What is recommendation?
A recommendation is a suggestion made by a computer system that tells you which product, movie, song, article, or any other item you might like, based on information about you and other users.
Let's break it down
- Data collection: the system gathers information such as what you’ve clicked, bought, rated, or watched.
- User profile: it builds a simple picture of your tastes from that data.
- Item profile: it also creates a picture of each item’s characteristics.
- Matching algorithm: it compares your profile with item profiles to find good fits.
- Ranking: the matched items are ordered from most to least likely to interest you.
- Delivery: the final list is shown to you on a website or app.
Why does it matter?
Recommendations help you find things you’ll enjoy without searching through everything. For businesses, they increase sales, keep users on the platform longer, and make the experience feel personal.
Where is it used?
- Online stores (e.g., “Customers who bought this also bought…”)
- Video and music streaming services (e.g., “Because you watched X, you might like Y”)
- News feeds and social media (e.g., “Posts you may find interesting”)
- Advertising platforms (e.g., “Ads tailored to your interests”)
- Job portals (e.g., “Jobs matching your profile”)
Good things about it
- Saves time by filtering out irrelevant options.
- Helps discover new products, songs, or content you might never have found.
- Increases user satisfaction and loyalty.
- Boosts revenue for companies through higher conversion rates.
Not-so-good things
- Can create “filter bubbles” where you only see similar content and miss diverse viewpoints.
- Relies on personal data, raising privacy concerns.
- May inherit biases from the data, leading to unfair or skewed suggestions.
- Struggles with new users or items (“cold start” problem) because there’s little data to work with.