Scaling beyond one: How Airbnb evolved its data architecture for a multi-product world

2026-06-09 · Source: The Airbnb Tech Blog - Medium · Field: Technology & Digital — Data Science & Analytics, Software Development & Engineering, Cloud Computing & IT Infrastructure · Depth: Intermediate, long

Summary

Airbnb evolved its decade-old offline data architecture to support its May 2025 Summer Release, which introduced a redesigned app, relaunched Experiences, and debuted Services, moving beyond its traditional Homes focus. Data engineers and analytics engineers developed a consistent and flexible data modeling framework to integrate these three product pillars. This involved addressing the core dilemma of separate versus monolithic data models, ultimately balancing centralized principles with decentralized guidelines. Three foundational principles were established: no hybrid data models, consistent identifier naming (e.g., id_experience vs. id_product_listing with dim_product_type), and clear namespace organization. Modeling guidelines, considering shared attributes, future evolution, and downstream consumers, helped teams decide between separate models for product-specific logic (like Listings, Availability, Location, Guests) and monolithic models for cross-cutting concepts (like Messaging, Payments, Customer Support). The initiative also navigated challenges of translating raw online data for analytics and managing legacy data debt.

Key takeaway

For Data Engineers or Analytics Engineers building multi-product data platforms, you should adopt a hybrid strategy that combines central architectural principles with domain-specific modeling guidelines. This approach allows you to maintain consistency for core elements while providing flexibility for unique product attributes, preventing data silos and technical debt. Prioritize clear identifier conventions and namespace organization to ensure scalability and ease of understanding for downstream consumers, and plan for continuous migration of legacy assets.

Key insights

Balancing centralized data consistency with domain-specific flexibility is crucial for scalable multi-product data architectures.

Principles

• Avoid hybrid data models; choose fully separate or monolithic.
• Enforce consistent identifier naming based on modeling choice (e.g., id_experience, id_service).
• Organize data with clear namespaces for product-specific and global tables.

Method

Establish foundational principles and provide modeling guidelines to empower domain teams in choosing between separate or monolithic data models based on product attribute commonality and future scalability.

In practice

• Use separate models for distinct product attributes (e.g., Service offerings, flexible availability).
• Apply monolithic models for cross-cutting concepts (e.g., messaging, payments, customer support).
• Treat the offline data warehouse as a crucial translation layer for online production data.

Topics

• Data Architecture
• Data Modeling
• Offline Data Warehouse
• Multi-Product Strategy
• Airbnb
• Data Governance

Best for: Data Engineer, Analytics Engineer, Director of AI/ML

Related on AIssential

• Stop building data products. Start building data services. — Traditional data product models hinder rapid growth and AI agent consumption; shift to open, governed data services.
• Data Mesh, Data Platform, or Data Chaos? Choosing the Right Ownership Model (Data #3) — Effective data ownership scales through intentional hybrid structures, not by choosing between centralized or decentralized models.
• Article: Lakehouse Tower of Babel: Handling Identifier Resolution Rules Across Database Engines — Inconsistent SQL identifier resolution across lakehouse engines and catalogs creates significant interoperability challenges.
• Multimodal Data Integration: Production Architectures for Healthcare AI — Multimodal data integration in healthcare requires a unified, governed lakehouse architecture to overcome production challenges.
• Your Data Layer Used to Hide Behind Your Product. Now It Is the Product. With Firebolt’s CEO — The data layer is shifting from a hidden backend to the primary product interface, driven by agent-centric interactions and deployment demands.
• Data Strategy and Data Governance, reimagined. — Effective data management requires distinct, interconnected layers for strategy, governance, and operational execution.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Airbnb Tech Blog - Medium.