Ready to dive into the intricate world of team dynamics and software systems? Discover how industry giants like Amazon and Google optimise their teams for peak performance. Learn from real-world case studies from Poppulo, uSwitch, TransUnion, Sky Betting & Gaming and more.

Team Topologies: Organizing Business and Technology Teams for Fast Flow

In today's fast-paced technological environment, the way organizations structure their teams is crucial for delivering effective software. Traditional org charts often fail to capture the real communication dynamics within companies, leading to inefficiencies and bottlenecks. Naomi Stanford’s insights from "Guide to Organisation Design" suggest that viewing organizations as complex, adaptive systems rather than linear entities is essential for success. 

Software impacts all aspects of daily life, so designing these systems well is critical. Poorly designed software can have severe consequences, from stranding people to endangering lives. Therefore, building and running these systems requires the combined efforts of diverse, skilled teams. Organizations must balance speed and stability, adapting to changes while fostering innovation.

However, many organizations still rely on outdated structures. Org charts can be misleading, showing hierarchical lines that don't reflect actual communication flows. Instead, organizations need realistic pictures of how communication happens to optimize team interactions and software delivery.

Geary Rummler and Alan Brache's book, "Improving Performance," and the shift from project to product-centricity highlight the importance of continuous improvement and realistic organizational structures. Niels Pflaeging identifies three structures within organizations: formal (org chart), informal (influence), and value creation (actual work). Effective organizations leverage interactions between informal and value creation structures.

Team Topologies introduces a dynamic model for team structures, emphasizing adaptable design and realistic team interactions. This approach respects team cognitive load and aligns team structures with desired software architectures, leveraging Conway’s law. Mel Conway’s 1968 observation highlights the strong correlation between organizational communication paths and resulting software designs.

Cognitive load is a critical, often overlooked factor. Teams burdened with too many responsibilities can't perform optimally, leading to delays and quality issues. By respecting cognitive load and designing clear team boundaries, organizations can optimize for flow of change and feedback.

In short, successful software delivery requires rethinking team structures, purposes, and interactions. Embracing Team Topologies helps organizations dynamically adapt, promoting effective collaboration and strategic adaptability. Just like groupers and moray eels collaborate in the wild, teams within organizations must also work together to achieve better outcomes.

Can Naomi Stanford and Mel Conway's Insights Revolutionize Your Team Structures?

Conway’s Law, which posits that the structure of software reflects the structure of the organization that built it, underpins this idea. This law, formulated in 1968, still holds relevance today despite technological advancements like microservices and cloud computing.

Take Adidas, for example. They transformed their IT department from a cost center to a product-oriented team, boosting their software release frequency sixtyfold. This shift was driven by aligning their team structure with their software goals, a testament to Conway’s Law.

Harvard Business School research further supports this, revealing that a product’s architecture mirrors its organizational structure. Similar patterns have been observed in industries like vehicle manufacturing. Ruth Malan modernizes Conway’s Law: if the system’s architecture and the organization are misaligned, the organization’s structure will prevail.

The “reverse Conway maneuver” is a strategy to align team and software architecture. This involves reconfiguring teams to match the desired software architecture before it’s built. For instance, avoiding a monolithic database by embedding database skills within application teams can lead to more autonomous and efficient software solutions.

Effective team design also requires restricting unnecessary communication. Not all interaction is beneficial; focused communication between specific teams is key. This can be managed by optimizing the use of communication tools and ensuring that only collaborative teams share tools.

Adopting Conway’s Law in organizational design is crucial. It ensures that the structure of teams and their communication paths align with and enhance the software architecture, leading to faster, more efficient software delivery. By embracing this principle, organizations can navigate the complexities of modern software development and achieve significant improvements in performance and quality.

Adidas' IT Transformation Echoes Conway’s Law and Harvard Insights

In today’s complex work environments, especially in fields like software development, effective team performance is crucial. Research shows that cohesive teams outperform individuals when tackling knowledge-rich, problem-solving tasks. Even traditionally hierarchical organizations like the US Army recognize the power of well-functioning teams, as highlighted by retired General Stanley McChrystal in his book "Team of Teams."

Google's research affirms that team dynamics are more critical than the individual members. In software development, where rapid, diverse changes are the norm, teams are indispensable. The concept of “team-first” emphasizes stable, small teams of five to nine people, which fosters trust and innovation. Amazon’s “two-pizza” rule and frameworks like Scrum support this by limiting team sizes to maintain trust and effective collaboration.

