Aligned Autonomy at Scale (from Spotify Model)
- Roland Flemm
- Jul 18, 2024
- 7 min read
Updated: Aug 29
“Aligned Autonomy” is a concept that aims to strike a balance between autonomous decision-making and alignment with organizational goals and values.
The concept suggests that employees should have the freedom to make decisions and take actions autonomously, while also being aligned with the overarching objectives and principles of the organization.
"Aligned Autonomy" as a concept has been referred to in various sources, to name a few:
“Drive: The Surprising Truth About What Motivates Us” by Daniel Pink
“Turn the Ship Around!” by L. David Marquet
“Empowerment and Organizational Development” by Bernard Bass and Bruce Avolio
Another source is Henrik Kniberg, a famous Swedish agile coach and org consultant. Henrik coached at Spotify around 2012, where he described their unique scaling model, initially proposed by Joakim Sundén and his colleagues. Henrik created a cartoon series that sparked worldwide interest and became known as the "Spotify Model."
As a part of explaining how Spotify's ways of working were unique, Kniberg drew an image that describes the special relationships between the teams and management at Spotify:
This image describes the relationship between team alignment and autonomy as a two-dimensional matrix. Team autonomy is on the horizontal axis, and team alignment is on the vertical axis. According to the matrix, autonomy marks the extent to which a team can make decisions about its work. Alignment symbolizes having a common purpose. Upon examining the image, it is evident that, according to the creator, management plays a crucial role in connecting the two.
Key insights to derive from this matrix:
Alignment and autonomy are not different extremes of the same continuum
More alignment doesn't mean less autonomy
We need both (alignment and autonomy) to achieve a high-performing organizational setup.
In this article, we ask ourselves:
Which organizational design allows multiple teams to collaborate at scale with high autonomy and full alignment?
Alignment (for Purpose)
Alignment is having a shared purpose. When teams are aligned, they pursue the same goal (as exemplified by Kniberg's "crossing the river"). Alignment is about getting the noses in the same direction and working toward a shared purpose.
Management sets the boundaries and decides on how much autonomy is given to the "workers". Also, they are responsible for laying the groundwork for alignment. In other words, managers must effectively convey the "why", with strategy and purpose. Alignment and autonomy are closely intertwined, with management playing a crucial role in connecting the two. Alignment can be structured by management in many ways (by assigning coordinators or organizing for self-regulation), and it needs to match the granted level of autonomy (people cannot self-organize when a manager is (micro)-managing them).
Autonomy (for Agency)
For Spotify and other great companies, autonomy is a degree of freedom that teams have to act within the aligned purpose. Management "gives" autonomy to teams.
These days, a common term for this phenomenon is agency-degree of freedom. Intelligent agents are given the freedom to act independently for the benefit of a shared purpose.
While developing organizational topologies, we have discovered new insights about Autonomy and Alignment that differ from the Aligned Autonomy model presented by Kniberg.
Dimensions of Autonomy
Trying to establish autonomy in a group, for example, by creating autonomous teams, has become commonplace in most organizations. The management discussion on controlling work and outputs has shifted toward granting mandates and determining the degree of self-organization. The question is no longer whether teams should be more autonomous or not. The question is where that boundary lies.
Richard Hackman, Professor of Social and Organizational Psychology, provides a powerful illustration of the options:
By "autonomy of teams," most organizations mean the second column per Richard Hackman's model above: Teams must be granted the authority to "execute the task" and "monitor and manage work processes." This is the bare minimum of self-management. There is no self-organization when these two permissions are not granted to the teams.
Some organizations have tried the next level of "self-designing teams," where the teams not only execute and monitor but also "design the team and its context." We have experienced these experiments in LeSS-inspired adoptions and can confirm that this significantly enhances teams' agency.
Designing for Autonomy in a Multi-Team Environment
Do you agree that an individual team's speed is an interesting indicator of organizational performance? Are we truly measuring organizational performance when we examine a single team? Consider that the Billing team can be extremely fast at implementing a new Billing rule, but delivering the overarching customer journey, which requires contributions from many other teams, may take months. For a fact: the fewer (blocking) dependencies a team has, the more autonomous it is, the faster it can deliver product backlog items end-to-end.
On the Org Topologies map, making org units (such as individuals, teams, or AI agents) faster is a horizontal move. Units can be formed more quickly by acquiring skills, insourcing skills, removing dependencies, and reducing the need for certain skills through automation, among other methods.
In a multi-team environment, such a change is not a single-team transition but a coherent transformation of the larger ecosystem. In the picture below, three teams have merged to create a more autonomous CAPS-2 team with reduced skill dependencies on other teams. Analysts, testers, and front-enders now work as a team in the same cadence. The new CAPS-2 team possesses a higher degree of autonomy. They can decide on solutions as a group and handle more complex work at the capabilities level without needing assistance from others. However, they remain dependent on the back-end team.
