Abstract: One of the primary drivers, if not THE central driver, behind any Enterprise DevOps transformation is the organizations need to optimize the flow of business value (in the form of incremental software) between developers and end-users. Within many organizations, and particularly within larger enterprise organizations, lack of trust between organizational stakeholders tends to sub-optimize the flow of value.
This technology agnostic talk will explore how organizations such as Aflac, Boeing, BNSF Railways, Service Master and Fanatics were able to dramatically accelerate evolution from low-trust to high-trust cultures. The common thread between these organizations is how they leverage data-driven instrumentation to dispel myths, break down organizational/political barriers and guide trust-building transformation.
What is DevOps Performance Measurement?
- Value Stream Performance vs Operational Performance - Precisely measuring the DevOps machine itself and not it’s outputs
- How to describe DevOps performance using measures the entire organization truly cares about:
- The Flow of Value
- Delivery Risk
- Process Compliance
- Rework
Data is Truth: Why DevOps Performance Measurement is so critical to building trust across the enterprise?
- How objective data can diffuse and ultimately eliminate the blame game
- Removing corporate politics with the ultimate equal opportunity enabler
- The relationship between Batch size and Trust
- Leveraging objective performance insights to foster organizational creativity
- Using metrics to create shared incentives and common agenda’s
- Quantitative results speak best to executive leadership
It all starts with Value Stream Mapping
- What is The DevOps Unit of Flow?
- Phases, Activities and Controls - DevOps specific templates and guidelines to simplify value stream mapping
- Manual vs automated activities
Top Ten DevOps Performance Metrics That Raise Organizational Trust (including…)
- Activity duration and DevOps Wait-Time
- Value Stream Bottleneck Analysis
- Release Candidate and/or Feature Risk Analysis
- Waste and Rework Metrics
- Quality Assurance Effectivness Index
- Measuring Code Stability/Complexity by Feature or Release
- Real-time compliance measurement
Some Lessons Learned:
- Choose Measures Wisely: Trust can only be gained when adopting DevOps performance metrics that are organizational strategic and highly valued by business stakeholders. Stay high level and business focused.
- Operational metrics such as broken-build percentage, deployment frequency and test coverage can lead to sub-optimal behavior and a reduction in trust.
- The initial performance baseline is not good or bad - it is the foundation for future improvement.
- Warning: The data must guide investment in people, process and tools - adopting a course of investment despite the data can nullify the entire effort.
Note: This presentation is vender and technology independent. Our findings are based on direct experience gathered from over a dozen performance measurement engagements with enterprise sized customers.
Learning Outcomes: - How to effectively map value streams in the context of DevOps
- How to distinguish between metrics and measures that build trust and those that erode trust
- Technology agnostic approaches to tracking business value (in the form of incremental software capabilities) thru the delivery value stream
- How performance metrics can help identify DevOps waste and sub-optimization
- How to use common DevOps data to objectively measure delivery risk - before software is released
Attachments:
