What Google's new Knowledge Graph says about journalists
Search Engine Land | The Huffington Post | Techmeme
Google's new "Knowledge Graph," introduced Wednesday, could have some unintended consequences for journalists and news organizations.
I searched about two dozen news organizations, journalists' names and program titles to see what information would appear in these new "knowledge panels" that display alongside search results.
- Generally, the knowledge panels displayed people more frequently than publications, which could reinforce the value of individual journalists' brands over institutional ones. Nicholas Kristof has an extensive entry, while his employer The New York Times (like similar publications) has minimal results drawn from its Google+ presence.
- Knowledge panels are self-contained. You can keep reading and clicking inside them and forget the links to the search results that will take you to more information about your original search term. Users may appreciate this and Google will enjoy more time on its site, but news organizations may see less traffic as questions are increasingly answered inside search rather than via search.
- Portions of the knowledge panels rely on Wikipedia, which can be confusing or inaccurate (Michael Huffington is still listed as Arianna Huffington's spouse). The excerpts are also by definition incomplete, which can create a false sense of neutrality. For example, results for Rush Limbaugh and Rachel Maddow (see below) describe them with the same broad term, "political commentator," entirely missing the dramatic differences in their political positions.
Sometimes, the most basic information was the most interesting: the journalists' full names, height, partners, parents and education.
As Bianca Bosker wrote, "The summaries will also add some surprises into users' search results, surfacing information users didn't know they wanted but may be interested in. ... The change underscores the steps tech companies are taking to engineer serendipity and introduce us to ideas we might not have discovered otherwise."
Here's a sampling of the Knowledge Graph results I saw Thursday morning: