Introduction

Netflix, founded by Reed Hastings and Marc Randolph in 1997, allows users to stream movies and TV shows. As of Q2 2024, Netflix had 277.65 million paid subscribers globally, an increase of over eight million from the previous quarter (Netflix, 2023). Operating in over 190 countries, it’s one of the most accessible streaming platforms, with international markets contributing over 65% of its user base and more than 60% of its revenue (KPMG, 2023).

Netflix leverages data for personalization, content discovery, user retention, and content production. Its recommendation algorithm enhances user engagement by offering tailored content, which is one of Netflix's key competitive advantages (Netflix, 2023).

Data Requirement(s)

To improve its personalized recommendations, Netflix needs to collect and analyze data from both internal and external sources. This data can be divided into two categories: internal data, and external data.

Internal Data

Internal data refers to information that Netflix collects from its platform and users. This data is critical for analyzing customer behavior, preferences, and engagement patterns, helping Netflix make data-driven decisions that improve personalization, content recommendations, and business strategies.

User Data

It collects user profile information, such as name, email address, payment details, location, viewing preferences, and account settings. This data is gathered during user sign-up and stored in their profile.

Viewing History

This includes the TV shows, movies, and genres they’ve watched, the time spent watching each title, and how often they engage with the platform.

External Data

In addition to internal data, Netflix also gathers and utilizes external data from various sources to enhance its platform and personalized recommendations. External data allows Netflix to gain deeper insights into user preferences, content trends, and global market behaviors, complementing its internal data sources.

Demographic Data

Netflix collects demographic information, including age, gender, income level, and geographic location, from external sources like third-party data providers and public records.

By combining internal and external data, Netflix refines its recommendation algorithms, provides relevant content, and improves the user experience. External data enhances Netflix’s ability to offer personalized suggestions, track global trends, and optimize its marketing strategies. The platform adheres to strict data privacy regulations, ensuring user data is handled securely and with transparency (Netflix Privacy Policy, 2023).

Data Warehouse

Netflix uses both internal and external data to enhance its recommendation system and user experience. Internally, it collects data like user profiles, viewing history, and content interaction, while externally, it gathers demographic and behavioral data from partners.

Netflix stores this data on Amazon Web Services (AWS), utilizing a scalable data warehouse to manage both structured and unstructured data. This infrastructure supports recommendation algorithms and helps data scientists analyze user behavior, optimize recommendations, and make strategic decisions to maintain Netflix’s competitive edge in the streaming market.

Data Warehouse Schema

Netflix uses the Snowflake schema, a multi-dimensional data model, to enhance personalized recommendations. This schema is ideal for supporting complex queries related to user viewing habits, content performance, and regional preferences. As an extension of the star schema, the Snowflake schema normalizes dimension tables into related tables, improving storage efficiency and query performance (Smith, 2023; Snowflake schema, 2023). This architecture allows Netflix to offer personalized content recommendations based on viewing history, genres, location, and device behavior.

ETL Process

The ETL (Extract, Transform, Load) process is fundamental to Netflix’s personalized recommendation system (What is ETL? - Extract Transform Load Explained - AWS, no date; Rouse, 2023). It allows Netflix to extract, process, and integrate data from multiple sources into a data warehouse, ensuring data quality and consistency for accurate recommendations.

OLAP (Online Analytical Processing)

OLAP systems allow Netflix to analyze large amounts of data from multiple perspectives or dimensions, helping the platform to make better decisions about content recommendations. The system enables Netflix to quickly aggregate and analyze user data, viewing history, genres, actors, and more, enabling the platform to generate complex queries that lead to more personalized recommendations.

Big Data

Netflix extensively uses big data to improve its content recommendation system and user experience by focusing on volume, velocity, and variety.

 Volume: Netflix gathers large amounts of data, such as viewing history, search queries, and device usage.

 Velocity: It processes data in real-time, continuously updating recommendations based on new behaviors.

 Variety: Data comes from various sources, including structured data (user demographics), semi-structured data (search logs), and unstructured data (movie synopses).

By analyzing user history, contextual factors, and social interactions, Netflix creates highly personalized recommendations, introducing users to new content and genres.

Limitations of Data Warehouse for Handling Big Data

Data warehouses are ideal for structured data, like user demographics and viewing history, which Netflix utilizes. However, they struggle with unstructured or semi-structured data, such as video metadata and user reviews, due to irregular formats and scalability challenges. Additionally, traditional data warehouses are not optimized for real-time processing, which Netflix's recommendation engine requires. Data silos and limited flexibility in analysis further hinder their ability to provide deeper insights. While on-premises setups can be costly, Netflix mitigates these limitations by using advanced big data technologies like Apache Kafka and Hadoop to efficiently manage its vast data and personalize recommendations.

Hadoop

Data warehouses are effective for structured data like user demographics and viewing history, which Netflix uses. However, they face challenges with unstructured or semi-structured data, such as video metadata and user reviews, due to scalability issues and irregular formats. Traditional data warehouses also struggle with real-time processing, which is essential for Netflix's recommendation engine. Additionally, data silos and limited analytical flexibility restrict deeper insights. While on-premises systems are expensive, Netflix overcomes these limitations by leveraging big data technologies like Apache Kafka and Hadoop to manage vast data and personalize recommendations (Bennett, 2023; Ghosh, 2023).

Cloud Computing

Cloud computing enables users to access remote computer resources managed by a cloud services provider (CSP) on a subscription or usage-based model. It offers virtualized IT infrastructure like servers and networking, allowing resource division across hardware boundaries (Ian Smalley, 2024; What is cloud computing? | IBM, no date). Various industries, such as healthcare and finance, use cloud computing for personalized treatments and fraud detection (What Is Cloud Computing? | Microsoft Azure, no date; what-is-cloud-computing - AWS, no date). Businesses benefit from cloud computing through cost savings, scalability, agility, reliability, security, and innovation. It eliminates large hardware investments, enables rapid deployment of services, and promotes access to new technologies (Stephanie Susnjara, Ian Smalley 2024).

Pros and Cons of Using Cloud for Data Warehouse and Big Data Systems

Pros

Scalability and elasticity

Netflix can leverage cloud-based data warehouses like Amazon Redshift or Google Big Query to scale its infrastructure up or down, depending on demand. This is just one they are many more.

Cons

Cost

While cloud services can be cost-effective, unexpected spikes in usage (due to high data loads or increased viewership) could lead to higher expenses for Netflix, making cost management crucial.

Conclusion

Netflix leverages data to deliver highly personalized content recommendations, using a comprehensive strategy that integrates data from user history, demographics, and social interactions. With a robust ETL process and advanced OLAP operations, Netflix gains meaningful insights to segment users and refine recommendations through machine learning and filtering techniques.

By adopting a hybrid data strategy, Netflix combines the strengths of data warehouses and big data systems, analyzing both historical and real-time data for dynamic content suggestions. This approach enhances user experience, helps in content acquisition, and supports scalability, ensuring long-term viewer satisfaction.