14
Implementing Agile Release Trains
“Train everyone and launch trains.”
—SAFe advice
Introduction
In the previous two chapters of this part of the book, we described steps 1–6 of the implementation roadmap. In this chapter, we’ll cover the next three steps, which together is arguably the most important part of the transformation, the part where the enterprise implements Agile Release Trains (ARTs). The steps described in this chapter include the following:
Step 7: Prepare for the ART launch.
Step 8: Train teams and launch the ART.
Step 9: Coach ART execution.
In addition, we’ll describe an accelerated, one-week ‘quick-start’ approach to the ART launch, which—after some preparation—is the fastest way for an ART to begin delivering value.
Step 7: Prepare for ART Launch
By now, the enterprise will have identified their value streams and established an implementation plan. It will also have loosely defined the first ART. This is a pivotal moment, as plans are now moving toward implementation. From a change-management perspective, the first ART launch is critical. This will be the first major change to the way of working and will generate the initial short-term wins that help build momentum.
SAFe Program Consultants (SPCs) often lead the implementation of the initial ARTs, supported by SAFe-trained ART stakeholders and members of the Lean-Agile Center of Excellence (LACE). Whoever leads the preparation of the ART launch facilitates the activities described in the following sections.
Defining the ART
In Chapter 13, Designing the Implementation, we described the process for defining the first value stream and the ART. At that stage of planning, the ART is expressed with just enough detail to determine that it’s a potential ART. However, its details and boundaries are left to those who better understand the local context. An ART canvas1 (Figure 14-1) offers a template for this definition.
1. Thanks to SPCT Mark Richards for the ART canvas inspiration.
Figure 14-1. Agile Release Train canvas
A key goal of the ART canvas is to help teams identify the principal ART roles, the vision for the train, what solutions it provides, who the customers are, and more. We know from systems thinking that the people, management, and processes of the organization that builds the solution are also a system. By that definition, the ART itself is also a system. For this system to function properly, the responsibilities of system definition, building, validation, and deployment have to be realized. Filling in the essential roles on the ART canvas fosters these discussions and highlights the new responsibilities.
Setting the Launch Date and Program Cadence
With the ART definition in hand, the next step is to set a date for the first PI planning event. This creates a forcing function, a ‘date-certain’ deadline for the launch, which will create a starting point and define the planning timeline.
A development cadence must also be established, including the PI and iteration lengths. Although the SAFe Big Picture shows a ten-week PI, consisting of four regular iterations and one Innovation and Planning (IP) iteration, there is no fixed rule for the PI cadence, nor for how much time should be reserved for the IP iteration.
The recommended duration of a PI is between eight to twelve weeks. Once the cadence is chosen, it should remain stable from one PI to another. This allows the ART to have a predictable rhythm and velocity. The fixed cadence also allows the train’s team members and stakeholders to schedule a full year of ART events on their calendars. The PI calendar usually includes the following activities:
PI planning event
System demos
Scrum of Scrums, PO sync, and ART sync meetings
Inspect and Adapt (I&A) event
The advanced notice provided by the PI calendar reduces travel and facility costs and helps assure that most of the stakeholders will be available to participate. Once the PI calendar is set, team events can also be scheduled, with each team defining the time and place for their DSU, iteration planning, review, and retrospective events. All teams on the train should use the same iteration start and end dates, which facilitates synchronization across the ART.
Training the ART Leaders and Stakeholders
Depending on the scope and timing of the rollout, there may be several ART leaders—Release Train Engineer (RTE), Product Managers, System Architects/Engineers, Business Owners, and other stakeholders—who have not attended Leading SAFe training.
It’s likely they will be unfamiliar with SAFe, will be unclear on expectations, and may not understand the need and benefits of their participation. It’s vital that they understand and support the new model, as well as the responsibilities of their new roles. SPCs will often arrange a Leading SAFe class to educate these stakeholders and motivate their participation. This is usually followed by a one-day implementation workshop, where newly trained stakeholders and SPCs can create the specifics of the launch plan. After all, it’s their ART; only they can plan for the best outcomes. Essentially, this is the handoff of primary responsibility from the change agents to the stakeholders of the newly formed ART.
