In recent years, Mediahuis Netherlands regional brands have jumped on the digitalisation bandwagon — not by simply publishing all print articles online, but by completely changing work practices optimised for online readers. It has helped our newsrooms drastically improve their performance by listening to what online audience data was showing.
Optimising for online news consumers
In December 2019, the regional newsrooms decided to fully change news production practices. One reason behind this change was reader insights learned from analysing Web data.
A first, not shocking but powerful, insight revealed the publication process was not at all tailored to online behaviour. As figure 1 shows, the production process was still very much based on print principles: Most online news articles were published based on print deadlines, with the bulk appearing online between 3 p.m. and 7 p.m. However, the Web data showed reader activity peaks at various moments throughout the day.
So, we cut our editors off the old print workflow and forced them to work only in the online CMS. We installed various online daytime deadlines by writing and publishing news stories throughout the next day. From that moment on, regional chief editors complied to publication times around 7 a.m., 11 a.m., 1 p.m., 3 p.m., 6 p.m., and 8 p.m.
Another obvious but impactful improvement related to the publication day. At first, articles were published on journalist-convenient days, with publication dips on Monday and peaks on Wednesday-Friday and Sunday (production for Monday’s newspaper). For print production, this made sense. But learning from the data, we understood most transactions happen on Saturday mornings, where our “fresh” news publication was the lowest. So, we further improved our news distribution by not only installing deadline times but also deadline days.
By better distributing our news stories over the week and saving the “best” articles for Saturdays, we increasingly started to find our online readers. This led to a significant increase in transactions.
Improving topic selection
A second major improvement we have made is researching and learning what news topics are popular amongst our readers (subscribers and potential customers). By automatically adding IPTC labels to our online news stories, an international standard for news classification, and a tree-shaped tagging structure, we were able to understand what news topics were most popular amongst our potential subscribers (measured via paywall hits) and our current subscribers (measured via reads).
We also were able to identify where transactions to memberships were coming from.
Now that we understood what news topics various audiences prefer, we were able to tweak our production accordingly. Analysing the subtopics of, for example, arts, culture, entertainment & media, we started to understand we needed to produce less news on literature and more on fashion. Respecting our journalistic principles and responsibilities, there are definitely choices that can help determine how to produce news stories. Simply by activating the insights from the IPTC-label analysis, we learned that many topics can be written with a human-interest frame, making topics like politics far more appealing to our readers.
Thanks to our Web-optimised approach, we were able to better design our news site to the behaviour of the online audience. In just two years, we drastically increased our membership sales and now sell more digital memberships than paper or hybrid memberships.
We’re far from done improving our online practices. There is still a huge amount of work to do translating data insights to improve our journalism. But having taken these first steps showing early success, we have proven to ourselves that small newsrooms can be very agile and quick to adopt change, improving the overall news delivery to our audiences.
We are not afraid of new challenges and deeply implementing data as a source of learning and improving our content. Our motto: We never lose, we win or we learn.
You can see the full Web site here.