Nieman Foundation at Harvard
What’s in your news diet? Sure, we can name the sites, papers, and stations that pepper us with news through the day and week. But we can’t easily sum up what we’ve read and how much of it, or really get an accurate sense of the balance between serious Times or Guardian fare versus the clickbait du jour.
What if we could know what we’ve read — and tune it — over time? What if we could count and categorize our news consumption, as we try to do with our meals, to become better versions of ourselves?
The Daily Beast, the hard-to-classify six-year-old love child of Barry Diller and Tina Brown, believes it’s time for the measured news diet. Check out the Beast’s mobile products and you know you’re not in website Kansas anymore. Both offer profoundly mobile-first, slimmed-down versions of the Daily Beast website (20 million unique visitors, 75 million pageviews). Pick “article feed” or “full view” on the tablet or just move through the article feed on the phone: You won’t find any of the website navigation — topical pulldowns, columnists, “what we like” — on those smaller screens, just the in-your-face Titling Gothic headline font that’s marked the Beast’s unique, in-your-eyes design. Why the sparer design? “Thoughtful reduction” is a driving point of its mobile redevelopment.
What you will find, in addition to the story roll are some unusual numbers at the top of the mobile screens, as in 51 STORIES, 7 READ, 3 SKIPPED. Touch any of those numbers, and a new screen draws down, with the site’s trademark blood red background. It’s entitled “The Daily Breakdown” and it’s an individual analytics console. The Daily Beast calls it a dashboard. You could also call it the dawn of the quantified news reader.
As a species, we’re going quantified at an astounding rate, able to record our life’s data in ways never before possible. We do it with footsteps, calories, online dates, check-ins, Facebook friends and Twitter followers. Why not news reading? Can we really learn what we’re reading and skipping? Do we want to?
Mike Dyer, co-managing director and chief product officer of the Beast, believes the answer is yes. Beast readers like The Daily Breakdown; since it launched in August on the smartphone (and a month ago on the tablet), the Breakdown now ranks among third or better in customer visits, among the site’s pages. More importantly, Dyer believes he’s found a vital engagement driver, with page consumption per session up 2× to 3× since August, he says. Fully 80 percent of mobile users swipe at least once per session. It’s reflexive and habit-forming.
Consequently, the mobile Beasts offer the seemingly Pandora-like innovation of SKIP (yes, in huge letters) on each and every story. Swipe to the left, the word, SKIP appears and then in one fell swoosh is gone, consigned to the SKIP count. Algorithmically, the SKIP is a minus-1, the opposite of a story read. Skip a story, and count the kinds of stories (by topic) you are skipping, and compare your own skips and reads to the crowd within the Beast app within the last hour. You get dynamically generated percentages of the kinds of stories you’re skipping and reading. A read story gets a big red checkmark imposed on it, to remind you what you’ve read and haven’t.
Dyer calls his skipping mechanism the Nudge Engine — nudge as in softly signaling interest.
Is it customization or personalization? Dyer prefers to call it “individuation,” intended to be a lighter touch on personalizing. Skipping stories provides data to the Beast’s content-serving apparatus, but doesn’t currently alter the feed of stories you now get. That’s where the nudge comes in: Red and blue nudge boxes pop up increasingly as readers more actively use the site, offering suggestions to readers. “We may say: ‘You’ve read a lot of politics stories, maybe it’s time for an entertainment story?’ Or, “You’ve read a lot of stories by this writer, do you want to follow him?'” says Dyer. It’s the data readers generate (+1 for reading more than half of a story, -1 for “skipping”) that fuel the kind of individuated nudges readers get. At this point, if readers use both an iPhone and an iPad, the Beast recognizes them as two separate readers, with no cross-platform intelligence gained; it could offer (or require) registration and authentication to provide a smarter cross-platform view, but that would serve as a speed bump in user adoption.
Is the Nudge Engine a new or good idea? It seems to be.
Other big publishers have whiteboarded the words “Pandora for news” and decided against it. They believe it’s just too blunt an instrument. If a reader skips a Bill Cosby story on the Beast, is that because she’s appalled by his behavior, or because she knows a lot about the story but can’t stomach anymore on it at that moment? Does she not want to read about rape, or race, or think too much attention is paid to famous people? There can be a lot of nuance in that simple “skip,” and it’s nuance that won’t be captured on any individual swipe.
That’s why what the Beast has done seems like Pandora’s iconic thumbs up or down but isn’t. “Pandora overplays the value of a thumbs down,” says Dyer. One translation: Journalism isn’t music. Understanding consumer desire in a flow of stories is far more complex than in picking tunes.
Bigger publishers may well consider the nuanced thinking The Daily Beast is bringing to the couple-of-decades-old debate on personalization.
