In 2007, lexicographer and dictionary editor Erin McKean gave a TED Talk calling dictionaries "the lossiest compression format," because the English language is too big to be kept in a book and so much is lost. Her vision of mapping the English language to scale--a sort of linguistic genome project--inspired Elevation Partners founder Roger McNamee to fund McKean's creation of Wordnik, the world's largest online dictionary.
Now that dictionary powers Reverb, launched as an iPad app in November and now available for iPhone, which uses a map of the English language to recommend news stories based on connections between words. Instead of relying on friends' social media timelines, title-based RSS feeds, or recommendations on article pages that are usually part of paid content-sharing arrangements, users are pointed to articles based on topics and areas they've already expressed interest in, and led to new topics through text connections.
"People are so used to the concept of dictionaries that they didn't get how a dictionary could do this," says McKean, who initially tried to get Reverb integrated with existing sites' article recommendation engines. "We'd go to a publisher and they'd say, 'We don't like the recommendation engine we use on our site now, but they pay us.' So we decided to go to the consumer. The longer you read, the more we know what you're about. Sometimes when you look at a new app or reader, they give you 20 checkboxes up front for what you want to read about--but that is not necessarily going to give you the best content."
For the user, Reverb works by developing a personalized word wall that is "like a set of bookmarks, based not on feeds but ideas," says McKean. The starting point is a top news feed generated by Reverb's algorithms for a wide range of topics, and then curated further by a human editor. From that essentially unfiltered selection of current topics, "there's always something that catches your eye," says McKean. "Once you tap on that, you'll see a list of articles on that topic or interest, and then a clean article view, and at the end of that you'll see related topics or ideas. It's very easy to get caught up in going from interesting idea to interesting idea. People spend about twice as long in our app as they do in competing apps, and our click-through rate from one article to another is about eight times as high as the industry standard."
But your trusty social media feeds aren't left out of the equation. One of the most attractive features of Reverb is the social news view, which pulls in your Twitter and/or Facebook feeds and all the articles your contacts are sharing, and organizes them into a word wall to see all the topics they are reading about. This means you can pick and choose topics from a single view ("It's a nice way to get over that fear of missing out," says McKean), and skip over topics you don't care about ("For me it's really useful during March Madness."). Another feature enables users to "favorite" topics to remind them to read later.
And even though Reverb automatically personalizes your news wall, "you're not trapped," says McKean. "We let you add an interest even if you don't see an article on that topic, and if you click on something clickbaity that you don't want the app to integrate, you can long-tap it and get rid of it." This is also useful for things you want to keep private or are personal--say, if you don't want your girlfriend to know you're reading up on engagement rings.
McKean says she launched Reverb as an iPad app first to give users a "bigger space to play" and lead the way in terms of showing the company what features really resonated. Before launching Reverb for iPhone, "we wanted to know what things the iPhone version absolutely had to do when you're standing in line at Starbucks."