Imagine if an article would adapt itself based on your preferences. Don't have a lot of time for some in-depth reading? Prefer to directly see the headlines? Or would you rather discover the news as video? In today's information overload, we are exploring the technological possibilities to better personalise news, not only in terms of content, but also as a format. In the NewsTAPAS project, we are therefore exploring how we can better inform end users by adapting the presentation of a news article to their preferences and context, such as a particular time or place. In order to test this directly with the end user, we developed a prototype app that we shared via our research platform, the VRT Pilot Zone. In this article, you can discover some of the insights and results.
News articles in different versions
As a starting point, we interviewed a number of end users about how they access and read news during the summer of 2020. With these insights, we created a test app in which we could offer the same news article in different versions.
We ended up making three versions of our news articles for our pilot test:
- The full article, as you can already see it on VRT NWS.
- A concise version that offers the short content of the news article in a number of key sentences.
- A version that only offers audiovisual content in video fragments (when available).
The pilot study
The NewsTAPAS app was shared via the platform of the VRT Pilot Zone. Testers could download the app there and participate in the pilot study. This study included two consecutive phases:
During the first weeks of the test, you could choose which version of an article you would like to read. For each article, you were shown buttons with the available versions. This way, the app could learn about your preferences when reading news articles.
Then, in the second phase of the pilot study, the app automatically suggested a version. The app learned from the choices you made in the first phase. When you clicked through to an article, you could still consult the other formats as well.
What did we aim to learn from this?
During the second phase of the test, we checked, among other things, which versions the end users chose and whether they often switched between versions.
At different times, we also asked the users a few questions:
- Did they find it useful and pleasant to use?
- Did it meet their news needs?
- What did they think of an app that automatically offered them suggestions?
This way, we could find out what our test persons thought of NewsTAPAS as a way of presenting news in different ways.
Some important insights:
- Our end users were predominantly positive about the idea of offering an article in different versions. They saw it as an interesting way of consulting 'tailor-made news'. They found it a suitable way for consuming news quickly, in a format that is adapted to a specific time or interest.
- A remark we often heard was that the content of these various versions needed to be of the same quality. Especially in the short version, a lot of participants thought there was still room for improvement, as the summary did not always reflect the content of the full article very well. According to these users, this was a condition that needed to be met when putting the idea into practice.
- We also learned that not everyone found it necessary for the app to make automatic suggestions. A number of users indicated that they preferred to decide on the moment themselves (and thus preferred the choices they were given in the first phase). We also saw this when we compared the data of the use of the test app across the two phases.
How do we take this further?
With the insights gained, we now aim to explore how we can further improve the content of the summary or short version. One of the things we are looking at is artificial intelligence to automatically generate summaries.