Google is never short on ideas how to improve its search system and break ground in new areas of user satisfaction. Sometimes, these ideas are large and revolutionary, like its embedment of maps into local searches. Its Knowledge Graph, a system of direct information provision (forgoing the need to peruse traditional search entries) is one of the most robust and useful additions in recent years, and it keeps evolving in new, unique ways.
Rather than solely providing word definitions, or numerical unit conversions, or even filmography information, the Knowledge Graph can provide unique insights on news and upcoming events. Take its entry on the Super Bowl, for example (keep in mind this was written just before the actual Super Bowl):
Presumably, this entry will self-update as teams score throughout the evening, and in the next week, will instead offer retrospective analysis of what is currently a forthcoming event. As a user, this doesn’t leave much to be desired; I can even scroll down to find additional results.
But a recent feature of the Google Knowledge Graph has made a much bigger impact, and reveals one of the biggest current flaws of the direct-information model. It’s all about Google’s portrayal of candidates in what has undoubtedly been one of the most eventful, peculiar election seasons of the past few decades.
Google’s politics-centric new feature, candidate cards, has begun the same way all its features begin: as an experiment. Accordingly, let’s refrain from judging this system too harshly.
The idea was to give the American public more transparency on their leading presidential candidates—which sounds great in theory. Google’s idea was to give each significant candidate a designated position in their search results for certain queries. These “candidate” cards would appear in a carousel to potential voters, giving them a direct line of insight into the candidates’ actions and positions. This feature was rolled out as a trial for the recent “undercard” Republican debate, along with YouTube integration and live tracking via Google Trends.
(Image Source: Google blog)
Here’s the issue: if you followed along with this experiment during the actual debate, you wouldn’t see multiple candidates’ positions. You only would have seen one, at least for the bulk of the time and for most queries.
As SearchEngineLand’s Danny Sullivan noted in a blog post on the issue, the carousel of cards that appeared, for practically any search, only showed posts and positions by one candidate: Carly Fiorina.
A handful of general searches like “gop debate” or even just “debate” returned the same carousel. Likewise for any undercard candidate-specific searches, such as “Huckabee” or “Santorum.” At first glance, you would assume that this is some type of error with Google’s system, that somehow these posts were “stuck” as the top results for any query that tapped into the feature. Could this mean that Google was unfairly favoring one candidate over the others?
Google would later confirm that nothing was wrong with the feature. Each candidate had the same ability to upload information to this system; Fiorina was the only candidate who made use of the system, and therefore had substantial ground to gain.
Candidate cards for the main GOP candidates appeared not long after the undercard debate ended, including Donald Trump, who was absent from the “main” debate. Take a quick look at these and take note of anything peculiar that stands out:
(Image Source: SearchEngineLand)
Look at the center post, which features a link to donate $5 to Mark Rubio’s campaign, and consider the nature of the query: 2016 Republican debate. If you’re like me, this raises some questions about the card system and whether it goes “too far” for search results.
I don’t care who you support, which party you belong to, or what you think about this election. For the purpose of this article, I’m assuming every candidate on both ends of the political spectrum is equally unqualified to lead the country, and so should you. Remove your biases and consider the following dilemmas this system presents, for any candidate in the running:
What does the candidate card system say about Google? What does it mean for the political system? Is it a useful tool that needs refinement or a total travesty that should be scrapped? I’m not quite sure myself, but you can be sure this experiment didn’t quite go the way Google originally intended. Keep your eyes peeled for how this feature develops—it could have a massive impact on how this and even future elections pan out. In the meantime, you better hope your favorite candidate is as skilled at SEO as you are.