“We have seen a huge uptick in the numbers of people who either want some form of personalisation or are taking steps to personalise their experience,” says Tanya Cordrey, The Guardian’s chief digital officer. “It is now around 20 percent of users on our Guardian apps.” Cordrey is rightly tired of the man vs. machine debate:
It is a topic that already feels slightly jaded as the past few years has seen a constant debate centering around whether algorithms can or will replace journalists. The debate is discussed at the extreme ends of the spectrum with not much in between. The debate, however, will move on in 2015. Clearly there are times when algorithms can work more quickly and more efficiently than humans. But humans are better at handling tone, serendipity, and context. Focusing solely on the tension between human editors and algorithms can also be a dangerous rabbit hole. I believe passionately that success will come with giving much more control to our users. They know best what they want.
The quantified reader idea makes sense to Cordrey: “News organisations will see more users wanting to understand their engagement and interactions with our products. Many of the most passionate users will also want to see more clearly how they contribute to our business models and our social ecosystems. We are currently discussing several experiments in this area.”
At The New York Times, chief information officer Marc Frons continues to test lighter alternatives as well:
The trick here is in inferring reader likes and dislikes, as in the Cosby story example. Says Frons: “Subject-based personalization limits serendipity — one of the main pleasures of social feeds in particular and the Internet in general…For content creators, I am not sure that slicing the report up into micro-individuated bundles is ever going to make business or product sense. But a little bit of personalization within a product can go a long way.”I think passive personalization such as recommendation engines and minimally active personalization (reordered navigation, some subject filtering) can be useful to deepen engagement, and perhaps more closely match readers with advertisers and thus command higher CPMs. I also think people are getting “personalized” news from Facebook and to a lesser extent from Twitter, and that social filters will become increasingly important as they are better refined, as they are in apps like Nuzzel.
Then there’s the balance of customization and high-end curation, he says:
The challenge for The Times and other large, serious news organizations is to sharpen our focus on what people want to read/watch/interact without giving an inch on what is journalistically important…At The Times, our most loyal readers come to us (as opposed to reaching us through the usual social or search channels) for our news judgment as much as the individual articles. Personalizing the home page, for example, would fracture that shared experience of what is important and newsworthy. But a great recommendation engine that surfaces content you might otherwise have missed can be a powerful asset for publishers hoping to hold their readers’ attention for more than the scant minutes most people spend on any given site.
The Economist works with a similar balance in mind, and a different value proposition: finiteness. Here’s Tom Standage, The Economist’s digital editor:
Our deal with the reader is that they rely on us to be a trusted filter on world events: so we tell them what’s important. That’s rather difficult to square with a system that lets readers decide what they do and don’t want: our readers are in effect saying “I don’t have much time, so tell me what I need to know, even if I don’t think I’m interested in it or don’t realise it’s important.”
So we’re not doing any personalisation at the moment…That said, I think there is some scope for personalisation. Things we are thinking about: a “read this first” feature that looks at past reader behaviour to present what they’re most likely to want to read first, in app or on the web. (Some readers always start with the obituary, or the business pages.) Another area this might be useful is for audio, and in particular in-car audio: a smart playlisting feature, if you like, that plays the stories you probably most want to hear first.
The Daily Breakdown is a good 1.0 product. It’s simple and intuitive. Expect refinements to roll out next year, providing more quantification alternatives (the possibilities are staggering, given the complexity of news flow), undo functionality, and more. The Breakdown is just one of several Beast innovations in mobile. In January, the website’s Read This feature — highlighting five algorithmically picked stories — will debut on mobile as well. In allowing readers to “follow” individual writers, a feature The Guardian added this year as well, the Beast is getting closer to understanding what readers want.
Why offer quantification? The early engagement numbers tell one part of the story. Publishers can quantify for same reason Weight Watchers does: Keep the customers coming back. In the hypercompetition of digital news, what distinguishes one site from another? Certainly, unique stories and above-the-crowd voices, but tools also encourage habit. If you’re sold (or semi-sold) on the Beast, or Slate, or The Washington Post, or WBUR, or Time, might you visit more often to check your running totals? Loyalty and direct traffic become greater issues in the mobile news age, with both more potential and more peril. (Digital veteran Cory Bergman’s recent Mobile Media Memo piece urging publishers to make their mobile front doors more appealing is right on the mark.)
We could get greedy and want such quantifying across news websites, tracking our whole news diet. The Daily Beast’s trick is site-specific, suiting the tech of the moment and, of course, its own business needs. If indeed we want to become quantified readers — do we? — we’d need better collection, aggregation, and sorting. How might that work, and who would do it? Facebook, Twitter, Evernote, and a host of curating/aggregating businesses would love that role. But that’s a question for another day. We’ve long talked about the nature of quality in the digital news landscape. Now we can add new measures of quantity, personal and collective, to that discussion.
