This post is part 3 of a 4 part-series. You can start over or read the introduction in this post.
Here are all the posts in this series.
- Software Delivery Performance
- Organizational Culture
- Continuous Delivery (this post)
- Lean Management
Part 3 - Continuous Delivery
From the last post about Organizational Culture, I concluded that 2 practices, Continuous Delivery, and Lean Management, can impact and change the culture. In this post, I will talk about Continuous Delivery and its impacts.
Figure 1 - Continuous Delivery and its impacts
What is Continuous Delivery
Let’s start with the definition of Continuous Delivery. Technically, you might know this term as CD, however, that’s not the whole story. The core of Continuous Delivery is not about releasing software to production continuously, but it is a mindset, culture, and practices that enabling us to release the software to customers continuously which includes:
Built-in quality. Thinking about quality in every process from start to end, not only do the testing after the product was finished.
Work in small batches. We don’t want too many incomplete works waiting for the next person to finish it. What we need is the work that flows from start to end as fast as possible.
Automation. repeated tasks must be done by computer not human.
Continuous improvement. We always need a new way of work in order to release a high-quality product faster and faster.
Team responsibility. Mistakes, failure, or even success should be owned by the team, not any individual.
Capabilities needed for Continuous Delivery
In the book, they discovered the key capabilities that drive continuous delivery in practice.
Version control. Everything related to the product from coding, infrastructure to configurations should be versioned and managed using a tool like Git.
Test automation. You need reliable test automation at all levels; unit test, API test, UI test, and many more to give the team the confidence to make a change. Also important, make sure that the testing codes are maintainable and not overload the team.
Test data management. Test automation needs data. If the test data is not reliable, tests won’t be reliable either.
Trunk-based development. Trunk-based means a product code lives mainly in one branch most of the time. The research from the book found that this approach is better than the long-lived feature branch-based approach. I absolutely agree with this based on my own experience. Merging from branch to branch is risky and error-prone and adds more friction to the delivery process.
Continuous integration. The process from code committed to product release should be automated using the CI/CD tool. It needs to also include automated testing and auditing as part of the process.
Built-in security. Security is more and more important these days but has been treated as an afterthought. Fixing the security later costs too much and blocks the flow of work. The team should think about it and built it into the system early from the beginning.
Loosely Coupled Architecture. Meaning that the structure of the system should have as little dependency as possible. Product components and teams should be designed in a way that only necessary communications are needed between components and teams.
Empowered Teams. The team should be able to decide their own path, the way of work, tools, or anything they need to be successful. The success of the team must be aligned with organization goals.
Monitoring. Important product indicators should be ready for the team to monitor or act proactively. Metrics include product-oriented such as error rate, performance, auditing, customer-oriented such reported issues, usage, and revenue.
Impacts of Continuous Delivery
The Continuous Delivery capabilities above are used to correlate with other measurements. What they found is that Continuous Delivery could directly impact Organizational Culture, Software Delivery Performance, Team member (see figure 1).
The impact on team member are less deployment pain, less burnout, bond with the organization, and more job satisfaction.
Conclusion
From this analysis, you can see that Continuous Delivery plays an important role in high-performance organizations. Capabilities or practices described above can be a clear guideline on what you can do to move your organization forward. Next, we will talk about another important factor that is also important, if not more, Lean Management.