---
title: "Design: a recommendation system"
description: "Design recommendations for an article app: find candidates, score and pick a list, handle new readers and items, tie feedback to an offer, and stay fast."
canonical_url: "https://fanout.sh/system/archive/recommendation-system-design"
md_url: "https://fanout.sh/system/archive/recommendation-system-design.md"
last_updated: "2026-11-11"
access: "pro-preview"
---

# Design: a recommendation system

Design recommendations for an article app: find candidates, score and pick a list, handle new readers and items, tie feedback to an offer, and stay fast.

- Access: Fanout Pro (public preview)

## The core idea

A recommendation service selects useful items before the reader asks for a particular one.

For an article app, that means choosing what belongs on the home page from the reader's interests, the available catalog and recent feedback.

## In this lesson

- What the home page promises
- Finding enough useful candidates
- Scoring an item and choosing a list
- A small implementation we can inspect
- Following Ada's request
- New readers, new items and short lists
- Feedback belongs to a particular offer
- Replacing prepared recommendations
- Keeping the service useful under load
- Learning from later observations
- Run the lifecycle

## Continue reading

Pro unlocks video lessons, visual walkthroughs, full notes, diagrams, case studies, and implementations across the System Design archive.

---
This representation contains public Fanout content only. Protected Pro lessons, account data, billing, checkout, and pricing are not included.

Browse the public content map: https://fanout.sh/sitemap.md
