In the book “Accelerate”, the authors Dr. Nicole Forsgren, Jez Humble and Gene Kim present the findings of rigorous scientific research to identify the factors that drive software delivery performance.

Measuring What Matters

Measuring software development productivity has always been difficult. Lines of code? More isn’t better and the metric encourages bad coding over elegant solutions. Utilization? Too high utilization leads to no time for improvement work.

“Accelerate” identifies four key metrics that actually predict organizational performance:

1. Delivery Lead Time: How quickly can you go from code committed to code running in production
2. Deployment Frequency: How often do you deploy code
3. Time to Restore Service: How quickly can you recover from failures
4. Change Fail Rate: What percentage of changes causes failures

High performing organizations generally excel in all four areas → There is no trade-off between speed and stability.

The book also focusses on continuous improvement rather than reaching an artificial final “mature” state. This creates a mindset of ongoing evolution and improvement.

The Technical Foundation of Excellence

Continuous delivery emerges as one of the cornerstones of high performing organizations. It is built on five key principles:

1. Building quality in from the start
2. Working in small batches for faster feedback
3. Automating repetitive tasks to free humans for complex work
4. Chase continuous improvement constantly
5. Creating shared responsibility across teams

The research shows a clear correlation between the following specific technical practices and performance:

• Version control (keeping application and system configuration in version control correlated even more with high performing teams than application code versioning)
• Reliable automated tests written by the developers themselves (flaky and unreliable tests have a negative effect and should be fixed or deleted)
• Trunk-based development with short-lived branches

High performing teams spend 49% of their time on new work compared to just 38% for low performers which spend more time in rework and unplanned tasks.

Architecture Matters More Than Technology

One of the books most interesting findings is that high performance is possible with nearly all technology stack or system types - from software development starting on the greenfield to legacy mainframe development. What matters isn’t what you build with, but how you build it.

Loosely coupled architectures that allow teams to test and deploy independently improve performance. Organizations with this architecture can actually scale engineering teams while increasing productivity rather than seeing it decline due to more communication and overhead.

The research also shows that teams should choose their own tools based on what works best for them, not through top-down standardization.

Management That Enables, Not Controls

“Accelerate” also shows that external approval processes decrease performance without improving stability. Peer reviews and selective approval only for high risk changes are the way to go.

Lean management practices also showed powerful results, but only when implemented as a complete system by:

• Limiting work in progress
• Creating visual displays of metrics / work status
• Using monitoring data for daily decisions

Lean product management and development form a virtuous cycle → delivery improvements allows working in smaller batches → allows to shorten the feedback loops → improves product decisions → enhances delivery.

Culture, Burnouts and Diversity

Deployment pain - the fear engineers feel when deploying code to production - strongly predicts poor performance. This pain comes from three sources: Software not designed for deployability, manual deployment steps, and handing deployment work off to other teams. All of these can be addressed through better practices and better design.

Organizations can fight employees burning out through creating a supportive no-blame culture, minimizing deployment pain investing in skills and improving delivery performance to create space for creative work.

High-performing teams also lead to more loyal employees who are 2.2 times more likely to recommend their organization to friends.

The data also shows that diverse teams (mainly gender-diversity was investigated) achieve higher collective intelligence and performance.

Leadership as a Multiplier

Transformational leadership amplifies all these effects. Leaders who provide vision, inspirational communication, supportive leadership and personal recognition fuel high performing teams.

Managers can support DevOps by investing in both technical practices and people, by providing accessible resources, dedicated training budgets, hackathons and time for experimentation.

Conclusion

“Accelerate” provides a scientific foundation for understanding what actually drives software delivery and organizational performance. The evidence-based approach leaves no place for opinions and reveals the capabilities that really matter.

It also provides a clear roadmap how to become a high performing team:

• Build technical capabilities around continuous delivery
• Create loosely coupled architectures
• Implement lean management in its entirety
• Empower teams with autonomy and authority

These practices lead to faster innovation, better products, engaged employees and ultimately better business performance.