Amazon.com, Inc.

In the late 1990s, Amazon was a fast-growing internet retailer navigating the transition from a tightly coordinated startup to a company that would eventually employ hundreds of thousands of people. The core challenge was coordination overhead: as teams grew, communication costs grew quadratically. Every additional person added to a team required more meetings, more alignment, and slower decisions. By the early 2000s, Amazon had begun operating more like a traditional large company — with functional silos, slow approvals, and innovation bottlenecks. Internal tools and services were being built redundantly by different teams with no shared interfaces. Engineering releases required coordination across teams, creating dependencies that slowed shipping velocity. Jeff Bezos observed that teams were spending more time coordinating with each other than shipping product.

Jeff Bezos instituted the two-pizza rule as a structural mandate: any team that required more than two pizzas to feed was too large and needed to be split. The rule was operationalized through a broader organizational architecture:

• Team size limit: Teams capped at 6–10 people, including engineers, product manager, and operations — small enough to share clear ownership and ship independently.
• Single-threaded ownership: Each team was assigned one and only one business goal, eliminating divided attention. The team owned its metric, its roadmap, and its technology — with no shared responsibilities across other teams.
• Service-oriented architecture mandate (API mandate, 2002): Bezos issued an internal memo requiring all teams to expose their data and functionality through APIs — no direct database access, no shared libraries, no backdoors. This forced clean interfaces between teams and made each team's work consumable by others without coordination.
• Autonomy over consensus: Two-pizza teams were empowered to make architectural decisions, set their own sprints, and deploy independently — eliminating the need for cross-team alignment on most decisions.
• Separation of data ownership: Each team owned its own data store; other teams accessed it only through the team's published API.