
The Science of Lean Software and DevOps: Building and Scaling High Performing Technology Organizations
Nicole Forsgren, Jez Humble, and Gene Kim · 2018 · Technology
Original summary · AI-drafted, human-published · added by Library
Drawing on four years of survey data from thousands of technology professionals, the authors argue that software delivery performance can be measured rigorously and that a specific set of technical, process, and cultural capabilities causally drives it. They show that speed and stability are not opposing goals, and that firms which excel at both outperform competitors on profitability, market share, and employee wellbeing. The book replaced folklore about IT performance with statistically validated evidence.
Pick a finish date and Genius lays out the days — the plan shows today's target and keeps you honest.
Start a circle and share the code — everyone sees everyone's honest place in the book. Accountability, not leaderboards.
- Engineering leaders who want evidence, not opinion, for how to invest scarce improvement effort - Product and operations executives deciding whether DevOps investment actually pays off financially - Researchers and consultants who want a rigorous methodology for studying organizational and technical performance together
Technology performance can be quantified with a small, validated set of metrics, and organizations that refuse to measure it are managing blind.
Software delivery performance reduces to just four numbers: deployment frequency, lead time for changes, mean time to restore service, and change failure rate.
Speed and stability move together rather than trading off against each other, so slowing down deliberately to be safer is usually a mistake.
A specific bundle of technical practices, especially trunk-based development and automated testing, is what makes frequent low-risk deployment possible, not heroics or better people.
Loosely coupled, well-encapsulated system architecture is a prerequisite for high delivery performance, because tightly coupled systems force coordination overhead that no amount of process improvement can fix.
Working in small batches and limiting work in progress, borrowed from lean manufacturing, is what actually enables fast feedback in software, not simply asking teams to work faster.
An organization's culture, specifically whether information flows freely and failure is treated as a learning opportunity rather than a blame target, predicts delivery performance as strongly as its technical practices do.
Transformational leadership, specifically leaders who articulate vision and provide intellectual stimulation, sets the ceiling for how far technical and cultural improvements can go.
High software delivery performance causally improves organizational profitability, market share, and productivity, and also reduces employee burnout, making the earlier capabilities a business and human imperative, not just an engineering preference.
Nicole Forsgren, PhD, co-founded DORA and led the State of DevOps research before roles at Google and Microsoft Research. Jez Humble co-wrote the earlier book Continuous Delivery and worked as a software engineer and consultant across large-scale systems. Gene Kim founded the security firm Tripwire and wrote The Phoenix Project, and runs IT Revolution, which studies high-performing technology organizations.