Dunbar’s number, a concept from anthropology, suggests that humans can only maintain a certain number of deep relationships—about five closely trusted individuals, extending to 150 people whose capabilities we can remember. This principle, applied to teams, helps maintain high trust and efficient communication.

Organizations should flow work to stable teams instead of frequently reshuffling members. This stability allows teams to reach high performance, as seen in companies like Pivotal, where engineers switch teams roughly every 9 to 12 months without disrupting performance.

Ownership of software by long-lived teams is vital. Teams should have autonomy and responsibility over specific systems, fostering better care and continuous improvement. Rewarding the team as a whole rather than individuals aligns incentives with collective success, a practice that companies like Nokia have successfully implemented.

Managing cognitive load is another critical aspect. Teams should only be responsible for domains they can handle without being overwhelmed. IKEA’s example shows that even high-performing teams can suffer from cognitive overload, necessitating team splits to maintain efficiency.

The physical and virtual work environments should be designed to support team interactions and reduce cognitive load. Companies like Spotify and Auto Trader have optimized their office layouts to foster collaboration and flow, while remote-first strategies emphasize clear communication rules and efficient digital spaces.

Ultimately, a team-first approach, supported by proven engineering practices and thoughtful organizational design, leads to faster, more reliable software delivery and a healthier work environment. This approach not only enhances performance but also nurtures a diverse range of communication styles and reduces the impact of potentially destructive individuals.

Can Google's Team Dynamics and Amazon's Two-Pizza Rule Revolutionize Your Workplace Efficiency?

When software delivery is slowed down by ad hoc or constantly changing team structures, it's crucial to adopt intentional team designs, known as team topologies. There isn't a one-size-fits-all solution; instead, organizations should consider their technical and cultural maturity, size, and engineering discipline when choosing a topology. Feature teams can be highly effective if supported by the right environment, and splitting responsibilities can empower other teams and break down silos.

Organizations should avoid common anti-patterns like ad hoc team design and shuffling team members, which create volatility and inefficiency.

Organizational designs that emphasize the flow of change and reduce handovers between teams lead to faster, safer software delivery. The Spotify model, with its autonomous, cross-functional squads, is an example of explicit design aimed at improving software delivery effectiveness. However, blindly copying such models without understanding the underlying principles can lead to failure.

Effective team interactions and feedback loops from live systems are crucial for rapid problem-solving and responsiveness to customer needs. DevOps topologies, which emerged to bridge the gap between development and operations, highlight the importance of addressing team misalignments for successful software delivery.

Feature teams, product teams, and cloud teams each have specific requirements for autonomy and support. For example, feature teams need high engineering maturity and trust, while product teams need a support system to avoid bottlenecks. Cloud teams must enable product teams to provision their own environments to leverage cloud benefits fully.

Site Reliability Engineering (SRE) is a dynamic approach that balances operability and feature delivery, suitable for large-scale systems with high engineering discipline.

When choosing a team topology, consider factors like technical and cultural maturity, organization size, software scale, and dependencies. Detecting and tracking dependencies is essential for optimizing team flow.

Organizations should adopt and evolve team topologies that match their current context, taking a multi-stage approach to ensure adaptability to future changes. The DevOps Topologies catalog provides patterns and anti-patterns to guide organizations in structuring teams for effective software delivery.

Mastering Software Delivery: Lessons from the Spotify Model and DevOps Topologies

In many organizations, team structures can be complex and fragmented, making it hard to visualize the organizational landscape. However, simplifying team types into four fundamental topologies—stream-aligned, enabling, complicated-subsystem, and platform teams—can bring clarity and efficiency.

Stream-aligned teams are dedicated to continuous work flows aligned with business domains or capabilities, such as a specific product or user journey. These teams are empowered to build and deliver value independently, reducing the need for hand-offs. For instance, Amazon's "two-pizza" teams, which are small and independent, embody this approach, ensuring rapid delivery and accountability.

Enabling teams bridge capability gaps for stream-aligned teams, offering specialist expertise and helping them acquire new skills. They operate with a collaborative spirit, avoiding the "ivory tower" mentality by working closely with stream-aligned teams to improve their practices. For example, a legal organization formed an engineering enablement team to tackle long feature lead times and improve deployment efficiency, leading to significant improvements in their development process.

