SAFe, a large-scale agile framework, and an introduction to the differences from other large-scale agile frameworks

The Scaled Agile Framework (SAFe) is an organizational structure and workflow model for an enterprise to implement agile practices, mainly based on the three knowledge systems of agile software development, lean and systems thinking.

The SAFe framework provides guidance for agile implementation including roles, responsibilities, planning, management, values, and team structure. At the same time, it also provides a set of structural methods for large enterprises to achieve large-scale agility . Four configurations are designed to suit teams of different sizes: Basic SAFe (Essential SAFe), Large Solution SAFe (Large Solution SAFe), Portfolio SAFe (Portfolio SAFe) and full SAFe (Full SAFe), effectively promoting the consistency, collaboration and delivery of large-scale agile teams.

With the development of time, enterprises have been unable to cope with the fierce market competition with traditional software project management methods. So agile coaches Dean Leffingwell and Drew Jemilo launched the SAFe framework in 2011, aiming to help companies optimize their organizational structure and improve the quality of software delivery to meet changing customer needs.

Today, SAFe has become one of the most popular scaled agile frameworks, and agile coaches practice and continuously improve the SAFe framework in enterprises around the world.

1. SAFe Core Principles and Values

1. Core Values of the SAFe Framework

SAFe takes enhancing leadership and improving team operational efficiency as its core values.

Consistency : SAFe requires the enterprise to formulate a unified work plan and conduct periodic summaries at all levels, so that each member must understand the current business progress and business goals, and how each person will contribute to the realization of the goals. The advantage of this is that it can unify the work rhythm and work tasks of the team, and improve the coordination of resource allocation at all levels. Different from the traditional top-down control mode of information transmission, the information in SAFe mode can flow from top to bottom in a timely manner.

Built-in quality: Increased agility should not come at the expense of quality of work. SAFe requires teams at all levels to define "done" metrics for each task or project and to incorporate high-quality development practices into each work process. According to SAFe's description, built-in quality has five key dimensions to measure the quality of work: process, architecture and design quality, code quality, system quality and release quality.

Transparency: SAFe encourages trust-building behaviors in the team, such as: planning work in smaller cycles so that problems can be found faster; to-do list progress is transparent in real time at all levels, so that norms can be checked and adjusted in a timely manner.

Program Execution: Program execution is the core of SAFe and the driving force of all rules in the framework. Individual teams and project groups must be able to create value and regularly deliver high-quality, high-performance software.

Leadership: The realization of SAFe requires the support of lean and agile leadership behaviors, because only the leadership has the power to change the organizational system and create a working environment conducive to the realization of core values.

2. The nine principles of the SAFe framework

All principles of the large-scale agile framework are designed to improve the operating efficiency of the entire company by inspiring lean decision-making in cross-functional and cross-organizational collaboration . Its role is not only to influence the decision-making of leaders and management, but also through the practice of lean portfolio management, it will affect the decision-making and thinking mode of every member of the organization - from the traditional waterfall thinking to lean and agile .

Principle 1: Take an economic perspective

The theory in Donald G. Reinertsen's best-selling book on the product development process tells us that to minimize the lead time for sustainable delivery, everyone in the decision chain needs to understand the consequences of delayed delivery. economic impact. The approach of short-term frequent delivery is not always suitable. The SAFe framework stipulates that prioritizing work to maximize benefits, understanding factors that affect business returns, and thinking about how projects will work within a lean budget is a shared responsibility of all members. Many of the concepts and tools in the SAFe framework come from Leinazen's theory of the product development process.

Principle 2: Use systems thinking

The SAFe framework encourages organizations to apply systems thinking in three key areas: business solutions, enterprise architecture, and value streams . Solutions can refer to products, services or systems that are delivered internally or externally to customers.

The business solution of a large-scale project usually consists of many interrelated parts, and team members must have an overall view of how to integrate the parts they are responsible for into the whole. In terms of structure, enterprises should consider employees, management and processes at the same time. If an enterprise wants to improve the working methods of its employees by practicing SAFe, it must first clear the barriers to inter-departmental cooperation, enable teams to truly achieve cross-functional cooperation, and then sign new working agreements with suppliers and customers.

Finally, the business should clearly define how value is translated from concept to money in the solution development value stream. And leaders and management need to maximize the flow of value across functional and organizational boundaries.

Principle 3: Reserve space, embrace change

By default, systems and software development face many uncertainties. Therefore, this principle deals with uncertainty by introducing the concept of set-based design (Set-Based Design). In the early stage of research and development, don't just select a set of solutions that you think are correct, and then go all the way to the dark. Instead, keep a design set and collect empirical data in subsequent research and development. In this way, the more options are eliminated in the later stage, the easier it is to select the best solution, and even the final solution can be gradually determined after entering the market. So as to deliver the most satisfactory results for customers.