Establishing the Agile Teams
During the implementation planning, questions will arise regarding how to organize the Agile teams around the solution’s purpose and architecture. Similar to organizing the ARTs themselves, there are two primary patterns for organizing Agile teams.
Feature teams. Focusing on user functionality, feature teams are optimized for fast value delivery. This is the preferred approach, as each is capable of delivering end-to-end user value. They also facilitate the growth of ‘T-shaped’ (multiple) skills of the team members.
Component teams. Component teams are optimized to achieve architectural integrity, system robustness, and reuse of assets (e.g., code, components, services). This type of team should be limited to significant reuse opportunities, areas of high technical specialization, and critical Nonfunctional Requirements (NFRs).
Most ARTs have a mix of feature and component teams. However, ARTs should avoid organizing teams around a technical system infrastructure (e.g., architectural layer, programming language, middleware, user interface) as this creates unnecessary dependencies, which reduces the flow of new features and leads to fragile designs.
The next step is to form the Agile teams that will be on the train. One innovative solution is to enable the people on the ART to self-organize with a set of minimum constraints.
In other cases, management may lead initial team selections based on their objectives, knowledge of individual talents and aspirations, timing, and other factors. This typically requires significant collaboration between the teams and management.
Prior to PI planning, all practitioners who will be joining the ART need to be part of a cross-functional Agile team. The initial roles of Scrum Master and Product Owner also need to be established. The team roster template shown in Figure 14-2 is a simple tool that can help clarify and visualize the organization of each team.
Figure 14-2. An Agile team roster template
The simple act of filling out the roster can be quite informative, as it starts to make the more abstract concepts of Agile development concrete. After all, the ideal structure of an Agile team is fairly well defined; the question of who is on the team, and the nature of the specialty roles, can lead to revealing discussions. Even the seemingly simple act of dedicating an individual to a single Agile team can be an eye-opening experience. But there’s no going back. The proven success patterns of Agile, including ‘one person–one team,’ are clear.
The geographic location column is also interesting, as it defines the level of collocation and distribution for each team. Collocation is better, of course. But there may be cases where one or more individuals cannot be physically located with the others. That may evolve over time, but at least everyone understands where the current team members reside, so they can start thinking about Daily Stand-Up (DSU) times and other team events.
Training Product Owners and Product Managers
Product Owners and Product Managers provide the direction for the train and are critical to the success of the ART. So, the people fulfilling these roles have to be trained to learn the new way of working, ensure collaboration, and understand how to best accomplish their responsibilities. In addition, these roles will be responsible for building the initial program backlog, which is a key PI planning artifact.
The SAFe Product Owner/Product Manager course teaches Product Owners and Product Managers how to drive the delivery of value together in the SAFe enterprise.
Training Scrum Masters
Effective ARTs rely on the servant leadership of Scrum Masters and their ability to coach Agile team members to improve team performance. Scrum Masters play a vital part in PI planning and help coordinate value delivery through Scrum of Scrums meetings. It’s incredibly helpful if Scrum Masters receive appropriate training before starting the first PI.
The SAFe Scrum Master course teaches the fundamentals and explores the role of Scrum in the context of SAFe. This course is beneficial for both new and experienced Scrum Masters.
Training System Architects and Engineers
System Architects/Engineers support solution development by providing, communicating, and advancing the broader technology and architectural view of the solution.
The SAFe for Architects course teaches senior technical contributors the role of architecture in a Lean-Agile enterprise. Attendees will explore the principles underlying Lean-Agile architecture, DevOps, and continuous delivery. They will also learn how to lead and support Solution Trains and ARTs, extend the principles driving continuous flow to large systems of systems and enable an improved flow of value across an entire portfolio.