Note that such a movement is an organizational design change, not simply a policy or process change.
Designing for Alignment in a Multi-Team Environment
Similarly, creating alignment in a multi-team environment is not a mere manager's blessing: "thou shalt now be all aligned."
People need to be aligned to complete a shared challenge. The question is: how can we arrange this to happen? What kind of alignment can we think of?
There's alignment through external coordination and alignment through collaboration.
Imagine a big room planning event (similar to Product Increment Planning in SAFe, for instance). The teams are gathered for a few days. They are presented with a strategy. The product board is filled with cards of different sizes and values. Teams are facilitated to identify mutual dependencies. The dependency manager connects cards with strings, and a detailed three-month plan is crafted upfront. Managers, coordinators, and teams are now aligned and understand how to manage the plan's execution and track dependencies. The teams are now free to proceed to their desks and begin work.
Are those teams aligned?
Yes, they are aligned. There is a shared purpose and a common plan to achieve this goal.
Cross-team meetings will be organized by the coordinators to compare the plan with the actual state and make necessary course corrections. As a result, the teams' work queues may change, reflecting the actual progress toward delivering value as a group.
Is it good enough?
Not really. Remember that we need alignment to deliver a shared challenge. The concept of a detailed plan, upfront identified dependencies, and external coordinators to manage individual teams will reduce the possibilities for cross-team collaboration. It increases the team's ability to work independently, but does not equip the teams to see the broader picture of the shared challenge.
Setting up for alignment through collaboration.
We believe that external coordination as a means to align is weaker than giving teams the mandate to align with each other through collaboration. The success of this approach will depend highly on the team composition.
Designing cross-team collaboration requires thinking about autonomy and alignment. We want to be in the top-right corner of the Kniberg matrix, where teams are highly aligned and have high autonomy.
Org Topologies offers a model for designing collaboration at scale. It identifies four distinctive levels of a team's focus:
These four levels are essential for designing effective collaboration.
As the teams' scope of work expands from task focus and capabilities focus up to partial business and whole business focus, the teams' scopes of work start to overlap. This is traditionally seen as a problem we want to avoid. But we can also see this as an opportunity for teams to collaborate.
Overlapping responsibilities are traditionally viewed as a problem to be avoided.
However, teams now have a shared scope of work and can join forces to attack big common challenges:
Alignment through collaboration is a result that emerges from overlapping responsibilities and shared goals.
Aligned Autonomy at Scale
We have discussed the different forms and shapes of autonomy and alignment. To create an organization where teams have full autonomy (not isolation) and can remain perfectly aligned (through collaboration), we need an organizational design that provides wide mandates on both the scope of work and the skills required to perform that work.
It is the space where highly collaborative practices, such as Multi-Team Product Backlog Refinement and Joint Sprint Reviews, become the norm. Refer to Elevating Katas™ to learn more about designing collaboration at scale.
Such an organizational design requires organizations to experiment with abandoning the idea of Independent, autonomous teams and instead creating cohesive, interdependent multi-team units: a team of teams.
At Spotify they created a space for collaboration with many cross-team structures and meetings:
Over time, through practice and tight collaboration, a team of teams will develop a broader understanding of the product. They will start to think holistically from the business and customer perspectives and realize that they need each other to make significant changes in the product.
Autonomy at scale requires a broad scope of work and skills to facilitate alignment through collaboration.
This might sound utopian, but don't worry. This is a "perfection vision" for your transformation. You can aspire to it and gradually move your organization in this direction. You can also grow gradually by trying out practices from the Elevating Katas™.
At a super large scale (hundreds of teams), there can be multiple "teams of teams", each specializing in a business domain and a set of cohesive customer journeys. In the case of financial services, for instance, there can be a team of teams for Retail Banking and another team of teams for Business Banking. Management provides direction by setting strategy and priorities, along with product goals. The teams turn those goals into tangible products.
Aligned autonomy at scale can be achieved through broad product knowledge (encompassing the scope of work, represented on the vertical axis) and a comprehensive skill set (encompassing the scope of skills, represented on the horizontal axis).
Endnote:
We do not claim that every individual needs to know everything and possess every skill. We claim that large groups of people can be mandated to multi-learn in both directions to acquire the capability of knowing everything and having every skill collectively as a group.
© Roland Flemm and Alexey Krivitsky
This was such a thought-provoking read! I really like how the article distinguishes between alignment through coordination versus alignment through collaboration. The idea that overlapping responsibilities can actually increase shared ownership and collaboration is a refreshing perspective. Too often, organizations see overlap as inefficiency, when in reality it can drive innovation and stronger cross-team bonds.
The “team of teams” approach resonates a lot, especially for large-scale enterprises. It reminds me of how learning and development works too—you don’t need every individual to know everything, but collectively, the group builds the capacity to handle complex challenges. That’s where autonomy at scale truly starts to show its value.
As someone who often dives into organizational design concepts for case studies (sometimes even through…