After the dust settled last week, everyone was asking the same question: how did the media get it so wrong? Why did it portray the election of Hillary Clinton as inevitable? Two primary critiques emerged. The first was that there was a media bias—that the industry was “so in bed” with Hillary that they wanted her to win and didn’t properly report on or vet both sides. The second was around the flaws of social media—that due to the nature of newsfeed algorithms, we all found ourselves in echo chambers of the same ideas from like-minded people, from the same old places.

The way I see it, there isn't enough content being created from diverse communities, and it's too hard for consumers to discover interesting content from multiple viewpoints. Media has to shout so loud to get through the noise, it resorts to radical headlines and outlandish commentators just to get our attention.

What if we were informed by the honest, unfiltered sentiment of people around the country instead? What if news feeds showed multiple perspectives around a topic instead of an echo chamber reinforcing our current beliefs? The good news is that our ability to do this exists now. If the media industry had embraced new media models that bring those perspectives to the world, everyone might have had a much more realistic understanding of how the election would unfold.

This is one of the reasons I created Odyssey, our social media platform of more than 15,000 content creators in 1,000+ communities and over 30 million readers. We analyzed nearly 11,000 pieces of election-related content posted on Odyssey over the past year. The results show, among a number of interesting data points, that while feelings about the two candidates were relatively balanced, a distinct anti-Hillary sentiment existed as well.

People engaged more with anti-Hillary content

While there was a higher percentage of positive Clinton articles posted on Odyssey, anti-Clinton content generated substantially more social engagements per read (while ‘engagement’ encompasses ‘likes,’ comments or shares per pageview, likes, which more obviously indicate sentiment, are the primary driver of the metric).

Trump was a more interesting subject

Of all election-related content generated since November 2015, 80 percent was related to Trump. Still, contrary to what the media might have thought, the posts were fairly evenly split between pro and anti-Trump sentiment.

Trump content inspired more response

Thirty-five percent of reads of a Trump-focused article resulted in the reader taking some social action, such as liking, sharing or commenting. Twenty-eight percent of Clinton-focused articles inspired that kind of response.

While “polls and predictive models failed to predict Trump’s strength” (Politico), an examination of the Odyssey community could have given everyone a more accurate sense of how the vote was going to turn out. Why?

How the democratization of content fixes media

The community model fosters trust: People are starving to understand what’s going on in their community. They also want to hear more global viewpoints through the lens of their community peers. When people understand that those perspectives represent nothing more than authentic ideas that the community deems important, they trust it.

Diverse viewpoints breed understanding: When you enable people everywhere to create content, the perspectives you get represent a wide range of opinions, not just what a few minds decide is newsworthy. There are a lot of people who have a lot to say out there. When that becomes thousands of millennial content creators representing more than 1,000+ local communities across the nation, you start to see just how wide that range--and how diverse that perspective--can be. Singular perspectives get woven into a rich fabric of ideas.

Personalized discovery helps people form a new worldview: Once that new fabric of ideas is surfaced, people need to make it their own. With noise-free exposure to these ideas, each person can then internalize what resonates the most with them and approach every issue with their own, new, broader perspective.

The combination of these three concepts solves media’s creation and distribution problem, breaking the echo chamber and bringing more empathy, value, and understanding.

It should be apparent by now that media’s role in influencing elections needs to change drastically. It bombed in this election because it didn’t understand or disseminate what people across the nation really thought. How do we ensure that media and social media are much more constructive next time around? At some level, there’s a fairly simple bottom line here: want to understand what the people really think? Go to the people! Hear their voices not only across the U.S., but across the entire globe.

Study & Methodology

*Study based on 10,990 total pieces of content. All articles mention Trump/Clinton in the headline and/or brief. Articles categorized as Trump/Clinton defined as content with mention of either candidate's name in the headline and/or brief of the piece. Engagements defined as likes, comments or shares to per page view per article. Pro/Anti categorization dependent on positive/negative descriptors commonly used for each candidate.