Principle 4: Establish a continuous learning closed loop and carry out incremental construction

Similar to principle 3, this principle calls for addressing risk and uncertainty through “learning milestones”. Because the development team can't just focus on whether some functions in the system work well, but must test the entire system in time to evaluate whether the current solution design choice is feasible. Learning milestones require teams to plan learning integration points to accelerate the learning cycle. The integration point concept comes from Shewhart's PDCA (Plan-Do-Check-Adjust) cycle, which is a mechanism for continuous quality improvement and control of variables in development.

Principle 5: Establish milestones based on an objective assessment of feasibility

Compared with superficial results evaluation such as reports, regular presentation of work results is more conducive to final decision-making. Involving stakeholders early in actionable decisions builds trust and systems thinking.

Principle 6: Visualize and limit WIP, take small steps

Limiting work in progress helps stakeholders understand exactly how work is progressing. The three elements of this principle are the primary means of achieving maximum throughput and accelerating value delivery. Applying visualization and limiting WIP to software development means: reducing the amount of overlapping work, reducing the complexity of each work and the total amount of work that needs to be processed in a given amount of time. Reducing the batch size allows the team to continuously verify that the work is going in the right direction.

Principle 7: Establish a synchronized working rhythm through cross-cutting initiatives

The general way agile teams work is paced through sprint or iteration meetings. Setting a cadence for all eventualities reduces complexity, reduces uncertainty, builds muscle memory, enhances quality of work, and helps spread the spirit of collaboration across teams. If the work rhythm is synchronized and information is shared, decision-making and incremental planning can be facilitated, and the work of team members will mesh with each other like gears, which will help accelerate the progress of the entire project.

Principle 8: Unleash the Intrinsic Motivation of Knowledge Workers

Inspired by renowned management consultant Peter Drucker and best-selling author Daniel Pink, this principle is designed to unleash the inner dynamics of teams and help leadership through the lens of mentoring and serving teams Departure, not a command and control mentality.

Principle 9: Decentralized decision making

By decentralizing decision-making to shorten process length and adopt an appropriate approach, teams can be given more autonomy. Leaders should retain decision-making authority for team members on topics of strategic importance, enabling the team to make informed choices about everything else.

3. Implementation steps of SAFe

Organizations that want to practice SAFe need executive support, a strong desire to change, and a Scrum foundation.

SAFe presents a detailed implementation roadmap of how to get started and how to set up an organization to achieve agile at scale across portfolios.

The 12 steps of SAFe include:

Reaching the tipping point
Train lean-agile change agents
Train executives, managers, and leaders
Create a lean-agile center of excellence
Identify value streams and Agile Release Trains
Create the implementation plan
Prepare to start ART (Prepare for ART launch)
Train teams and launch the ART (Train teams and launch the ART)
Coach the ART execution
Launch more ARTs and value streams
Extend to the portfolio
Sustain and improve

4. The similarities and differences between SAFe and other large-scale agile frameworks

Although SAFe has been widely practiced in many large software development enterprises, there are also many large-scale agile frameworks that have gradually become favored by enterprises over time. All large-scale agile frameworks have five common building blocks: the twelve principles of the agile manifesto , cadence, synchronization, scrum, and quality development practices. Understanding the origins of other frameworks, their key differences, and the conditions under which they work can help organizations choose which framework is best for them.

1.SAFe vs. Scrum@Scale

In the Scrum@Scale (S@S) framework , each member is part of an interchangeable Scrum team, and the decentralized Scrum team network can be aggregated into an ecosystem according to the goals. S@S aims to form a network of Scrum teams through a "freely scalable architecture" to achieve linear scalability without introducing new dynamic processes. For example, for a business with 25 Scrum teams, a Scrum of Scrum (SoS) may not be able to handle very complex product deliveries, so a Scrum of Scrum of Scrums (SoSoS) or a Scrum of Scrum of Scrums Master (SoSM ) to complete.

S@S does not have many normative standards like SAFe, but it gives a guiding question - if more people are added to the organization, will the operating efficiency of the enterprise increase exponentially, or will the productivity be affected? Influence? Thus inspiring enterprises to think about whether they are ready to implement large-scale agile.

Like SAFe, there are many references to S@S on the web, such as the frequently cited Scrum@Scale Guide.

Conditions for S@S to work:

Technology stack is user-oriented (e.g. vertical user stories can be delivered in two weeks)
Functional teams with T-shaped skills, product-centric values, and minimal bureaucracy
Agile or Agile Lifecycle Management (ALM) tools are not needed until S@S practices become second nature
The executive team is willing to practice Scrum and is committed to removing impediments in the organizational structure

2.SAFe vs. Large-Scale Scrum (LeSS)

The LeSS framework adopts a simplified strategy for roles, structures, and artifacts. While SAFe offers four configurations to accommodate increasingly large and complex solution teams, LeSS has only two:

LeSS model: suitable for the size of 2-8 teams;
LeSS Huge mode: adapt to the size of more than 8 teams.

