Always looking for ways to improve its search engine, Google Search’s latest move is offering more
about the sites you are visiting. And it’s all made possible using the Knowledge Graph.
The latest tweak plays off Google Search’s major algorithm upgrade in September 2013. Code-named Hummingbird, it was the biggest morph since the Caffeine update in 2010. Hummingbird expanded the features of the Knowledge Graph so it can answer more questions -- even questions that don't have simple answers.
“As you choose the right search result for you -- be that about the American Civil War or back pain -- you want to know where the results come from,” said Bart Niechwiej, a software engineer at Google. “To help you learn more about the Web sites you see in your search results, starting today you may see more information about them directly on the results page when you search on your desktop.”
Choosing the Right Results
Here’s how it works: When you get your search results you can click on the name next to a link for more information. So, using the American Civil War example, you will see the source right under the URL. When you click on the name of the source, it takes you directly to its Web site.
“You’ll see this extra information when a site is widely recognized as notable online, when there is enough information to show or when the content may be handy for you,” Niechwiej said.
“The information you’ll see is based on the Knowledge Graph, Google’s interconnected understanding of the things that exist in the world. As we expand the Knowledge Graph, we expect to give you more information about more Web sites -- making it easier for you to choose the right result,” he added.
A Reliability Score
We caught up with Rob Enderle, principal analyst at The Enderle Group, to get his take on the new Google Search feature. He told is it’s a good idea that makes it much easier for consumers to click through and ensure they are not being fooled. He pointed to one valuable example: rumors about presidents and urban legends, which run rampant online.
“People are easily fooled and they still may not take the trouble to click on the link,” Enderle said. “If we are expecting people to take that extra step and click through, we’re probably expecting too much.”
Enderle proposed what he thinks is a better approach: a score that tells the reader the probability that what he is reading is accurate, such as a Yelp-like quality rating.
“The rating would need to show the number of people that voted and the quality of the score. It would need to be an algorithm because it could be gamed,” Enderle said. “A reliability score that speaks to the validity of the site and the article would be valuable. So this is an improvement and not a big one.”