Assessing and Evolving Launch Readiness
Training people in their new roles and responsibilities is key to ART readiness, but it’s only one element of a successful ART launch. PI planning is a significant event, and preparation is required. The ART Readiness Workbook in the SAFe PI planning toolkit provides a checklist for that purpose. It’s available to SPCs at https://community.scaledagile.com/.
However, since SAFe is based on the empirical Plan–Do–Check–Adjust (PDCA) model, there is no such thing as perfect readiness for a launch. Trying to be too perfect up front will delay learning, postponing the transformation and benefits realization.
Preparing the Program Backlog
Using the launch date as a forcing function increases the urgency to determine the scope and vision of the PI. This is defined by the vision and program backlog—the set of upcoming features, NFRs, and architectural work that outline the future behavior of the system. Consequently, SPCs and LACE stakeholders often bring the ART stake-holders together to prepare a common backlog. This is often done through a series of backlog workshops and related activities (Figure 14-3).
Figure 14-3. Preparing the program backlog and related activities
It’s easy to over-invest in backlog readiness, so don’t let that slow the process down, as the act of planning with the teams will sort out many issues. Experience shows that a list of well-written features with initial acceptance criteria is sufficient. There can be a tendency to over-plan and create user stories ahead of time, but that often creates waste and disappointment when the vision changes.
Step 8: Train Teams and Launch the ART
Now is the time to focus on the new and tentatively identified Agile teams who will make up the majority of the ART. Since these are the people who create the systems needed by the business, it’s critical that they understand their role in the ART and gain the Lean-Agile skills needed to be more effective. These team members may come from different parts of the organization—business, development, operations, support, and other domains—to define, build, test, and deploy their solutions. So, the next significant task is to train all the teams in the SAFe way of working.
Training Teams
It’s common that some people will feel that they do not need Agile team training. However, this course is critically important for digital transformation success, as the learning describes how Agile teams work together as a team of Agile teams.
The SAFe for Teams course offers an opportunity for team building and provides an introduction to Agile development. It covers the Agile Manifesto, Scrum, Kanban, and built-in quality practices, along with an overview of the Scrum Master and Product Owner roles. It also includes preparation for PI planning and building a Kanban board for tracking stories. In addition, teams prepare their backlog, which helps identify some of the work needed for the upcoming PI planning event.
Understanding the Benefits of Big-Room Training
In some rollouts, teams may be trained separately over an extended time period. However, we recommend a more accelerated approach, where all teams on the ART are trained simultaneously. This practice has created some controversy in the industry. Many compare this approach to the intimate setting of a small group with a single instructor and can’t imagine that it delivers equivalent benefits. In reality, it delivers far more. But this is something you have to experience to grasp its full effect, as the following story describes.
The ‘big-room’ training approach offers the following benefits:
Accelerated learning. This training happens in two days, rather than a few months, which helps the train come together as a team of Agile teams, accelerating the ART launch.
A common model for scaling Agile. All team members receive the same training, at the same time, from the same instructor. This eliminates the variability of different training sessions over time, by different instructors, using different courseware materials.
Cost-effectiveness. One of the challenges with Agile implementation at scale has been the availability and expense of training. Talented, proven instructors are hard to find and not consistently available, and their costs are correspondingly high. The big-room approach is typically three to five times more cost-effective than training one team at a time.
Collective learning. There is no substitute for the face-to-face interaction and learning experience of big-room training. It starts building the social network that the ART relies upon and creates a far better experience than what can be accomplished when working separately from each other. There can be a transformative aspect to it, something you have to experience to believe.
As different as it is, the ‘all-in, big-room training’ approach is one of SAFe’s most cost-effective and valuable implementation strategies.
Launching the ART
There are many ways to start an ART successfully, and there’s no specific timeline for the preparation activities we described earlier. However, our experience has shown that the easiest and fastest way to launch an ART is through the ART quick-start approach (Figure 14-4).
