Mediahuis newsroom uses Article DNA to help each story find its audience

By Yves Van Dooren

Mediahuis NV

Antwerp, Antwerp, Belgium


If you’re a journalist or a content creator, you know how important it is to write stories that resonate with your audience. You want to capture their attention, inform them, and inspire them. As a publisher, you want to invest in journalism that supports a sustainable future.

Article DNA tries to deliver better answers to that question. What are the missing characteristics that define the stories that really resonate with our readers?

If we have better insights into the intrinsic value of our stories, we should be able to align the way we write and publish stories with the likes of our audience. Hence, “The secret life of article characteristics.”

As a start, we needed to look at our stories from a different perspective. Instead of focusing on the topics we cover, which are largely determined by the news market, we decided to focus on how we write our stories.

We first made an inventory of all kinds of attributes. Some can be determined with the help of natural language processing (NLP) techniques. Some turned out to be better added by the editors themselves (e.g., story nature, user need).

In the first stage, we analysed performance using only automatically derived characteristics (e.g., tonality, emotion, style). We found out that it was difficult to get clear results from those.

So, we turned our heads to the newsroom.

Introducing Article DNA to the newsroom

The newsroom is an operational environment with tight deadlines and schedules. Asking journalists to fill in extra information next to writing their stories is a daunting task.

First, we held discussions with the editors-in-chief to explain the different characteristics we wanted to add and the effort it would require from the journalists.

We ended up with this bare minimum of three article DNA elements:

  • Story nature: Why did we write this story in the first place?
  • Genre: What is the format of this story?)
  • User need: How and for what purpose is this story written?)

We integrated this information into our content management system so every article has a dedicated pane where the journalists enter the newsroom Article DNA.

Mediahuis developed an Article DNA pane for its CMS.
Mediahuis developed an Article DNA pane for its CMS.

We started with a pilot project involving one Mediahuis brand and small groups of journalists from different sections. We trained them on how to use Article DNA and how to assign the appropriate characteristics to each article they wrote. We also created a set of guidelines and best practices to ensure consistency and accuracy.

After weeks of testing and fine-tuning, we rolled out Article DNA to the entire newsroom. We organised demos to inform and train our journalists on using it effectively.

Analysing and optimising with Article DNA

Initial results after implementing Article DNA were promising.

It was easy to detect positive and negative combination “outliers” during those first analysis efforts. We saw that inspirational sports pieces were read extensively,  and those insights were immediately translated into concrete newsroom actions.

After the initial weeks, we made the Article DNA information mandatory, and an editor had to review and approve the characteristics before publishing an article. 

From then on, everything that got published included the new characteristics and more in-depth analysis became possible.

At the same time, it also became clear that those what/why/how combinations need to be approached with a certain nuance. A strong news site or app is not just a bunch of stories; it’s a well-balanced mix of topics and story types that suits your audiences’ needs. One-trick pony news sites don’t exist.

Creating a content mix

That’s where our Content Mix Dashboard came in. For the first time, we could visualise the mix of different formats and user needs for the different brands and sections.

The content mix overview of the sports section at De Limburger.
The content mix overview of the sports section at De Limburger.

That gave us an idea of how we write and present the news to our readers. It also told us how our readers and prospects react to a specific mix of stories by evaluating the attention time and subscriptions sold. This allowed us to develop specific use cases to inform our newsroom on how to bring stories that are best suited for our brands and audiences.

One of the examples is the “Reading Clock,” which shows specific news teams the best times to publish for a specific user need. 

The Reading Clock shows the best times to publish stories based on specific user needs.
The Reading Clock shows the best times to publish stories based on specific user needs.

Newsroom impact of Article DNA

Honestly, asking our journalists to take care of extra meta information was not easy. It took the belief and persuasion of our editors-in-chief to get things rolling. Getting everyone to add high-quality data needed constant attention.

But once the newsrooms got the hang of it, we saw new types of discussions popping up. Our journalists and editors were using the new vocabulary that was introduced. There were discussions on what user need was best to cover a specific story while audience teams were monitoring whether they offered a healthy content mix on the homepage and newsletters before sending them out.

The most surprising result of the Article DNA initiative was that the added characteristics, originally intended for better analysis, became part of the daily discussions in the newsroom. Terms like connecting or inspirational pieces are now used widely by editors; previously, this terminology didn’t even exist.

For the first time, analytics input is now actually used to support how stories are best written to fit our brand and audiences. For our newsrooms, this was truly innovative.

About Yves Van Dooren

By continuing to browse or by clicking “ACCEPT,” you agree to the storing of cookies on your device to enhance your site experience. To learn more about how we use cookies, please see our privacy policy.