There are also big differences between LeSS and SAFe when it comes to decision-making. The product owner ( Product Owner ) of LeSS has absolute strategic decision-making power, while SAFe is more democratic in decision-making. The generation of SAFe decision-making planning is affected by many factors, while LeSS is more focused on customer-centricity and focusing on serving paying customers.

Like S@S, LeSS is also based on the Scrum framework, with large-scale and agile extensions on Scrum's events, artifacts, and responsibilities. The guiding principles of SAFe and LeSS are similar, emphasizing both systems and lean thinking. But LeSS aims at continuous improvement and pays more attention to reducing the waste of resources in the organizational system.

Conditions for LeSS to work:

Scrum teams have mastered Scrum agile practices
Leadership is willing to rebuild and experiment for the long run
Have a consistent product definition
There is a standard definition of "done"
Hire an external coach to participate in the collaborative improvement of the organization, team and technical team
Functional and component teams with T-shaped skills
Organizations are willing to move away from traditional project management models entirely

3.SAFE vs. IN

Unlike any of the other frameworks above, Disciplined Agile (DA) is a toolkit that lets businesses decide which way of working is best for them . It is a lightweight agile management method rooted in Scrum and Kanban, including human resources, finance, management, DevOps, portfolio management and other fields of knowledge. DA sets different levels of agile management methods according to different projects, focusing on guiding enterprises how to make decisions on strategic directions.

Conditions for DA to function:

Enterprises want to define their own scalable and agile expansion path
Enterprises want to maintain the flexibility of the organizational system
Businesses don't want to decide quickly which workflow or management framework to choose

4. SAFe vs Spotify mode

Spotify "patterns" are a set of human-driven, highly autonomous, scalable agile management methods designed to facilitate collaboration among agile teams . Strictly speaking, the Spotify model cannot be called a model or framework, but some companies practice it as a framework. The Spotify model emphasizes autonomy and cross-functional collaboration, and calls teams working in the same location "squads" (equivalent to Scrum teams), while SAFe only encourages teams to perform PI planning, and does not require teams to work in the same location.

In the Spotify model, large teams of multiple "squads" are called "tribes." There is not too much interdependence between the "Squads", and they collaborate through SOS (Scrum of Scrums) when encountering complex tasks. "Chapter" and "Association" are informal groups established by skill classification and interest, and internal knowledge can be shared.

While the other models mentioned in this article have online resources, training courses, and certification bodies, the material on the Spotify model is limited to published blog content, and experience sharing from businesses and fans who have practiced it first. The Spotify model is becoming more and more popular, and I believe there will be richer theoretical summaries in the future.

Conditions for Spoify mode to work:

The practice of this mode should be combined with the actual situation of the company.
Organizational culture focused on learning, allowing for mistakes, and taking manageable risks
Teams and products are "loosely coupled, tightly aligned" to avoid dependency conflicts

5.SAFe 5.0

One of the core principles of the SAFe framework is to continuously optimize and upgrade after learning from the practical experience of companies around the world. Recently, the community Scaled Agile, Inc. launched SAFe version 5.0, which added the tenth principle: "Organize around value" and added the twelfth step Changed from "maintain and improve" to "accelerate".

epilogue

SAFe, and the other frameworks described in this article, provide a practical approach to how organizations can achieve agile transformation internally and achieve business results. Thinking about how to get the most out of the agile framework you're following is as important as deciding which tool to choose.

Further reading: Guide to Agile Development

Scrum Development Guide: Detailed Explanation of Scrum Framework | Four Scrum Meetings and Correct Ways to Hold | Correct Way to Plan and Execute Sprint | 4 Key Points for Doing Iterative Planning | Do these 4 points to make daily stand-up meetings more suitable for agile Team | 3 Key Steps to Opening a Good Iteration Review | Why Hold an Iteration Review | The 3 Scrum Roles and Their Specific Responsibilities | The Role of the Scrum Three Artifacts in Agile Development | 14 Best Scrums in 2022 Agile Project Management Software | More

Kanban Agile Guide: 5 Benefits of Using Kanban (Kanban) Management Method | Kanban VS Scrum: How to choose? | Which scenarios are suitable for the mixed mode of Kanban and Scrum | More

Product Management: How to Build a Qualified Product Roadmap | How to Become a Good Product Manager | The Importance of Agile Roadmaps and Construction | How to Build Simple and Effective Product Requirements Documents | Use NPS to Determine Functional Priority | Every Product Manager Needs Learn Product Analysis Skills | More