How to use data to understand readers' needs

When the BBC World Service looked into how its content performs, it realized that 70 percent of articles contributed to only 7 percent of traffic. Simply put, most of the work that journalists put their time and energy into sat on the website barely read.

That was five years ago and the newsroom has since adopted an audience-first model, using the concept of six user needs. The basic premise is that readers need more than just breaking news: they want to understand the context, be able to make up their own mind about a topic and also be inspired or entertained. How much content a newsroom needs to produce in each category depends on the preferences of the audience but, on the whole, the public expects more from journalists than just news updates.

At the BBC, this approach was championed by Dmitry Shishkin who later joined travel journalism start-up Culture Trip and is now an independent digital publishing consultant.

Most recently, Shishkin has been working with editorial analytics company smartocto on what became known as the Triple N project - News, Needs, Notifications - that wants to make the user needs approach more manageable for busy editors. The tool uses an algorithm to identify user needs and it then sends notifications to suggest actions a publisher can take to increase reader engagement.

After about a year of research, smartocto has published the first white paper 'Actionable user needs'. We unpack some of the findings.

Newsrooms and user needs

Newsrooms used to be a place where a wise editor just "knew" what the audience "wants" and the journalists were writing for said editor rather than for their readers.

From there, we leapfrogged to an era with such an abundance of data on user behavior, preferences and dislikes we struggle to wrap our heads around it. What is important for your editorial strategy? Page views or returning readers? The number of articles read in a day or social media shares?

To make these decisions, editors have to first understand what the readers really need. Only then can they commission articles that people will find useful and engaging.

Now here is the more difficult bit. Even when a publication knows its users' needs, how does it go about writing individual articles? How do you decide whether a story should be a brief update or an in-depth piece providing wider context? And how about a listicle or an expert comment?

The most common problem is that newsrooms often massively overproduce news updates and ignore user needs for other types of stories. They lack insights on timing, format, channels and topics that users prefer which diminishes the impact of their work. Journalists also often fail to write follow-ups to stories the audience is engaging with, thus missing out on the opportunity to produce more useful content.

From data to action

For the white paper, smartocto worked with three publishers: Omroep Barabant, a public broadcasting corporation in the Netherlands, Belgian newspaper De Morgen and IDN Times, an Indonesian digital media publication aimed at millennials and Gen Z. With each of these companies, it identified their pain points and roadblocks to their growth.

So far, smartocto developed nine notifications to help editors see how a story is performing. Some of them are pretty straightforward - exceptional page views or Twitter usage - while others alert you when your story is still going strong after a major peak and recommends a follow-up from a different user need.

The algorithm proved to be mostly successful for the three publishers although it still needs some work. What is more important is that everyone saw a growth in audience engagement and size once they started to tailor their content to user needs.

All that may leave you wondering - what will happen to journalistic instincts? Are journalists now going to be taken hostages by the data and only write something because the computer says so?

The simple answer is no. Metrics are there to give journalists guidance, not to dictate what and how to write. After all, we want our work to have an impact, not to sit on a website and gather virtual dust.

"The beauty of a user need model is simple - it comes from users," writes Shishkin in the report.

"The model itself is external, not internal - it challenges newsrooms by reminding them not to preach but to listen, before starting on. And having worked with various newsrooms in my life, I can see a direct correlation between strong performance, finding your audience and satisfying its needs properly."

You can download and read the full report here.

This article was originally published by Journalism.co.uk. It was republished on IJNet with permission.

Photo by Serpstat from Pexels.