Platforms such as Netflix, Spotify, Facebook or Amazon use personalized recommendation systems based on the user’s activity. Recommender systems process all the information related to users’ online activity: their preferences, their interests, the things they purchase, the content they consume… in order to show them personalized advertising or recommendations on specific news or products.


Nowadays, the variety of online products and services is so wide that recommender systems have become the key to success:

  • They improve customer satisfaction by facilitating their search and making them discover new products.
  • This customer satisfaction translates into increased time spent on the platform and increased sales.


Recommender systems use different types of machine learning algorithms:

  • Collaborative filtering, which collect information from multiple users from various sources and establish a correlation of preferences.
  • Content-based filtering, which identifies certain features in a product or in users’ preferences, to make recommendations that include those same features.
  • Demographic filtering, which base their recommendations on the specific interests of a certain geographic region.
  • Hybrid filtering, which combine different types of filters.


Recommender systems have a great potential growth and will become increasingly sophisticated. An example of this is Netflix, which shows different thumbnails for the same film or series based on the user’s profile. The goal is to catch your attention by showing you a thumbnail with your favorite character or with an action scene.