Welcome to the first part of this News Personalization Series. In this first part of the series, our Account Manager, Thomas Steisel will dive into the five main reasons why personalizing news is so challenging.

Introduction‍

Model changing events are unpredictable news that has a direct and violent impact on your interest in a certain topic. Even though you might not be interested in that topic, the news will grab your attention. In this first episode, Thomas Steisel introduces you to the concept of "trendingness", a feature that predicts these changing events.

The way users consume movies is completely different from the way they consume news. News needs to provide relevant content in less time and less digital space. Learn how to face the challenge of real-time news recommendations.

The filter bubbles are known in the news industry for isolating the readers’ opinions. What happens is that the algorithm only shows the reader the same kind of article based on their behavior and therefore, reinforces their own opinion. To solve this problem, you need to feed balanced pro and counter-opinion topics. In this video, we will explain what a filter bubble is and how to overcome it.

The combination of the Editor-In-Chief efforts and AI support is possible when implementing a hybrid personalization in your newsroom. But first, you should first put into consideration other aspects that Thomas mentions in the following video.

The way a reader gives us feedback about the articles he likes, it’s not comparable to a movie, which they can rate. A "like" is a clear proof of satisfaction in the entertainment industry. In this video, we introduce you to the term “impression discounting”, and how we use it to give “implicit feedback” when personalizing news.

News is harder than Netflix
News personalization is on the roadmap, but progress is going slow? That’s because the news industry is harder than personalizing platforms like Netflix.
I get it. You’ve evaluated so many personalization solutions. You’ve chosen and installed the one that promised 30% CTR increased and 2-minute integration time. You’ve set it up to your CMS.
Or perhaps - You’ve hired a team of data scientists to develop your personalization in-house... for the last few years.
You’ve ticked all the boxes on the ‘personalization roadmap’, yet somehow you’re still struggling to show the most relevant article to the right reader - like a total beginner.
Your loyal subscriber-base isn’t increasing.
The reason why your personalization has low impact is because it wasn’t designed for publishers.
You’ve probably selected a great solution for e-commerce, but not one that’s designed to solve publishers’ complex problems.
Here are some specific challenges that news brands are facing when personalizing news.

News catalog is changing rapidly
While Netflix has days, weeks, months, or sometimes years to learn user behaviour - you have only seconds. News is instant and most articles become irrelevant in a few hours. You need to be able to correctly recommend articles within minutes after publication.


The blogpost continues here.

In this new video, Anton Smet, identifies three factors you should consider when building your personalization strategy. By implementing a conversion module, you will gain a new idea of what a recommender system can do. It goes beyond just recommending interesting articles but also increases subscriptions to your newspaper.

Regarding the previous video, Froomle has created a module that will help you increase your subscriptions. This module will make you increase revenue from subscriptions and turn your newsroom into a subscription machine. How? Showing to your non-paying readers premium articles that will convice them to subscribe.

In the last episode of the series, we put into consideration the possible reason why news personalization isn't working. A big part of the problem is the high expectations of the tech giants, who are personalizing to such a high level. In order to overcome this issue, we recommend you apply the advice that Anton mentions in this video.

Watch our CEO, Koen Verstrepen, talk to Pieter De Smet from Mediahuis about how they are making a difference by increasing relevance.

In this case study, learn how Mediahuis is using a self-learning recommendation engine to increase reader loyalty, deliver a better news reading experience and build relevance at scale.