Algorithmic curation is now the most important factor when it comes to personalization and recommendations, and even more so when it comes to news websites. A recommendation system pretty much does a librarian’s job: if the librarian knows you well, they know what book to recommend based on your preferences. If they don’t, then they usually recommend the top selling book or the next crowd pleaser.
News organizations publish a huge amount of content daily, and filtering this content is of the utmost importance to make sure they can direct users to the most relevant article, to make sure they stay longer and consume more content. Therefore, recommendations are crucial to audience engagement.
The traditional approach for news websites has been the use of tags to categorize content. Editors would add tags manually to each article and then the system would recommend other articles with similar tags. But this approach has failed over time mainly because it is too time consuming, lacks in consistency, and it is subjective. Automated tagging has been a solution to tagging articles seamlessly and consistently, all while saving editors’ time and improving accuracy. But are tags enough to really grasp what the reader is interested in?
The advantage of topic based recommendations is that you don’t need information about other users since the recommendations are specific to this user based on the type of content they are consuming. The model can capture the specific interests of a user, and can recommend niche items that very few other users are interested in.
When the system is limited to recommending the content of the same type as the user is already using, the value from the recommendation system is significantly less when other content types from other services can be recommended. For example, recommending news articles based on browsing of news is useful, but wouldn’t it be much more useful when music, videos from different services can be recommended based on the news browsing.
A more sophisticated approach involves going beyond tags and content. Instead, natural language processing has been used to create a more in-depth understanding of both content and the reader’s profile.
For example, when it comes to personalized recommendations, Magnet profiles users based on the content (video, articles, photo galleries, etc) they consume overtime and accordingly makes contextual and personalized recommendations using 16 algorithms. These algorithms cover the trending aspect, collaborative learning techniques, long-term vs. short-term aspect, etc to name a few.
The result is increased loyalty and satisfaction of your users with your web services. Typically, users also interact with more items and this behavior leads to increased consumption and higher profits.
A good recommendation engine should be in a position to learn, adapt and deliver the best recommendation always. The strength of Magnet’s personalized recommendations lies in the fact that it understands the “aboutness” of text, is able to categorize, recognize named entities and extract key topics. All of these factors are scored in terms of relatedness to offer the most relevant recommendations in the most seamless way possible.
Recently, our team has announced the official launch of a new service that allows Publishers to affect what is returned to their visitors when using Magnet Personalized Recommendations: Magnet Rules Engine. Get in touch to see what Magnet by Klangoo can do for your news website.