Complicated-subsystem teams handle parts of the system that require specialized knowledge, reducing the cognitive load on stream-aligned teams. These teams focus on areas like video codecs or financial algorithms, where deep expertise is essential.

Platform teams provide internal services that enable stream-aligned teams to work with autonomy. They offer self-service APIs, tools, and support, treating these services as products to ensure reliability and ease of use. For example, Auto Trader evolved its infrastructure team into a platform team to support its rapid growth and improve service reliability.

By adopting these four team types, organizations can optimize for rapid, safe flow of change, ensuring that most teams are stream-aligned and supported by enabling, complicated-subsystem, and platform teams. This approach not only simplifies organizational structure but also enhances collaboration and efficiency, driving better outcomes for the entire organization.

How Did Amazon's Two-Pizza Teams Revolutionize Organizational Efficiency?

When software systems struggle, the root cause often lies in how teams are organized and interact. Eric Evans, in "Domain-Driven Design," emphasizes the importance of aligning teams to the main streams of change within an organization to facilitate a smooth flow of software development. When team boundaries are poorly defined, it leads to a lack of ownership, disengagement, and a slow rate of delivery.

To improve this, organizations need to explore suitable boundaries within software systems that empower teams to effectively own and evolve their parts. This approach applies to both monolithic software and more modular systems. By aligning software boundaries to the capabilities of a single team, ownership and sustainable evolution become much more feasible.

Organizations like Amazon have demonstrated the need for guiding team interactions to achieve service independence. However, many organizations fall into the trap of creating "distributed monoliths" where services are interdependent, causing autonomy issues and slowing down changes.

There are various types of monolithic systems that can hinder progress:
1. Application Monolith: A single large application with many dependencies.

2. Joined-at-the-Database Monolith: Multiple services coupled to the same database schema.

3. Monolithic Builds: One large continuous integration build.

4. Monolithic Releases: Bundling smaller components into a single release.

5. Monolithic Model: A single domain language across many contexts.

6. Monolithic Thinking: Standardizing everything, which reduces team flexibility and motivation.

7. Monolithic Workplace: Standardized office layouts that can negatively affect collaboration.

To break down monoliths, we need to identify "fracture planes"—natural seams in the software that allow for clean splitting. These planes can align with various factors such as:
- Business Domain Bounded Contexts: Mapping to internally consistent business areas.
- Regulatory Compliance: Splitting systems based on regulatory requirements.
- Change Cadence: Different parts of the system changing at different frequencies.
- Team Location: Geographically distributed teams.

- Risk Profiles: Different risk tolerances within the system.
- Performance Isolation: Parts of the system with varying performance demands.
- Technology: Splitting based on older or less automatable technology.
- User Personas: Different user groups relying on different features.

Real-world examples like Poppulo show the benefits of aligning software boundaries with business domains and adopting practices like continuous delivery and DevOps. By focusing on team ownership and clear domain separation, they scaled their engineering division effectively.

Ultimately, choosing software boundaries that match team cognitive load and align with business domains is crucial. This approach supports autonomous teams, reduces dependencies, and promotes a faster flow of changes, ultimately leading to more efficient and sustainable software development.

Eric Evans' Blueprint for Team-Centric Software Success

In the realm of software delivery, effective team interactions are pivotal for innovation and rapid outcomes. Through well-defined modes of interaction—collaboration, X-as-a-Service, and facilitating—teams can achieve superior collective performance and adapt to evolving needs.

Collaboration is essential when high adaptability or discovery is needed, particularly when exploring new technologies. This mode fosters rapid innovation by minimizing costly hand-offs but requires a deep alignment and mutual respect between teams. However, it can blur boundaries and increase cognitive load, so it’s best used for short-term projects or specific discovery phases.

X-as-a-Service helps teams deliver quickly by consuming services or components with minimal collaboration. This mode emphasizes clear boundaries and ownership, reducing the cognitive load on teams and enabling them to focus on core tasks. Effective product management and service reliability are crucial for this mode to succeed.

Facilitating mode is about one team helping another to overcome impediments and learn new approaches. This mode increases overall productivity by detecting and resolving cross-team issues. Teams providing facilitation aim to make other teams self-sufficient, fostering a culture of continuous improvement and learning.

A case study from IBM illustrates the practical application of diverse team interaction modes to drive organizational change. IBM used a team of "DevOps advocates" to spread new practices and facilitate learning across global teams, showcasing the importance of adopting suitable team interaction modes for strategic advantage.

