A user-to-item recommendation selects catalog items for a particular user using information such as their profile, history, and current context. Different users may therefore see different items. section 1.2.1
The recommendation seed shapes how personalization is applied:
- User-seeded recommendations use user data to find candidates; item-seeded recommendations start from a specific item and find related items. section 1.2
- A popular-content list can also be personalized: a system may start with popular items and reorder them according to a user’s tastes. section 1.2.1
Personalizing popular content
Netflix’s “Popular on Netflix” carousel can show different selections or ordering for different profiles, reflecting their viewing histories. section 1.2.1
Figure 1.1 A typical e-commerce landing page showing two types of recommendations: personalized ("Recommended for you") and non-personalized ("Popular Products"). The same page uses different recommendation strategies simultaneously.

Compare personalized “Recommended for you” results with a non-personalized “Popular Products” list on the same page. section 1.2

