An Elegant Puzzle, in detail
An Elegant Puzzle is Will Larson's guide to the craft of engineering management, written from his experience leading engineering teams at Digg, Uber, and Stripe. The title reflects the book's central sensibility: most management problems are well-defined puzzles with multiple valid solutions, and the manager's job is to understand the constraints well enough to find the most elegant one.
The book is organized around the actual challenges engineering managers face rather than around inspirational principles. Larson writes about team sizing and dynamics, the difference between a team that's slowing down because it's too small and one that's slowing down because it's carrying technical debt, and why these require completely different interventions. He offers a model of team states — falling behind, treading water, repaying debt, innovating — and argues that misdiagnosing the state leads managers to apply the wrong fix.
A significant portion of the book covers organizational design at scale: how to structure teams to reduce coordination costs, how to think about platforms versus product teams, how to handle the political dynamics of growing an engineering organization without losing engineering quality. Larson is unusually frank about the gap between what organizations say they value and what they actually reward.
The final section addresses personal effectiveness: how to manage a large team without becoming a bottleneck, how to write well enough that your decisions travel through the organization without you in the room, and how to build a career that survives company transitions. The book is more useful for engineering managers than for general management, but the thinking on organizational systems applies more broadly. Larson writes in a dense, essay-like style that rewards slow reading and re-reading over the weekend before a difficult quarter.
The big ideas
- 1.
Teams move through four states: falling behind, treading water, repaying debt, and innovating. Each state requires a different intervention, and misdiagnosing the state leads to the wrong fix.
- 2.
Hiring into a team that's falling behind often slows it down further before it speeds up. Adding headcount doesn't solve problems that stem from technical debt or unclear ownership.
- 3.
Organizational design is about reducing coordination costs. The fewer dependencies a team has, the more it can ship. Structure should serve shipping, not org chart aesthetics.