Building Better Story Formats for Live Coverage
What we’re using and testing, from live blogs to mobile alerts
Live coverage is a big challenge for newsrooms. It sits at the intersection of high-stress moments and production-intensive story forms. We use a variety of tools to help us with breaking news, but they’re typically not the ones we use day to day. And on top of that, we still need to think about improvements and new ways to reach our readers.
SRCCON 2017 gave us the opportunity to discuss the pros and cons of story forms we use for breaking news and live coverage. It brought together people from newsrooms large and small, with a wide variety of experience. We the facilitators—editors Tiff Fehr and Hamilton Boardman from the New York Times and Alastair Coote from the Guardian’s Mobile Innovation Lab—were eager to talk with just such an audience about these persistent challenges.
Classic Story Forms
To establish a common vocabulary for our discussion, we first reviewed classic live-coverage forms: articles (when used for breaking news) and live blogs. We discussed the pros and cons of each form, both for newsroom staff and for readers.
Articles: The (Sometimes) Evil We Know
Articles are a time-honored and battle-tested breaking-news story form. No other form is more familiar to readers. Reporters and producers are also well versed in breaking-news articles. Yet it takes time to produce a good article. In the internet age, readers are not accustomed to waiting on (even speedy) article production when news is breaking. Our efforts to work faster are rarely a good fit for anyone’s needs.
In the worst-case scenario, a breaking-news push-alert on a reader’s mobile device may click through to a stub of an article that offers little more than the alert itself. These articles are published as quickly as possible, of course, but readers expect immediacy, and many will only ever see the stub. Some might remember to return again within minutes, and fewer still will re-read a developing breaking-news article, hunting for new reporting.
That’s a lot of cognitive effort in high-stress moments, for both readers and reporters. Readers may not get far into an article if the toll seems too large, no matter how familiar the form or comprehensive the reporting.
When articles don’t offer the right shape or pace, newsrooms often turn to live blogs so they can publish faster than their usual tools allow. But the amount of information in a live blog is frequently intimidating and poorly-organized for most types of readers.
Frequent exposure to live blogs means loyal readers expect something live-blog-shaped during breaking news events. Many readers end up skimming for new information, and our helpful features like sorting and filtering rarely meet their needs and are rarely used. They also expect to see signs of live-ness while they visit the page, however brief that may be. In some cases, newsrooms end up “feeding the beast” with information they might exclude in other formats. This may be a way to signal live-ness and freshness to readers despite incremental information.
The relationship between blogging and news is still evolving, despite more than 20 years of symbiosis. We seem to presume that live blogs are time-tested and fully evolved. (We also thought that about classified ads.) It helps to remember that blogs are essentially a hacky coincidence: little study has been put into the form of live blogging, and research on the usefulness of blogs to readers has been piecemeal, at best. Live-blog usability assessments are even more rare.
At the New York Times, we analyze reader behavior around our live blogs whenever possible and have found a vexing mismatch between our presumptions about live-blog readers and how those readers actually behave. The idea that readers stay on a live blog waiting for the most recent update is a compelling but mythical image. Few people outside of newsrooms read this way. A more representative pattern is that readers enter a live blog midstream, a few hours after news breaks, and encounter a bloggy deluge that is too high-context and therefore alienating. They skim and then leave for other news sources, with the goal of validating information gleaned from each.
Sidenote: The Mythical News Junkie
Throughout our SRCCON session, we considered what we currently know about the breaking-news audience. We discussed what forms seem to resonate, what traffic patterns emerge, and what myths we hold about reader activities.
One persistent myth we want to bust is the idea of news junkies who wait excitedly for every word of live coverage we can publish.
A brief analysis of a couple of breaking news events conducted at the Times in 2016 contrasted readers who viewed live blogs with those who read traditional breaking news articles. It found no indication that the live blog readers spent any more time on the site, read more articles, or came back more often than the traditional article readers. The cohort of live blog readers, in short, behaved almost identically to those who were article readers. If we were serving a “news junkie” audience with our live blogs, it certainly wasn’t showing up in our numbers.
With these tensions acknowledged for both reporters and readers, the Times has reduced its live blogging for the moment and expanded our breaking news forms overall. Our newer story forms aim to better align with different types of readers and production speeds. Some newer forms include live briefings and live chats—though, of course, neither are all that new. But their (re)application as part of the Times’ breaking news coverage has been constructive.
Twitter is an obvious inspiration for chat-shaped coverage. Many of the Times’s best chatters are also prolific on Twitter, and we needed to harness their abilities. A chatty, conversational tone is engaging for readers, and it can make explanations of complex news events feel more like banter between well-informed and incisive friends. But that ideal note is hard to strike; even our best efforts can wind up reading like a typed-out version of a cable news panel. Even our most adept chatters can also struggle to avoid lulls in conversation and overlapping threads.
And those challenges come before we consider heightened reader expectations. We believe expectations are higher for chats because readers have a strong idea of a good chat’s pace from non-news discussions. They expect to see a flowing, stimulating discussion, with frequent live-ness signals. The trap with live chats is that they still encourage “feeding the beast,” even if this particular beast encourages brevity.
For many readers, the tone and pace of a chat fit their needs, but perhaps not the amount of time they have to spend following along. Consequently, we see fewer readers opt for this form.
As the Times pulled back from reflexively spinning up a live blog, we have started to more frequently use a form that we’ve internally been calling the “live briefing.” It’s a bit of a hybrid of a traditional article and a live blog. The idea is to build a format that is more welcoming to a reader coming into the news midstream, while also allowing some of the flexibility of a live blog.