To adapt and evolve effectively, organizations must anticipate changes in team dynamics and use interaction modes to sense and respond to emerging challenges. Collaboration can help define viable service boundaries, while deliberate changes in interaction modes can refresh team experience and empathy.

By understanding and implementing these well-defined interaction modes, organizations can reduce uncertainty, enhance flow, and ultimately achieve better outcomes in software delivery.

How Did IBM Revolutionize Team Collaboration to Drive Innovation?

Modern organizations face rapid changes in technology, market conditions, and customer demands. To thrive, they must be adaptable, with flexible team structures and decision-making rules that can evolve as needed. The key is to design organizational rules rather than focusing solely on the structure itself.

One important aspect is determining the right level of collaboration for team interactions. Collaboration can drive rapid innovation but comes with a higher cognitive load. On the other hand, the "X-as-a-Service" model offers clarity and predictability but may slow down innovation.

Paul Ingles from uSwitch shared how adopting Kubernetes helped reduce complexity and cognitive load for development teams, illustrating how strategic technology changes can drive organizational improvements.

Another case study from TransUnion highlights the evolution of team topologies over several years. Initially, separate "system-build" and "platform-build" teams were formed to bridge the gap between development and operations. Over time, these teams merged back into their respective groups, improving operational awareness and accountability.

Sky Betting & Gaming experienced a similar evolution, with their "Platform Evolution" team transforming into "Platform Services" to better support product teams. This change involved reorganizing infrastructure teams to align with product goals, enhancing engagement and reducing friction.

The key takeaway is that team interactions and structures should continually evolve to match the organization's needs. Collaboration is essential for discovery and innovation, while the "X-as-a-Service" model is suited for predictable delivery. Organizations must balance these modes and adapt their team topologies to maintain agility and effectiveness.

By fostering a dynamic environment where team interactions are well-defined and purposeful, organizations can better sense and respond to changes, ultimately achieving greater success in a fast-paced world.

From Kubernetes to Team Topologies: Insights from uSwitch, TransUnion, and Sky Betting

In today's tech-driven world, team structure is vital for effective software delivery. Traditional organizational charts often miss the real communication flow, creating inefficiencies. Viewing organizations as complex, adaptive systems can help align team interactions and software outcomes. Successful software requires diverse, skilled teams balancing speed and stability while adapting to change.

Outdated structures and misleading org charts hinder communication. Realistic organizational designs, as outlined in "Team Topologies," respect team cognitive load and align structures with software architecture, leveraging Conway’s Law. Teams must manage cognitive load, ensuring clear boundaries and focusing on flow and feedback.

Effective team design includes stable, small teams fostering trust and innovation. Stable teams, like Amazon’s "two-pizza" rule or Scrum frameworks, maintain high performance by limiting size and reshuffling. Ownership and autonomy over specific systems drive continuous improvement.

Team topologies—stream-aligned, enabling, complicated-subsystem, and platform—streamline interactions and reduce cognitive load. Organizations should aim for stream-aligned teams supported by enabling and platform teams. Aligning software boundaries with team capabilities fosters ownership and evolution.

Adopting suitable team interaction modes—collaboration, X-as-a-Service, and facilitating—enhances collective performance. Collaboration drives innovation, while X-as-a-Service offers clarity and predictability. Facilitating mode boosts productivity through cross-team support.

Organizations must continually evolve team structures and interactions to match changing needs. Balancing collaboration and service models ensures agility and effectiveness, leading to better software delivery and overall success.

Home

Trending Books

News & Mags

Relax

Curated

Collections

Key Ideas & Notes

Insights

Highlights & Bookmarks

Listening History

Team Topologies: Organizing Business and Technology Teams for Fast Flow

Matthew Skelton

9 Key ideas

21 MINS

4.8 (490)

Ready to dive into the intricate world of team dynamics and software systems?
Discover how industry giants like Amazon and Google optimise their teams for peak performance. Learn from real-world case studies from Poppulo, uSwitch, TransUnion, Sky Betting & Gaming and more.

Ready to dive into the intricate world of team dynamics and software systems?

Discover how industry giants like Amazon and Google optimise their teams for peak performance. Learn from real-world case studies from Poppulo, uSwitch, TransUnion, Sky Betting & Gaming and more.

Key Ideas

#1