Establishing relevance to reduce churn was a theme during the INMA Consumer Engagement Summit in Miami, Florida, where three Scandinavian media companies discussed their strategies on Thursday.

Aftenposten CEO and Editor-in-Chief Espen Egil Hansen talked about redefining relevance in the age of algorithms.

There once was a time when the newspaper was both the habit and the reward. “It was all in one package,” Hansen said. “We are not in a monopoly situation anymore. There are so many ways of getting information and using our time, so now relevance is really important.”

During the last 10 years, Aftenposten has flipped its business model from 70% of revenue generated by ad sales to 70% of revenue coming from subscriptions.

“Next year it will be 80%,” Hansen said. “Subscriptions are approaching an all-time high in Norway. Part of the story is decline in advertising. The other part is growth in subscriptions. This strong growth period comes with new challenges.”

In an overwhelming media landscape, establishing relevance is the key to keeping subscribers engaged and reducing churn. Algorithms are useful for identifying content that leads to conversion and high engagement, and they can support the journalistic mission.

That said, Hansen believes the algorithm strategy needs to be owned by the newsroom.

Aftenposten uses an “editorially guided algorithm” that takes into account user behaviour and preferences, but also is influenced by editorial staff deciding what content is important. This helps to avoid filter bubbles or echo chambers.

Hansen said Aftenposten also is building relevance and loyalty based on its journalistic tradition “in a time when news institutions are being attacked all over the world.” As a symbol of its commitment, the original newspaper logo from 1861 was at the center of a recent redesign.

The addition of a brand new podcast is driving engagement and subscriptions with a new audience, Hansen said: “After five weeks, it is incredibly popular, reaching younger listeners and 70% are already subscribers. They are well educated and female. It’s a dream group. We are on the brink now of understanding that Aftenposten cannot be about one channel. It’s about being who we are in different situations.”

Events are also a key part of Aftenposten’s engagement strategy. The company now organises about 30 events per year, Hansen said. And the Aftenposten Jr. children’s newspaper is proving a good engagement tool with 30,000 subscribers, mostly 7- to 12-year-olds.

All of these initiatives have been welcomed and supported by the newsroom because journalists want their articles to be read, Hansen said.

Martin Jönsson, head of editorial development at Dagens Nyheter in Sweden, shared its strategy for fighting churn with digital quality and data — lots of data.

“What we focus on is building the loyalty. We believe we can only achieve loyalty by focusing on quality,” Jönsson said.

Total newspaper subscriptions in Sweden fell by 11% in 2017. Yet during the past four years, digital-only subscriptions have grown steadily, and in 2017, digital reader revenue increased by 92%.

“People are more willing to pay for content now, whether it is journalism or music or film,” Jönsson said, adding that there still is a fairly low threshold for the price people are willing to pay for these subscriptions. “It is a very narrow price range — approximately USD$11 — so price increases are a challenge we face.”

Research about Sweden’s digital subscription market shows that 13% of those who subscribe to content channels are paying for a mix of film, music, and newspapers – indicating that consumers are more willing to pay for content and even will pay for multiple types at once.

This has led the company to emphasise quality of digital products and improve their offerings. Dagens Nyheter offers three main subscription levels:

• Base (about 60% of subscriptions are base level).
• Medium, which includes e-paper, crossword, and archive access.
• Bundles with other publications such as The New York Times, audiobooks, and another local newspaper.

The company also has three kinds of paywalls:

• Metered model (three free articles per week), which represents 10% of all conversions.
• Premium model (select content with two to three articles per day), which represents 35% of conversions.
• Dashboard model, in which successful articles are put behind a paywall after several hours once they meet certain indicators set in the algorithm (15 articles per day), which is 55% of all conversions.

“We test and tweak these all the time,” Jönsson said. “And we work a lot with tools to make data accessible to everyone.”

For instance, a writer receives an e-mail report for each article showing metrics that measure how successful it is. Editors can see dashboards and top lists that track the most important KPIs. Everyone in the organisation can see such measurements as total engagement scores, pageviews, external traffic shares, and other data that helps to identify the most successful and promising content.

Jönsson also emphasised the design of these tools is important. Data is represented using graphs and graphics with clear measurements and colours that are quickly understood — for example, green, yellow, and red to indicate whether a particular goal has been met.

The data collected has helped to predict churn by looking at such indicators as methods of payment, how often or how long a reader spends on the page, etc.

“Then we looked at micro segments to see how many subscribers we have in risky segments, Jönsson said. “Anyone who is high risk we would communicate with a different approach.”

These combination of these efforts has reduced churn, Jönsson said. “We improved to a record low level of churn 8.2% and returned to growth rates with lower costs. We are down to 1.5% churn among longtime subscribers.”

The entire organisation has shifted away from the traditional focus on acquisitions, customer service, and engagement, Jönsson: “Now we are agile and truly cross functional, and integrated with the newsroom to get them involved in the entire process.”

MittMedia’s Thomas Sundgren, head of platforms and strategic parterships, and Magnus Engstrom, head of data strategy, share how the company has transitioned to machine-driven content personalisation.

“Machines are now running the show for 150,000 daily active customers every day, every hour, every week. Every night, algorithms are recalibrated. Machines are doing a far better job than humans, and this frees up our humans who are producing unique journalism.” 

The mission at MittMedia comes into play with this journey, Sundgren explained: “We have a mission to contribute and uphold local democracy by creating and distributing journalism and information. To fulfill that democratic mission, we need to transform ourselves and do it pretty fast and very agile. The first step is transitioning to the idea of a media company that we want to be.” 

Sundren and Engstrom explained how MittMedia is transitioning from a legacy media company to one that is a partner to local people. Doing so is the “only way for us to continue to stay relevant,” Sundgren said. 

Key to this is understanding the routines and habits of customers and users. Last year, the UX (user experience) team at MittMedia began collecting stories of actual users. Then data specialists mapped those stories with data patterns. 

“Suddenly we had a machine learning problem which we could use to fit into models,” Engstrom said.   

This is the foundation for MittMedia’s personalisation initiative, which it began 18 months ago. 

“If we want to be able to monetise, we need to create products where supply of product meets customer demands,” Sundgren said. “It’s working. Relative consumption is going way up. We are now live with personalisation with online and newsletters; apps will come later.” 

Engstrom shared one key takeaway from MittMedia’s journey: “Take a step back. You need to build something that will last for some time. Otherwise, it’s very hard to get back the user groups to move to another type of device.”

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .