Shaun Abram
Technology and Leadership Blog
Book Summary: Accelerate
Accelerate: Building and Scaling High Performing Technology Organizations is a book by by Nicole Forsgren, Jez Humble and Gene Kim. It is a follow on from the State of DevOps Reports that Forsgren and Humble used to publish (and which I wrote about before in Development and delivery practices for team success). I highly recommend buying the book, but here are some chapter summaries for the highlights.
Contents
- 1 Foreword
- 2 Preface – The Research
- 3 Chapter 1 – Accelerate
- 4 Chapter 2 – Measuring Performance
- 5 Chapter 3 – Measuring and changing Culture
- 6 Chapter 4 – Technical Practices
- 7 Chapter 5 – Architecture
- 8 Chapter 6 – Integrating Infosec into the delivery cycle
- 9 Chapter 7 – Management practices for software
- 10 Chapter 8 – Product Development
- 11 Chapter 9 – Making Work Sustainable
- 12 Chapter 10 – Employee Satisfaction, Identity, and Engagement
- 13 Chapter 11 – Leaders and Managers
Foreword
The State of DevOps Report talks about techniques such as getting code into production faster & more frequently and finding & fixing issues sooner, all of which have been proven to correlate with strong business performance. This book follows those reports and anyone working in software should use the recommended techniques to improve, along with a program of continuous improvement. Adopting this guidance will make your organization higher performing.
Preface – The Research
Our research collected over 23,000 survey responses, from over 2,000 organizations. The research results show that software delivery needs to be an exercise in continuous improvement and that improvements in software delivery will have a very real impact on overall company performance such as profitability and market share.
Chapter 1 – Accelerate
Use small teams that work in short cycles and measure feedback from users. Focus on capabilities (outcome and improvements based), rather than antiquated maturity models. Instead of focusing on factors that don’t predict performance, such as age or types technology, or who performs deployments (ops or dev), focus on things that have been proven to predict performance such as frequency of code deployments, lead time from commit to deploy, mean time to recover from downtime, and change failure rates.
Chapter 2 – Measuring Performance
To measure the performance of software teams, do not focus on outputs or individual measures such as lines of code, velocity or utilization. Instead focus on outcomes and team or global measures such as lead time, deployment frequency, time to restore service, and change fail rate. Focusing on these can allow you to improve performance and achieve higher levels of stability and quality and, crucially, improve your organization’s performance in terms of profitability, market share, and productivity.
Chapter 3 – Measuring and changing Culture
Culture is of critical importance but can be intangible. Westrum developed a typology of organizational cultures in which organization cultures are classified as
– Pathological: power-oriented where information may be held for political reasons
– Bureaucratic: rule oriented where some departments may try to maintain their turf. or
– Generative: performance-oriented where everyone is focussed on a common mission
Culture can be measured using questions and surveys and that your classification predicts software delivery and organizational performance.
In a similar vein, while researching high performing teams, Google found that who is on the team is less important than how teams interact, and how organizations deal with failures is particularly instructive.
A consequence/predictor of good culture is trust and cooperation, better information being available for making decisions, and problems being more rapidly discovered and addressed.
Culture can be changed, by changing how people behave and what they do. The research shows that Lean management and continuous delivery have a real and meaningful impact on culture.
Chapter 4 – Technical Practices
Research shows that technical practices play a vital role in achieving successful Agile adoptions.
Continuous delivery is a set of capabilities that enable us to get changes into production safely, quickly, and sustainably. The five key principles are build quality in, work in small batches, automate, continuously improvement, collaborate across the org. To implement CD, we must create Configuration management, Continuous integration, and Continuous testing. Create multiple feedback loops to ensure that high-quality software gets delivered to users more frequently and more reliably.
Unplanned work and rework are useful proxies for quality because they represent a failure to build quality into our products. We found that continuous delivery predicts lower levels of unplanned work and rework in a statistically significant way.
The technical practices of continuous delivery have a huge impact on delivery performance, quality, and culture.
Chapter 5 – Architecture
High performance is possible with all kinds of systems, provided that the architecture, and the teams that build it, are loosely coupled. This means being able to change and deploy independently, and test without requiring an integrated environment. Teams should also be able to choose their own tools, and should build security into their work. Focus on engineers and outcomes, not tools or technologies.
Chapter 6 – Integrating Infosec into the delivery cycle
All the different functions within the software delivery value stream need to work together effectively, including Infosec. “Shift left” by having infosec involved with design, demos and testing throughout the software delivery process, instead of having security reviews as a separate downstream phase. Having standard Infosec libraries & tools can make it easier for developers to do the right thing too.
Chapter 7 – Management practices for software
Management practices derived from the Lean movement can drive software delivery performance. WIP limits can ensure that teams don’t become overburdened, but need to be combined with visual displays and a feedback loop back from production monitoring tools.
Chapter 8 – Product Development
Chapter TLDR; The four Lean practices of working in small batches, making flow of work visible, gather and using customer feedback, and team experimentation predict higher performance, as well as improving culture. Furthermore, there is a reciprocal model (a virtuous cycle) between increased delivery performance and Lean product management practices driving better outcomes for your organization.
Chapter 9 – Making Work Sustainable
Where code deployments are most painful, you’ll find the poorest software delivery performance, organizational performance, and culture. Both technical and Lean management practices, such as CICD, loosely coupled architectures, and supporting experimentation & learning all contribute to reductions in burnout.
Chapter 10 – Employee Satisfaction, Identity, and Engagement
Effective management practices combined with technical approaches, such as continuous delivery, impact performance and have a measurable effect on organizational culture. Automation (incl. tests and deployments) and proactive monitoring can improve engagement, which is not just a feel-good metric—it drives business outcomes.
Teams with more diversity achieve better business outcomes. It’s up to all of us to prioritize diversity and promote inclusive environments.
Chapter 11 – Leaders and Managers
Leadership has a powerful impact on results and is about inspiring and motivating those around you. Engaged leadership is essential for successful DevOps transformations. The 5 characteristics of a transformational leader are Vision, Inspirational communication, Intellectual stimulation, Supportive leadership, and Personal recognition. Managers can enable and invest in DevOps practices and & their employees’ professional development, take measures to make deployments less painful, and make performance metrics visible and monitoring a priority. Technology transformations benefit from truly engaged and transformational leaders who can support and amplify the work of their teams.
Tags: accelerate, books, cicd, devops, metrics, mttr, stateofdevops, summary