User Interest Predictions
2025 Finalist
Media associated with this campaign
Overview of this campaign
Schibsted Media is a media organisation that consists of many different news brands serving a plethora of users in the Nordics. Among these brands, you will find primary news destinations like VG and Aftonbladet, subscription newspapers such as Bergens Tidende, Svenska Dagbladet and Aftenposten, as well as niche brands like MinMote and TV.nu.
Schibsted’s winning ambition is to be the leading media destination in the Nordics, reaching and empowering millions of people in their daily lives through offerings with superior relevance. Users visiting any of our sites signal interest in specific content, whether it is politics, fashion or e-sport. Until now, Schibsted has not had a uniform way of identifying users’ interest across the organisation. User Interest Predictions (UIP) were created in order to meet our ambition and ensure that we not only deliver valued personalised experiences to our users, but also enable data-driven decision-making in the company.
Demographic data such as age, gender and location have long been used as a proxy for interest. Other attempts at building interest profiles have focused on inferring interest by using newspaper sections. Although there is overlap between section and interest, our findings showed that it is not readily interpretable as the same. Because Schibsted is a family of brands, each brand defines their sections independently based on their own needs, thus resulting in constraints created by brand-specific categorizations.
In order to solve the aforementioned problems, we leveraged AI. UIP is a machine-learning-based solution that uses data from across all of Schibsted to predict users’ interest. It uses natural language processing to classify content, then a model predicts and aggregates interest and interest levels for each user based on their browsing history. Interests are categorized based on a fixed, well-known taxonomy, to ensure flexibility and sustainability for a fast-changing news landscape. This approach allows new interests to emerge as they become more relevant and others can become dormant as their relevance decreases.
Results for this campaign
In the first iteration of UIP, the model was developed to predict interest in 25 categories – the top tier in our taxonomy. A series of experiments were conducted in collaboration with our subscription department to validate UIP’s utility. Users predicted to have high levels of interest in certain categories were sent personalised marketing messages through email and content cards embedded on our news sites.
In the first set of email experiments, users identified with a specific interest received emails pertaining to said interest. Click-to-open rate was significantly higher in the UIP group compared to the control group, indicating UIP can efficiently identify users’ interests.
The second set of email experiments sought to increase engagement and reduce churn. UIP was used in combination with churn predictions and other behavioural data. Within the sports segment, results showed a 3.06 percentage point increase in the number of people who read sports content within 7 days after communication compared to the control group.
The last set of experiments focused on content cards on the front page of our news sites and looked to increase the number of sales. All experiments returned significant results, with two resulting in an astonishing 167% and 256% improvement in new sales.
UIP was later expanded to Tier 3 in our taxonomy and the model can now predict interest in close to 1,000 categories. In order to democratize the insights this brings, interest dashboards have been created for employees to explore different combinations of brands, demographics and more in order to uncover the interest profiles that lie within our readerbase.
We sought to harness the power of being a family of brands and give our users personalised experiences based on their interests and not our guesswork, and our results prove that it is a winning recipe. UIP is currently being adopted by our advertising organisation, and as more corners of the organisation awake to the power of interest-based personalisation, we are seeing more UIP experiments on the horizon.
