Media

Editor vs. Algorithm: Why the Future of Media Needs Both

By Amanda Walgrove June 23rd, 2015

Last week, Fortune journalist Matthew Ingram asked an interesting question: “Would you rather have Apple’s human editors filtering your news or Facebook’s algorithms?”

With Apple hiring human editors for its News app and Facebook’s Instant Articles running in the almost entirely automated News Feed, this debate of editor versus algorithm has perhaps never been more pertinent.

But here’s a solution that may be simpler than it seems: We need both.

While it’s fun to pit minds and machines against each other, the future of media isn’t going to be dominated by companies that choose one over the other—it’s going to be driven by those that combine the efforts of editors and algorithms to anticipate reader demand and deliver to a unique audience.

But why? There’s no shortage of arguments for either side, so let’s lay out some generally agreed-upon facts.

Algorithms and data sets can provide editorial operations information they otherwise would never have, such as pinpointing which topics are bringing in traffic, when readers are tuning in, and what devices people are reading on.

Meanwhile, humans can anticipate big stories and empathize with the reader. After all, the human editor is a reader, one who responds emotionally to each story. Computers may digest information quicker, but they’re just scanning, not feeling. It’s also important not to forget that algorithms are only as efficient as the humans that create them—and are therefore prone to just as many biases and mistakes.

For these reasons, companies are mostly lauded for bringing in humans to curate their news platforms, like when Snapchat hired a CNN politics editor to serve as its head of news.

Still, we can’t throw the algorithm out with the bathwater.

Choosing humans alone to curate the news would be a big step backwards for journalism. As Matthew Yglesias wrote for Vox, “If it had actually been possible back in the day to algorithmically determine what choice of cover story was most likely to get subscribers to engage with newsstand readers to buy it, isn’t that what editors would have picked?”

Most likely, they would have. We can’t shun the progress we’ve made with data; we just have to learn how to use it best, and when humans need to intervene.

Upworthy’s recent traffic plummet, for example, teaches us a lesson on what happens when a publisher leans too hard on algorithmic magic. Editors gamed Facebook’s algorithm by packaging preexisting YouTube videos with clickbait headlines and sharing them on the social site. Facebook’s editors caught on and adjusted its algorithm to penalize sites that use clickbait, causing a dip in Upworthy’s traffic.

As entrepreneur Jason Calacanis put it, in order to succeed, we need “John Henry and the steam hammer versus the steam hammer alone.” For Upworthy, it’d grown to rely too much on a steam hammer that, in the end, it didn’t even own.

Perhaps the best example of a site that combines human editors and algorithms is Gabe Rivera’s tech news aggregator, Techmeme. Its content was curated by algorithms for three years until Rivera hired veteran tech reporter Megan McCarthy in 2008. The algorithm continued to run, with headlines moving up on the page if a blogger linked to it. McCarthy just trimmed the fat for readers.

Rivera wrote in his announcement: “Though the implicit edits conveyed via algorithm outnumber the explicit edits perhaps by 1000 to 1 or more, the impact of the human editor is nonetheless pronounced.” With an eye on Techmeme’s site, McCarthy could eliminate obsolete stories more quickly and push up potentially breaking stories before the computer even caught them.

More recently, LinkedIn started to use a multi-tiered editing system for its Pulse app. As Wired reports, human editors curate the big headlines each day and select specific stories for users based on their industries. Meanwhile, an algorithm takes into account users’ connections and company affiliations to deliver relevant news.

LinkedIn even goes so far as to tell readers who or what is curating their story. Above each piece of content, users might see “Editor’s pick,” “Trending in Internet industry,” or “Popular among employees of LinkedIn.”

As for Apple, it clearly doesn’t want to just rely on algorithms for its News app. The Apple News Editor job posting states: “[The candidate] will have great instincts for breaking news, but be equally able to recognize original, compelling stories unlikely to be identified by algorithms.” Still, Apple’s News app announcement made it clear that algorithmic software has a major part to play in its upcoming app: “The more you interact with News, the more it learns about you. As a result, the app will be able to create a personalized feed of curated articles based on your interests.”

This killer combo of human editors and data-driven news feeds is the natural evolution of media creation and curation. When you consider the fact that many of these human editors are increasingly millennials who have come to treat computers as extensions of their own brains, this isn’t too surprising. Leading tech companies that understand this are already building hybrid newsrooms to blow past competitors and appeal to new readership.

So hire your John Henrys, build your steam hammer, and start making noise.

Image by Deb Wenof
Tags: , , , , ,