How does TikTok know exactly what videos will keep you scrolling for hours? How does Amazon suggest products you didn't even know you wanted? How does Netflix queue up shows that match your taste perfectly?

Behind all these experiences is a recommendation system, and it's one of the most impactful patterns in modern system design. These systems drive engagement, revenue, and user satisfaction across virtually every major platform. Get them right, and users stay glued to your product. Get them wrong, and they leave.

This pattern shows up constantly in system design interviews, whether you're designing TikTok, Netflix, Amazon, a dating app, or a news feed. Interviewers want to see that you understand how to generate relevant recommendations at scale, how to balance personalization with freshness, and how to handle the tricky cold start problem when you have no data about a new user.

In this chapter, we'll explore the recommendation pattern in depth. We'll look at different algorithms, understand their trade-offs, and learn how to design recommendation systems for various use cases.

What are Recommendations?