As technology continues to advance, personalisation has become a buzzword in the publishing industry. And it’s easy to see why. Today’s readers expect tailored information that caters to their unique interests and preferences, and they’re not afraid to look elsewhere if their needs aren’t being met. 

At Corriere della Sera, we understand the importance of providing readers with the news they care about, which is why we’re proud to introduce “Le Tue Notizie” — Italian for “Your News” — our latest experiment in personalised news delivery. 

Thanks to the power of Artificial Intelligence, “Le Tue Notizie” allows our subscribers to tailor their news section according to their interests, complementing the traditional homepage. Using data about subscribers’ preferences, our content recommendation engine selects articles unique to their needs, delivering an experience that’s both satisfying and exciting.  

Our A/B testing approach

Corriere della Sera is committed to improving readers’ experiences through an A/B test approach and customer-driven evolution. Initially, we selected a cohort of 33% of our subscribers who visited our Web site to test our new content recommendation engine. We aimed to increase visit depth, frequency, and time spent per visit compared to other subscribers.

To achieve our objectives, we developed a new page managed by our content recommendation engine. We made it available to the test group through a link in the header on our Web site’s home page. We continuously optimised the recommendation algorithm by conducting A/B tests and differentiating the setting for mobile and desktop use. 

After seeing positive results in the first weeks, we expanded the test to 66% of our active subscribers. We added new communication touchpoints, such as a sticky ribbon and a pop-up alert on the homepage, to increase awareness of the page and its contents. 

A few months later we expanded to our entire customer base and added a “le tue notizie” module at the end of the article. 

During this first period of tests, we collected some interesting results that confirmed our initial hypothesis and allowed us to keep evolving the project.  

Key results in a nutshell: 

• +40% time spent for visits vs. control. 
• +69% visits depth vs control. 
• +18% access frequency vs control. 
•  70% of users refreshed the page +50 times to find more news. 

Evolving the project

Our next steps involved balancing active and passive personalisation. To achieve this, we conducted an internal design thinking workshop to define how we could improve engagement KPI through our “Le Tue Notizie” section. We relied on user research insights and quanti-data to define the main user needs and criticalities to address. 

However, during our testing phase, we realised our users preferred a passive personalisation approach instead of an active one. We learned the recommended feed did not always fit the user’s interests, and users wanted to express their preferences and have more control over the news they received. 

To meet our readers’ needs, we added a new personalisation feature inside the “Le Tue Notizie” section that allows users to follow their favourite journalists. Many of our subscribers are fans of specific journalists, and it was crucial for us to incorporate this personalisation into our news platform. 

Our project aimed to suggest news to our users based on their interests while guaranteeing discovery and serendipity without limiting the user feed. We accomplished this by creating a new design with several “strips” where the user could easily access their recommended news feed and top news, add their favourite local news, and bookmark articles of their interest.

Through this continuous innovation and customer-driven approach, we aim to deliver personalised, engaging, and relevant news experiences to our readers.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .