Will a Spotify for News Distribution Solution Work?

Question

Accepted Answer

Image source: Ever Wonder How Spotify Discover Weekly Works? Data Science, via Galvanize.Spotify's personalization approach is based on three recommendation components: 1) Collaborative filtering (having star ratings from users), 2) Natural Language Processing (NLP), and 3) Neural Networks (understanding tempo, loudness, time signature etc. to determine similar songs). There are two key factors to consider why news aggregators should potentially not adopt their model for news distribution. Firstly, Spotify still relies on having labeled data for its NLP which is why there are times when its recommendations aren't accurate. So, there's a huge cost in tagging and processing data in databases. Some of the news aggregators and apps in China and Asia, on the other hand, are using Dual Learning AI which provides a feedback loop and helps better personalize the content based on user searches and other applications.Secondly, content licensing and inventory building is something that Spotify relies upon in creating value for its end user. Netflix also does this and is only seeing the fruits of its labor after heavy investment in its own content and Open Connect (Content Delivery Network) in partnership with ISP's. Spotify doesn't appear to have a path to their own infrastructure or hardware that can develop usage habits and adoption.At the end of the day each market, as per the most recent table, shows different user preferences in news consumption. But with the promise of AI, the rise of mobile messaging apps and hardware from Asia, trends at this stage indicate the importance for publishers to partner with mobile news aggregators with a growing content personalization stack for better audience reach.

Share of respondents
Direct entry	32%
Search	25%
Social media	23%
E-mail	6%
Mobile alerts	5%
Aggregators	5%

Would a Spotify for news distribution model work?

Vahe Arabian

Vahe Arabian

Vahe Arabian

How news is currently being consumed