To truly pivot towards a digital subscription model, MittMedia needed to innovate, scale, and monetise the digital ecosystem. To be able to do this with a constantly experimental approach, we decided to build a data platform from the ground up. 

The innovation of the platform itself is MittMedia’s capital for future sustainability. Having a constantly improving foundation is also necessary for doing business development without constraints.

MittMedia wants to anticipate societal changes and rapidly adapt to them. Owning our own data and platforms was the first and most important step towards this goal. Therefore, we transitioned from external system silos to our own in-house platforms for everything we considered core functions:

  • Content production.
  • Recommendations.
  • Distribution.
  • Paywalls.
  • User and advertising data.

In addition, we built a central data platform called Soldr, which collects, processes, and sends out data between all our systems. Soldr is the “data brain” of MittMedia and acts as the hub that makes innovations possible. 

A well-implemented data platform at the centre of a news publishing organisation must be able to fuel strategic synergies between editorial staff, the paid content business, and the advertising business. 

Johannes Lindén, machine learning researcher at Mittmedia, presents a natural language processing model based on data stored within the Soldr platform.
Johannes Lindén, machine learning researcher at Mittmedia, presents a natural language processing model based on data stored within the Soldr platform.

As the data platform took shape, it had immediate effect on the organisation. In the wake of the first implementation, we were able to coordinate several working groups in MittMedia, resulting in a higher and more relevant content output. 

The most important aspect of company data is to find the right KPIs to guide the organisation, but Soldr also had many other direct effects on the strategic work, including:

  • Setup of new data team to take care of collecting, storing, and distributing data, data science, machine learning, and KPI processing based on Soldr data.
  • Automation of manual work in the content management system (CMS), freeing editorial resources to create more unique content.
  • Data analytic tools are based on Soldr data and allow insights on which B2B, B2C, and the editorial team can act to develop new strategies.
  • Setup of an editorial content optimisation team focusing on innovating approaches to content creation.
  • Aracua is content production and planning system for editorial. Using Soldr we are able to build a feedback loop, which helps journalists prioritise and optimise content ideas based on prior user insights. 
Profiling and quantifying user behaviours with UX research, later to be stored in the platform.
Profiling and quantifying user behaviours with UX research, later to be stored in the platform.

We are effectively using the Soldr platform to scale long-term innovation. After its implementation, we have observed measurable innovative results for our organisation:

  • By understanding content and advertising performance, editorial staff, B2B, and B2C, we can fine tune content consumption short term and develop innovative strategies long term. A new content optimisation team focuses only on finding data-driven innovative ways of using content that already resulted in new verticals and robot-text initiatives.
  • It has led to higher relevance of articles shown on the start page and thus increased click-through rate. By rolling out personalisation on all sites and apps, we could also automate 70% of all manual work in the CMS so editorial staff could focus more on content creation. 
  • Soldr also provides geographical data, resulting in innovations like the activation map for editorial staff and for a B2B map which shows their customers’ campaigns by visualising the reach-out-radius to their end customers. Soldr data is even used to fine-tune push notifications to activate geographical regions outside of MittMedia’s primary reach. 
  • Auto categorisation improves the basis for machine learning and will innovate the editorial content creation and planning module to automatically help journalists link their events to geographical spaces and physical entities.
  • Research into a content API providing Soldr data to external sources has begun. It allows external innovations on our data, and it allows us to access and combine external data with ours. The content API opens up for even more innovation potential and is considered a capital for future investment.
  • Soldr has also led to cultural changes in MittMedia. Based on Soldr data insights, new organisation wide KPIs have been set up, which result in new strategic alliances between earlier unconnected departments. Knowledge is spread in cross-departmental initiatives to tackle company-wide challenges together, for example, by developing a common churn and user model on which we can act following a read thread. 

To build a data platform to represent the entire digital ecosystem was a daunting undertaking, but it also was completely necessary to drive the digital transformation in line with the chosen strategy.

Since the systems are now in-house controlled, we are able to catch and flexibly act on the needs of our customers and the organisation. The result is an upstream of relevant ideas, needs, wishes, and insights. To understand what really is important to tackle first, a new aspect of company culture emerged: encouraging data-driven thinking.