Ranked feeds use metadata at the top level of an activity to reorganize activities in a feed based on your algorithm.
- Ranking algorithms need to use data that is at the top level of an activity, it cannot be nested.
- Ranking algorithms are unable to know what a user has previously interacted with. In fact, they are unable to take into account any specific user data - they only use data stored in an activity.
We are frequently asked how to create personalized feeds based on what a user is interested in. Although a "personalizaed" experience can be created through a user following topics (Feeds) they are interested in, we cannot take into account their specific engagement data when using ranking. Our Personalization product exists for solving this problem as it stores a database of user data such as engagements with previous content, impressions, and can access a graph of the users friends and followers to suggest content outside of what they follow.