“Look,” says Thomas Boeschoten, pointing to the collection of figures and graphs on his screen. “Here, halfway through the article, everyone suddenly stops reading. Let’s see what the problem is. A photo in portrait format, maybe?”

He opens the article on the website of a DPG newspaper, scrolls down to the halfway point and, indeed, a photo in portrait format has been inserted as an illustration. “That’s a real drop-off moment,” says Boeschoten. “The photo is so big that you can’t see the end of it. The text continues underneath but readers have gone before they get to it.”

Drop-off moments: Boeschoten, Team Lead of the News Analytics Team (NAT) of DPG Media, knows several. For example, an intro that does not match the header. “Then the reader’s expectations are not fulfilled.” Or an intro that contains a lot of institutional jargon. “That’s another reason for people to drop off.”

The News Analytics Team gains those insights by studying a variety of data. Everything is measured and displayed in graphs and tables, including what is known as the scroll depth: how far down an article do readers scroll? In that way you can see precisely where they stop reading. Preferably at the end of the article, of course, but the data shows that this is definitely not always the case. Indeed, if 50 percent of the readers finish reading your article, statistically you have done very well.

If the data shows that there is a ‘problem’ in an article or if there are any other ideas for improving an article or its scope, the News Analytics Team provides feedback to the editorial team in question. “We ourselves are not editors but we can give advice based on our knowledge.”

Aren’t they on shaky ground here? After all, editing relates to the content. “Yes, we hear that a lot,” says Boeschoten. “We believe that data is just one of the many sources from which you can derive information; you weigh it up against the other information at your disposal. Insights derived from data have to be subservient, not leading.”

That also applies to the digital newspaper, where the News Analytics Team can gauge how thoroughly articles are being read. “Sometimes we see way back on page 36 an article that’s very much in demand. Perhaps it ought to have been given a better position. Or vice versa: a front-page article that attracts very little attention.” The editorial team itself can obtain such insights from the daily reports, possibly using them to good effect when they put tomorrow’s newspaper together.

The News Analytics Team works closely with the editorial teams of news media and magazines in both the Netherlands and Belgium and in 2021 started providing training for hundreds of editors. The aim is to train data-savvy editors who are able to interpret data and take advantage of it in their own work and that of their colleagues. “We want to develop a culture in which we are receptive to data and can learn from it,” says Boeschoten.

Data also provides interesting insights in the domain of marketing. Margot Rozendaal, Director Data & Analytics at DPG Media, heads a Dutch-Belgian marketing team with more than 30 colleagues who work in support of the news media and the magazines. Their mission is to get a clearer image of the readers. “We use data to try and predict the future behaviour of visitors,” Rozendaal explains. “We can then focus our marketing campaigns on it.”

Data provides support to visitors throughout their visit, whether they are incidental visitors, frequent visitors, frequent visitors who log in, or frequent log-in visitors who ultimately decide to take out a subscription. “We predict what the next best action is for a reader. Should we offer you a subscription straight away or should we first get you to the point where you log into the website? Or if you haven’t got that far yet, should we first offer you a newsletter so that you are periodically reminded of our existence? We do this step by step and, needless to say, in line with privacy laws.”

The figures show that this approach pays off. “We see that visitors who emerge at the top of our prediction models with a high conversion probability are indeed more likely to convert and take out a subscription. Would we have found them otherwise? You can show everyone a subscribe banner, of course. Then you might find the person you are looking for, but you also miss out on many others who you may have invited to subscribe at too early a stage.”

The use of data is handy for advertising as well. Starting this year, advertisers can use the internally-developed data platform, which allows them to choose target groups so as to be able to target their advertising better. With Datalab, DPG Media is able to offer a secure environment, thereby guaranteeing to advertisers that data is reliable and is not being misused. Privacy laws are becoming stricter, so this is an important step going forward.

What is new is that advertisers can now also prevent their advertisements from turning up alongside content with which they do not wish to be associated. DPG Media is the first publisher in Europe to roll out its own system for scanning all the content that is published on DPG websites. It makes use of advanced natural language processing techniques (a form of artificial intelligence) to classify every new article accurately. This enables advertisers to ensure that their advertisements are not placed alongside articles that do not fit in with their brand.

“Our own brand suitability model is a major step in the strategy of setting up a trusted and secure advertising ecosystem,” says Pim van Boekhold, Business Development Manager Data & Demand at DPG Media. “We want to assure brands that advertising on our network is always appropriate and secure, at no extra cost.”