Can Users Trust Google’s Knowledge Graph Results?

Most of the time, Google’s Knowledge Graph is pretty good at what it does. It uses structured data to conveniently get searchers to the results for what they were actually referring to when they...
Can Users Trust Google’s Knowledge Graph Results?
Written by Chris Crum
  • Most of the time, Google’s Knowledge Graph is pretty good at what it does. It uses structured data to conveniently get searchers to the results for what they were actually referring to when they entered their query. This is particularly helpful in cases when names apply to multiple people or things.

    Sometimes, Google still has some trouble.

    Have you ever seen Google display erroneous information in its Knowledge Graph results? Have you ever felt like these results weren’t showing the best possible answer? Let us know in the comments.

    The latest example that has been spotted comes when the user searches for “brandy”. Brandy could mean the alcoholic beverage or it could mean the pop singer/actress (Brandy Norwood). If you go by Google’s Knowledge Graph results, they’re pretty much the same:

    Brandy results

    Brandy Results

    Brandy Results

    It’s also worth noting that after a day of this particular result being pointed out in multiple articles, it hasn’t changed.

    To be fair, Google’s Knowledge Graph results are getting the text right, but clearly there are issues with image association that can occur. Perhaps drawing images from Wikipedia when drawing text from it is the better way to go. Wikipedia gets it right:

    Brandy on Wikipedia

    Okay, so what? Errors happen, and this is a harmless one. But it’s not the first time one has been spotted. A while back, we reported on another one where it got a football player’s marital status wrong. While errors may very well be few and far between, how can users know for sure whether or not they can trust the information Google is providing as “knowledge”. Typically, users aren’t going to question the information they see here unless it’s obviously wrong.

    We’ve seen other cases where Google displayed images for Knowledge Graph results which probably weren’t optimal. For example, Google shows an image of Laura Prepon as the main image for the show Orange is the New Black. Sure, she’s on the show, but she won’t be for long, and she certainly isn’t representative of the entire show. She’s not even the lead actress. Why not show a poster or a picture of the cast?

    Laura Prepon

    There have also been cases where Google has shown nudity in these results, and this is obviously not something that Google wanted to do intentionally.

    Google knowledge graph nudity

    The point is that Google does not always display the best possible results for these highly prominent “Knowledge Panels”.

    This could become problematic when your consider that Google is continually expanding the type of information that appears in the Knowledge Graph, including much more important types of information (like medical and nutrition).

    Knowledge Graph nutrition

    Knowledge Graph medical

    It’s one thing to show questionable results to third-party sites. That’s just Google trying to deliver on information provided by others, but Knowledge Graph is supposed to give you accurate information so that you don’t have to hunt it down on other sites. It needs to be accurate. These results, if users are satisfied that they got the answer that they were looking for, are going to prevent a lot of users from going to other sites, which may or may not have better information.

    There’s also the fact that Google just changed its algorithm to apparently treat the rest of the web more like it does Knowledge Graph, at least in some ways.

    Mike Blumenthal, by the way, recently wrote a post called “Hummingbird, Local Knowledge Graph & Shitty Search Results,” suggesting that when Hummingbird launched it immediately led to a decline in quality of local results.

    “For example if you search on the phrase ‘Buffalo NY Diamonds’ it surfaces a second listing for a local jeweler at the same address that was created long ago for the purpose of keyword spamming ‘marketing’ in local,” he writes. “The problem of Google showing a single branded results was first spotted years ago. It subsequently lead to a spate of one box spam and then, for the most part, squelched by Google. For whatever reason, these spammy local knowledge graph entities seem to have made a come back.”

    “The timing and nature of the results makes me believe that we are seeing ‘the Hummingbird effect’,” he adds.

    Obviously this can be harmful for businesses who rely on local search visibility.

    Meanwhile, Google is apparently dumbing down some of its organic results, indicating that high quality, accurate content isn’t always enough.

    Regarding the Knowledge Graph, Google told us when it first launched that they realize it will never be perfect. Hopefully users realize that too.

    Do you think the Knowledge Graph has improved the Google experience? Do you think it has significant room for improvement, or is it doing a good job? Share your thoughts in the comments.

    Note: This article has been updated from its original form.

    Image: Google

    Get the WebProNews newsletter delivered to your inbox

    Get the free daily newsletter read by decision makers

    Advertise with Us

    Ready to get started?

    Get our media kit