Needle in a Haystack: Navigating Social Media in 2015

by William Fisher

Curation, curation, curation: those were three of the big themes for social media in 2015. Facebook, Instagram, Snapchat, Twitter and YouTube all launched major initiatives to help users find what they’re looking for – and, of course, to induce advertisers to open their wallets. But disappointments and mishaps (including a particularly embarrassing moment at Google Photos) have shown the limits of algorithmic curation, and are leading all these companies to re-evaluate the need for human editors.

Here’s a quick look at each – and thoughts on what lies ahead in 2016:


Launched in September, FB’s Signal allows journalists and marketers to source and embed content from both FB and Instagram. But Signal is a professional tool rather than an actively curated service. It was FB Newswire that emerged as the most interesting and successful source for curated user-generated content, both for consumers and media professionals. Newswire aggregates newsworthy content shared by individuals and organizations on FB into a single feed – with a hat tip to FB’s human editorial partners at News Corporation’s Storyful.


In June, Instagram released a “reimagined” Explore page by adding trending Tags and Places, as well as curated collections. These features have greatly enhanced navigation for app users. But if you’re really serious about content discovery on Instagram, request access to FB Signal. On the other hand, businesses have given the new Explore the thumbs up as a useful marketing tool.


If curation is about finding stuff, Snapchat’s Live Stories and Story Explorer (which debuted in November) are more about allowing users to get lost (read: “engaged”) in a more immersive navigational experience. An algorithm aggregates interesting content in the form of “Live Stories,” while “Story Explorer” lets you swipe on Live Stories content to view related shots. But without human editors to vet the results, is the content safe for brands? Partnerships with media companies like Buzzfeed, CNN and Vice seem more commercially promising than Snapchat’s 2015 user-generated content initiatives.


Twitter’s Curator was previously the go-to tool for professional journalists seeking to monitor and display trending social media content. This year, Twitter made Curator available free to all publishers. However, it still reserves advanced integrations, customizations and support services for its Certified Partners. For consumers, the launch of Twitter’s Project Lightning was bigger news this year: that lightning bolt icon at the bottom center of the app allows you to tap into breaking news and trending entertainment. Twitter hopes to grow the service by building out human newsrooms in major markets around the world.


In June YouTube announced its own partnership with News Corp’s Storyful to create a curated service called YouTube Newswire. Here too, Storyful’s human editorial team provides editorial triage and authentication services. However, notwithstanding its high minded intentions, YouTube Newswire is likely to remain just another news lite consumer platform (with limited ad revenue prospects) until they figure out how license content to media organizations.

Each of the above initiatives highlight the limits of machine learning for editorial decision-making. Of course, the superiority of human judgment in this area isn’t surprising. What is surprising is that the solutions implemented by the major social networks in 2015 amounted to little more than putting old school newsrooms in place. And, as technologists are fond of saying, humans don’t scale.

Launched in 2006, FaceBook’s basic newsfeed algorithm will be 10 years old in 2016.  And yet it arguably remains the most effective algorithmic curation tool out there.  Next year look for new solutions to emerge that will efficiently combine small human teams (co-located or distributed) with machine learning models to sort and deliver user-generated content more efficiently. Most likely those solutions won’t be developed by the social networks themselves, but rather by early-stage companies, and will be platform agnostic, deployable across the entire social Web to better enable consumers and media outlets find that needle in the haystack.