Tips for Getting Started
"It does not happen all at once. There is no instant pudding." - W. Edwards Deming
Adopting entirely new workflows is rife with opportunities to fail. These are a few tips that may help you avoid common pitfalls when getting started.
Agility Helps#
Agile methodologies can be essential when implementing a culture of full-service ownership. Work is often unpredictable and agile methodologies can help teams stay the course toward long-term goals, even when the unexpected occurs. Agile-type workflows can help identify things that are going well and potential blockers. Adopting agile-type workflows is typically a part of most digital transformation initiatives.
Start Small#
Organizational transformation takes both significant effort and executive support. Gaining executive support begins with demonstrating value. A useful method for demonstrating that value with mitigated risk is to start small. Choose a non-critical production system and begin the process of defining your service. Next, measure a baseline of performance for your current production system. Allow your team the necessary time to learn the stepping stones they will use to master the mechanics of full-service ownership. One of the more immediate gains you should see is faster incident acknowledgement and, therefore, resolution.
Greenfield vs. Brownfield#
Greenfield projects eschew the past and start free from inherited technical and organizational debt. It's simpler to start fresh and design a world free from legacy restrictions. However, if your goal is to prove organizational value, it may better serve your long-term transformational initiative to select an achievable brownfield project. Here's a simple test: Are you likely to hear the phrase, "That's never going to work here," in your organization when proposing a new project? If so, consider selecting an achievable brownfield project to start with.
Blameless Culture#
Mistakes will be made. Digital transformation and full-service ownership are about learning new models and adapting them to the needs of your organization. Individuals need to be empowered to make decisions and experiment, while not fearing blame or retribution when making the wrong choice. A culture of psychological safety is one of the leading indicators that predict high-performance among software delivery teams. For in-depth tips on creating a culture of blamelessness, check out our postmortem guide.
Small Batch Change#
We mean it when we say start small. Another leading indicator of high performance for software delivery is trunk-based development. New features and code changes are written in short-lived branches, tested, and frequently deployed in small batches. Small batch changes serve multiple purposes, not the least of which are both increasing the ability for service owners to more quickly ship features to customers and making it easier to identify and mitigate problems that occur in production. The shift to a full-service ownership model is a great time for teams revamping their current workflow to start adopting continuous delivery practices.
Practice Makes Progress#
As with any new skill, practicing in low-risk settings builds confidence necessary when the real thing happens. Simulating incidents is a great way for teams new to owning production services to become accustomed to managing incidents when they occur. At PagerDuty, we've been practicing by way of Failure Fridays for several years. But you also don't need to actually break running services in production as a way to prepare for the real thing. Check out this podcast episode with Stitchfix's Bruce Wong for tips on ways to run tabletop exercises that help your teams build the skills necessary to manage incidents when they occur.