the past two years, so it continues to rise in importance for Google and its users, Singhal said. In the past year, the use of the voice search feature in Google's mobile search has increased sixfold, officials said.
In particular, mobile search has emerged as the preferred search alternative at the times of day or seasons of the year when people are usually away from their desktop PCs, such as lunchtime during weekdays, as well as summer and winter vacations, officials said.
Sign up for Computerworld eNewsletters.