Figure 14-4. The one-week, all-in ART quick-start approach
After the preparation for the ART launch is complete, the quick-start approach trains people on Agile teams and holds the first PI planning event in a single week. Although that may seem daunting, numerous SAFe adoptions demonstrate this is the easiest and most practical way to transition 100-plus people to the new way of working.
Planning The First PI
During the ART quickstart, PI planning helps build team backlogs based on current priorities. It also reinforces the learning from the training. A successful PI planning event is essential to the success of the first PI. It demonstrates a commitment to the new way of working and offers the following benefits:
Builds confidence and enthusiasm for the SAFe adoption
Starts to establish the ART as a team of Agile teams and the social network that it relies on
Teaches the teams how they can take responsibility for planning and delivery
Creates full visibility into the mission and current context of the ART
Demonstrates the commitment of Lean-Agile leaders to the SAFe transformation
It’s recommended that an experienced SPC co-facilitate PI planning with the RTE to ensure success.
Step 9: Coach ART Execution
“Whenever you let up before the job is done, critical momentum can be lost, and regression may follow.”
—John Kotter, Leading Change
At this stage of the implementation, the first significant events are now in the rearview mirror. Teams have been trained, the first ART is launched, and the first PI planning event has been held. The result of all this effort is an empowered, engaged, and aligned team of Agile teams ready to begin building solutions that deliver value.
Before moving on to the critical work of the train, it’s important to understand that training and planning alone do not make the newly formed teams and ARTs Agile. They simply provide the opportunity to begin the journey. To support this effort, lead-ership—and SAFe Program Consultants (SPCs), in particular—need to be mindful that classroom knowledge does not equal understanding. It typically takes several PIs for effective Agile practices and behaviors to become the norm, which is why significant effort is required for coaching ART execution.
Coaching Teams
While SPCs often focus on enabling the ART roles and events, program execution ultimately relies on a competency in team and technical agility. So, SPCs are also needed to provide coaching to Agile teams in the following areas:
Initial iteration planning to learn how to refine and adjust plans developed during PI planning
Backlog refinement to adjust the scope and definition of user stories defined during PI planning
Daily meetings to help the team stay aligned on progress toward iteration goals, raise impediments, and get help
Iteration reviews and system demos to get feedback from stakeholders and assess progress toward PI objectives
Iteration retrospectives to review team practices and identify ways to improve
However, this is only the beginning. To establish a smooth and consistent flow of value, Agile software teams will need to become proficient in the built-in quality practices described in Chapter 6, Team and Technical Agility.
Most of the software practices were established in the Extreme Programming (XP) movement and remain a solid foundation for software craftsmanship. DevOps has made substantial contributions to these practices as well.
Training Software Engineers
The introduction of Lean-Agile and DevOps principles and practices has evolved the discipline of software engineering over the past decade. New skills and approaches will help organizations deliver software-centric solutions faster, more predictably, and with higher quality.
The SAFe Agile Software Engineering (ASE) course offers the foundational principles and practices that enable the continuous flow of value and built-in quality, including XP practices, such as BDD, TDD, and test automation.
Attendees also learn how to define, build, and test stories in a SAFe CDP. They will explore abstraction, encapsulation, intentional programming, and the SOLID design principles of object-oriented software development. They will understand how ASE fits into the solution context and learn to collaborate on intentional architecture and DevOps.
Coaching ARTs
Like the approach for Agile teams, coaching the SPCs typically coaches the ART through the following essential events:
PI planning. Creates alignment and shared commitment to a common set of objectives.
System demos. Closes the rapid feedback loop by integrating and validating working solutions
Inspect & Adapt (I&A) workshops. Enables relentless improvement and systems thinking.
Scrum of Scrums, PO sync, and ART sync. Maintains alignment, resolves issues, and facilitates achievement of PI objectives.
