The art of being ahead, consistently.
Toyota is one of the most intensively studied organizations, with thousands of researchers and scientists having investigated its success. And while other companies tried to imitate the Toyota product system, the competition improved but never caught up to Toyota’s success. Why?
Learning organizations like Toyota never sit still. They constantly improve with a decentralized approach as part of their culture and daily routines. This white paper translates the high-velocity edge management approach by S. Spear (MIT) [1] to a software development management system.
What is a learning organization?
In times when most software departments want to become more agile, the first steps usually include breaking work into smaller chunks and delivering added value in shorter iterations, thus, becoming more adaptable. A second dimension is increasing delivery speed, reliability, and focus. This is usually achieved via automation and, especially, structural improvements in quality.
While these steps might already sound challenging, they are usually not as soon as the people doing the work are enabled (can perform within their accountabilities) and are empowered (have the right to perform changes within their role). As a result, it is not the responsibility of the management to achieve the above-listed goals but rather the accountability to create the environment where the teams and individuals can make them happen. Two important elements for management to consider are the following.
Structures
Learning organizations tend to put great effort into developing technical competency and how the work of individuals, teams, and technology will contribute to the outcome. These organizations are process-oriented, in contrast to silo-oriented structures of other organizations.
Dynamics
Learning organizations are constantly experimenting and learning to understand and solve problems when and where they occur.
The principles of a learning organization
The following principles universally apply in all high-velocity organizations as per S. Spear [1]:
Create your organization in a way that work is done in the best-known approach, standardized across all teams with similar challenges, in a way that signals when and where problems occur.
Solve problems when and where they arise in a way that prevents the problem from recurring.
Multiply the impact of local discoveries through a sharing mechanism to make them systemically useful across all teams with common interests.
Lead your organization by enabling and empowering your knowledge workers to facilitate the constant development of great discoveries so that these discoveries happen all the time, everywhere.
Graphic 1: high velocity organization principles
It might sound trivial but implementing and living by these principles can be quite challenging for many organizations, as they might conflict with their norms, culture (e.g., command-and-control), or psychological barriers (loss of power and control).
“To be successful in a knowledge economy firms need to create learning organizations."
Don Tapscott
Capabilities of a learning organization
Like building cars at Toyota, software development is a constant and repetitive process, where business requirements must be estimated, confirmed, detailed, developed, tested, accepted, and released. One can use Scrum as a starting point to build a process around these activities. While the above principles can be applied completely independent of lean-agile, this paper combines already widely established concepts with ideas that allow learning organization principles to be established and improved.
The four steps to translate the learning organization principles into our lean-agile processes.
1. See
Our goal: establish the most effective approach currently known for achieving success at the task and build in that approach the capacity to detect failure when and where it occurs. [1]
During this step, we simplify and standardize all our product team activities. First, the same accountabilities are owned by the same roles (see graphic 2, and for more details on our authority model see [2]). Second, the same process steps are performed in standardized practices (ceremonies) across all teams. Third, transparency is created using similar artifacts to manage all stakeholder interests, including progress on delivery, incidents, dependencies, availability, the stability of the productive solution, and team improvements.
The governance roles interact with the teams continuously as part of the standard practices during any cycle (see graphic 3) to detect problems when and where they occur:
during daily: ask everyone about impediments
at end of sprint: during retro, discuss challenges and consider improvements to be implemented
at end of PI cycle: during PI retro, discuss intra-team and business collaboration challenges
The governance authority must actively capture known impediments in an improvement backlog. These improvement items consist of at least three information pieces: the problem statement, the expected benefits of resolving the problem, and a conceptual idea or potential solution approach.
Graphic 2: Separation of accountabilities into authorities
2. Solve
Our goal: detect problems at the time and place they occur. Then, contain those problems before they can spread. Do so by formulating an action and anticipating an outcome so you can see when you are right and wrong. Fixing a problem is not enough; understanding why a certain action has fixed a problem is also crucial. [1]
The governance roles in every team and at scale own all impediments; however, they must identify solutions collaboratively with their team and document the expected benefits of an identified action with others quickly after a problem is identified. In urgent cases, quickly means immediately. In other cases, this can happen during the sprint retrospective. Depending on the effort required and people involved, these improvements are realized immediately or are prioritized during the retrospective and committed during a subsequent cycle. After implementing an improvement, the benefit statement should be re-evaluated to ensure that the implemented solution prohibits the same problem from reoccurring in the future.
Graphic 3: Basic PI cycle with nested Scrum cycle.
3. Share
Our goal: multiply the power of new learning by making them available. Not only solutions but also the processes by which they were discovered. [1].
Often, challenges and impediments arise that are specific to a particular accountability and, therefore, to a specific role owning that accountability. Therefore, communities of common interests can be formed, bringing role holders with the same standardized role together. A similar concept is embedded in the Spotify model with chapters. These communities meet regularly, discuss identified impediments, and highlight how the team has experimented to find a solution that solves the problem sustainably. Other participants might then discuss their experiences that could further improve the approach to simplify and standardize the solution. Generically, such communities should have a community manager who defines the participation group, a channel manager responsible for keeping the knowledge persistent and accessible, and, finally, they need a content manager responsible for keeping the focus on meaningful problems, approaches, and answers.
4. Lead
Our goal: leaders of a high-velocity organization value both the delivery of a product or service and the continuous improvement of the processes by which those products or services are delivered. Such organizations become ever more self-diagnosing, self-improving, skilled at detecting and solving problems, and multiplying the effect by making the solutions available throughout the organization. [1]
A learning organization can only evolve if senior management creates the conditions in which everyone is enabled and empowered and where experimentation is explicitly encouraged to find superior solutions. Failure must be appreciated as a significant way of learning. Based on mutual trust and a centrally-governed knowledge-sharing platform, an ever-improving organization can be sustainably established. Hence, senior management has the primary accountability to enable a learning system as follows:
centrally manage and support communities (find role holders, provide channels and tools to persist accessible knowledge)
create new knowledge/ learning decentral when and where problems are identified (enable and empower employees to do so)
track the implemented improvements and validate expected benefits to further learn, simplify, and standardize a solution that can be reused by others
“Through learning we become able to do something we were never able to do."
Peter Senge
Conclusion
Despite working within the same market, the same access to technology, and a similar workforce, some companies can repeatedly set themselves apart from their competition. It is not the environment that allows them to thrive, it is their ability to work in a simpler, more standardized, and more stable way, and continuously succeed, not only with their product or service delivery but also with their continuous will to constantly improve their way of working. However, to get there, it sometimes takes a tremendous effort, especially at the leadership level. Besides the discussed amendments of the management system, these also include political and cultural hurdles.
As a leader, start with the conditions necessary for everyone to succeed (enablement and empowerment). Then, optimize the system by simplifying and standardizing the approach, before helping the teams to stabilize (to avoid problems from spilling over). Gather real-time information from their systems instead of manually provided, subjective management reports. Then, strengthen the dynamic of seeing and solving problems continuously. As a result, in high-velocity work environments, people tend to be more grateful and harmonic in their collaboration and supportive of each other.
Spark Mind has the knowledge and skills to support you in establishing an understanding of your organization’s situation and supporting a transition to a setup as discussed in this paper.
Download the PDF version of this article here:
_______________________________________________________
Comentários