In form, the live briefing generally begins with a short set of bulleted updates bearing the latest news and key facts, followed by a collection of short and independent posts that expand on the key points of news or offer color and context. In theory, this form allows editors to structure information much as they would in a classic inverted-pyramid news article, while removing the overhead of having to construct the connective tissue that ties a traditional article together. For better or worse, the live briefing dispenses with the reverse chronological order so common with live blogs, and (attempts to) focus on putting the most important news updates first.
Our live briefings, thus far, have not been without their pitfalls: At the Times, there is no custom software supporting them—they are produced as normal articles in our CMS. This leads to some workflow and publishing bottlenecks. And without a sharp and critical editorial hand, they can often sink under the weight of too many story-chunks. It also remains difficult for editors to resist the temptation to top their briefing with what looks a lot like a traditional article top.
And briefings can still leave repeat visitors with the same task classic articles require: mental sorting through what is new and what isn’t.
Our most recent experiments in live coverage forms are often tied to special events like elections or sports, because they benefit from having a strong data source. Meaningful updates for readers becomes a mix of news, analysis, and significant data changes (updated counts, shifts in momentum, winner calls, etc).
Another shared characteristic in these experiments is a focus on mobile. Alerting readers via notifications has become a huge part of breaking news, and alerts are just the start of our efforts to earn a reader’s attention.
Many readers learn about breaking news via an alert. A smart, thorough first impression counts. Looking at the different notification APIs available through the web (on Android devices) and in iOS apps, we saw that there was room to take these breaking news alerts beyond the form we all know. Alastair Coote has recently focused on this topic specifically, through his work at The Guardian’s Mobile Innovation Lab, and he led this part of our SRCCON discussion.
The first improvement was one the Lab called “self-updating notifications.” The title is technically misleading, since it involves pushing multiple notifications sequentially, but the appearance to a user is that of a single notification with changing content. At its most simple, it could be used to update the information in a breaking news alert, so that the notification is never out of date when a user reads it. But we took it one step further and started to think about the notification as standalone content, and what we could provide.
We conducted a number of experiments in “persistent” notifications that updated throughout the course of an event: the Brexit vote, the American presidential election, the Super Bowl, and the 2017 British election. In each case, we kept the notification up to date with the latest results—and used appropriate imagery to convey those results to users quickly. At any point, the user could tap on the notification to dive into our deeper coverage, but the persistent notification meant that our coverage was always on-hand, and we saw that users tapped through multiple times throughout each event.
We also experimented with some iOS-specific improvements. As part of our coverage for the inauguration this year, we pushed out a live video notification to users of our experimental app. Like the persistent notification, it was designed to be left on the user’s lockscreen throughout the day. Instead of the data-driven display we used in those events, we listed “key points” in the collapsed notification view, telling users what event they should expect to see if they opened notification. The expanded view (opened by swiping down or performing a “force press” on the notification) showed a video player with mute/unmute controls, as well as action buttons to open our live blog.
It’s still early days for these notifications, but our experiments were very encouraging. Users are familiar with notifications, and incremental improvements to functionality means we can meet users where they are, rather than force them to navigate their way to our coverage. Here’s some of what we learned:
- Web Notifications Introduction: News on Lock(screens)
- Now From Your Lock Screen
No matter how effective the notification, readers will still want to dive more deeply into coverage at some point. With that in mind, we re-examined the traditional live blog, and the assumptions that it makes. The result was an experiment we called “Shifting Lenses”—a mobile-first re-invention of the live blog that better reflects how they are used today.
The primary difference is the ability to swipe horizontally to view an event from different perspectives, or through different media. During the inauguration, we provided “on the podium” and “off the podium” categories, while during the Superbowl we split our coverage into “on the field” and “ads & half time.”
An important technical change was in how we handled social media embeds. Rather than rely on third-party widgets, we rendered them manually ourselves. This required a larger coding effort, but the result on mobile devices (lower-powered ones especially) was very noticeable: scrolling that had previously been slow and jerky was perfectly smooth.
Third-Party Distribution: The Only Semi-Discovered Country
Our spectrum of live coverage forms and near-future experiments has one large oversight: readers who encounter our breaking news offerings within other websites. We may have the smartest chat, briefing, and notifications, but that won’t matter if we can’t get in front of readers when they look for breaking news. In recent years, that means going beyond our own sites and onto platforms owned by others.
Search engines, social media and mobile feed integrations are the latest battlegrounds for reader attention. In order entice readers with fast results, Facebook, Google, and Apple each now request or expect news outlets to publish customized, performance-oriented versions of our articles for integration with their platforms—formats called Instant Articles, AMP, and Apple News, respectively. The tech giants incentivize use of those formats by offering preferred placement in their sites, with arguably improved reader engagement. But as we have laid out, we need the ability to use newer forms like briefings and chats. And these forms may not fit with strongly defined third-party formats.
We didn’t get to spend as much time as we would have liked on this topic during SRCCON. But we hope to resume the discussion in greater detail sometime soon. In the meantime, the world of third-party presentations continues to evolve rapidly. As do reader expectations on those platforms.
The form of our live coverage is the kind of topic that benefits from cross-newsroom information sharing as we all try new forms and experiments—and as we learn more about our readers during breaking news cycles. We hope to continue discussing it with more newsroom citizens as our breaking news forms evolve.
Hamilton Boardman is a senior editor on the news desk at The New York Times. He his currently the acting deputy Washington editor for digital. Previously, he spent nearly eight years editing The Times’s home page and coordinating live event and breaking news coverage.
Alastair is a developer with the Guardian Mobile Innovation Lab, where he experiments with better ways to deliver news to mobile devices.
Tiff Fehr is a senior interactive developer with the Interactive News team at The New York Times. Previously she worked at msnbc.com (now NBCNews.com) and various Seattle-area mediocre startups.