But these just scratch the surface of the ART’s purpose and potential. To help ARTs optimize the flow of value, SPCs coach ART leaders to look beyond the current PI and capabilities. As the team roles and events are mastered, the focus shifts to Agile product delivery and the CDP. This involves both managing and continuously improving the speed and quality of the ART’s capability to do the following:
Continuously explore. Sense and respond to market and business needs and apply design thinking to build and maintain the program vision, roadmap, backlog, and architectural runway.
Continuously integrate. Build, validate, and learn from working system increments.
Continuously deploy. Deliver validated features into production, where they are readied for release.
Releasing on demand. Release value to customers, with frequency and timing based on market and business needs.
While the program Kanban is the primary tool for visualizing and managing the CDP, SAFe DevOps, value stream mapping, and the I&A problem-solving workshop are the coach’s primary tools for enhancing these capabilities.
Training SAFe DevOps
To accelerate development of the CDP, DevOps training can be held during the first IP iteration or according to need and opportunity in subsequent PIs.
This course provides a comprehensive overview of the DevOps skills needed to accelerate time to market by improving the flow of value through the CDP. From concept to value, attendees will map the current value stream through their delivery pipeline and identify practices that will eliminate bottlenecks to flow.
The course will build an understanding of the complete flow of value, from continuous exploration to continuous integration and from continuous deployment to release on demand. Attendees will leave with the tools they need to execute an implementation plan for improving their CDP and the knowledge they need to support it.
Training Agile Product Management
Learning the right mindset, skills, and tools to create successful products—from inception to retirement—using Agile techniques to tap into new markets is critical in today’s fast-paced digital economy.
The SAFe Agile Product Management (APM) course offers advanced product management techniques for applying customer-centricity, design thinking, and continuous exploration to fuel innovation in the Lean enterprise.
Learners will understand how to accelerate the product life cycle to get faster feedback and quickly deliver exceptional products and solutions that delight customers—in alignment with your organization’s strategy, portfolio, evolving architecture, and solution intent. They will learn how to manage value stream economics, including licensing and pricing; how to use empathy to drive design; apply product strategy and vision; develop and evolve roadmaps; execute and deliver value using SAFe; and explore innovation in the value stream.
Conducting the Inspect and Adapt
There is no coaching opportunity for continuous improvement more critical than the first I&A event as described in the next section.
SPCs and coaches can assist the RTE in leading the first I&A event, where the current state of the solution is demonstrated and evaluated by the ART. Teams will then reflect and identify improvement backlog items during a structured, problem-solving workshop.
During the I&A everyone will learn the following:
How well the organization is adopting SAFe
How well the PI was executed and the quality of the system increment
How well the Agile teams and the ART performed against their PI objectives
In addition, SPCs and coaches can help lead the first problem-solving workshop, where corrective actions for the next PI are identified. This workshop gives teams the tools they need to relentlessly improve their performance independently. It also allows teams to work together—along with their management stakeholders—to collaboratively address the larger impediments that they face.
Summary
There is no such thing as ‘perfect readiness’ for an ART launch. Indeed, trying to be perfect upfront can potentially delay the transformation and the realization of its benefits. To avoid this, setting the launch date ahead of time provides an effective forcing function.
Launching an ART involves three main activities: preparation, training the Agile teams, and conducting the first PI planning. The preparation includes identifying and training ART leaders and specific roles such as Scrum Masters, Product Owners, Product Managers, and System Architects, as well as setting the program cadence and preparing the program backlog.
Following this preparation period, the one-week ART quick-start approach is a proven way to rapidly implement the new way of working. In this approach, the Agile teams are trained, and the first PI planning event is held, all in a single week. Once launched, opportunities are sought to coach the ART, focusing on supporting successful program execution alongside identifying improvements to the continuous delivery pipeline.
Launching and coaching the first ART creates an initial win, which ensures the energy and positive momentum needed to maintain the transformation and move onto the next step, which is to launch more ARTs and extend to the portfolio.