Has The Knowledge Graph Made Google Better At Search?
Sometimes Google’s Knowledge Graph is inaccurate. Google knows it’s not perfect, and provides a “feedback/more info” link at the bottom of Knowledge Graph results, which you can click to let Google know what it has wrong, if you believe there is an error. Of course, there’s no guarantee that Google will acknowledge your feedback or that you’ll hear back from them with any quickness.
Have you noticed errors in Google’s Knowledge Graph? Let us know in the comments.
Quality of the “knowledge” within the Knowledge Graph has been a question from the beginning. We asked Google after it launched how susceptible it would be to Wikipedia vandalism, given that Wikipedia is a major source of the info provided to the Knowledge Graph (though it’s only one of a variety of sources). At the time, a Google spokesperson told us, “I can’t share a ton of detail here, but we’ve got quality controls in place to try to mitigate this kind of issue. We’ve also included a link so users can tell us when we may have an inaccuracy in our information.”
“Our goal is to be useful; we realize we’ll never be perfect, just as a person’s or library’s knowledge is never complete,” he said. “But we will strive to be accurate. More broadly, this is why we engineer 500+ updates to our algorithms every year — we’re constantly working to improve search, and to make things easier for our users.”
Google is showing Knowledge Graph results for a whole lot of searches these days, and continues to expand the product all the time. Earlier this month, Google announced its expansion into more languages. You have to wonder how many errors are out there. It’s one thing for Google to try to deliver the best results in terms of other sites, but this is the information that is supposed to be “knowledge”.
Barry Schwartz at Search Engine Roundtable points to a Google forum thread where somebody claiming to be the agent for NFL player Nick Eason found an error with the player’s Knowledge Graph entry, and asked to have it corrected:
I am writing to get some information removed from one of my Client’s biogrpahy’s that is sponsored by Google. When you search the name, NICK EASON in Google, a little biogprahy box comes up under his picture and under the Wikipedia blurb.
Nick is no longer married and desperately wants his Spouse tab removed. Can you please let me know the most efficeient and effective way to get this done? Thanks!
By the time this article was first written, it had remained unchanged, but Google has since fixed it.
Wikipedia appears to be the main source of Google’s “knowledge” about Eason. That’s the only source Google has labeled, anyway. When we checked, Eason’s Wikipedia page had no mention of his marriage or Regina Eason, who Google’s result said was his spouse (before finally updating it).
As Schwartz writes, “With web search, the results can be within minutes updated – extremely fresh. But the knowledge graph, if it needs an update, it can take months and months.”
A lot can change in that amount of time, obviously. Google, per its mission statement, wants to organize the world’s information and make it universally accessible, but the ways in which it has been able to address the timeliness and/or freshness of data has left a great deal to be desired in recent years.
Google has placed such a great deal of emphasis on providing fresh search results, but then you have an example like this, where the result Google is actually emphasizing is simply outdated. Results, in general, have also become less fresh due to the lack of Google’s realtime search feature, which used to give you near-realtime content from Twitter as it appeared in the Twitterverse. This gave you fresh up to the second (or close to it at least) content about newsworthy topics right in your search results. Now, if you want that, you better just go to Twitter.
Meanwhile, Google is cramming organic search results with “fresh” content, often leaving more relevant content buried under the more recently written content. This was illustrated perfectly on the day of the Sandy Hook shooting, when reports came out that a person named Ryan Lanza was responsible for the horrible shooting in Newtown, Connecticut (granted, the killer was later reported to be Adam Lanza). If you tried searching Google for “Ryan Lanza,” you were likely getting little but few-minute-old reports, when you may have really been looking to find more about who this guy really was (a situation made even more confusing when media outlets started putting out photos and pointing to social media profiles for the wrong guy).
That has little to do with the Knowledge Graph, but it has everything to do with the relevance and usefulness of Google’s search results (which is being diminished in other areas of search as well).
One thing that’s mildly disturbing about this whole Knowledge Graph thing, in terms of the potential for errors, is that Google recently started including information about medications in the Knowledge Graph. This was an area of major concern as so-called “content farms” filled up Google’s search results with hardly credible articles about health issues. Google has made strides in that area, particularly with updates like Panda.
“This data comes from the U.S. FDA, the National Library of Medicine, and the Department of Veterans Affairs, among others,” said Google Search Senior Product Manager Aaron Brown of the recently added Knowledge Graph data. We hope you find this useful, but remember that these results do not act as medical advice.”
That last part is important.
Have Google’s search results gotten better since the launch of the Knowledge Graph? Have the results in general gotten better over the past year? Let us know what you think.