Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Presentations Architecture for Flow with Wardley Mapping, DDD, and Team Topologies

Architecture for Flow with Wardley Mapping, DDD, and Team Topologies



Susanne Kaiser illustrates the concepts of DDD, Wardley Mapping and Team Topologies, and demonstrates how these techniques help to evolve a fictitious legacy system for a fast flow of change.


Susanne Kaiser is an independent tech consultant supporting organizations with building socio-technical systems. She is passionate about connecting the dots between Wardley Mapping, DDD, and Team Topologies as a holistic approach to design and build adaptive systems. She is the author of the book "Adaptive Systems with DDD, Wardley Mapping, and Team Topologies: Architecture for Flow".

About the conference

QCon Plus is a virtual conference for senior software engineers and architects that covers the trends, best practices, and solutions leveraged by the world's most innovative software organizations.


Kaiser: Thanks a lot for joining me in this talk about architecture for flow with Wardley Mapping, domain-driven design, and team topologies. Most initiatives that try to change or to optimize a system are usually directed at improving parts taken separately. They tend to focus on local optimization of the separate parts of the system. From a system thinking lens, local optimization of separate parts will not improve the performance of the whole. Dr. Russell Ackoff, one of the pioneers of the system thinking movement stated that a system is more than the sum of its parts. It's a product of their interactions. The way parts fit together determines the performance of a system, not on how they perform taken separately. He also claimed, if the systemic nature of an organization is ignored, the effort of improving and changing that system are doomed to failure.

Challenges of Building Systems

When building systems in general, we are facing the challenge between building the right thing and building the thing right. Building the right thing addresses effectiveness and addresses questions such as, how aligned is our solution to our user needs? Are we creating value for our customers? Have we understood the problem and do we share a common understanding? Building the thing right, on the other hand, focuses on efficiency, for example, efficiency of engineering practices. It's not only crucial to generate value, but also being able to deliver that value, and about how fast can we deliver changes and how fast and easy can we make a change effective and adapt to new circumstances? The one does not go without the other, but as Dr. Russell Ackoff pointed out, doing the wrong thing right is not nearly as good as doing the right thing wrong.

Three Perspectives to Build Adaptive Systems

By considering the whole and having effectiveness and efficiency in mind, to build the right thing right, we need a holistic perspective to build adaptive systems. One approach that I would like to share with you is combining the perspectives of business strategy with Wardley Mapping, software design and architecture with domain-driven design, and team organization with team topologies in order to design, build, and evolve adaptive socio-technical systems that are optimized for fast flow of change.

Evolving a Legacy System

Let's put these three perspectives together by evolving a legacy system. We are using a fictitious example of an online school for junior students based on a legacy system built, run, and supported by functional silo teams. The first step to evolve our system is starting with a team and analyzing their current situation in regards to their team cognitive load and encounter delivery bottlenecks. Due to handover, the functional silo teams reported high communication and coordination effort between teams when implementing and releasing changes. Dealing with a monolithic big ball of mud with a messy model and no clear boundaries leads to high cognitive load and tight change coupling. It takes a high amount of effort to understand a piece of code. In addition, the teams reported about no clear ownership boundaries they are facing in general with delivery bottlenecks and impeding their software delivery performance. Due to on-premises infrastructure, the operational activities consist of configuration, backup and recovery mechanisms, maintenance, including upgrades and security updates, scaling, monitoring, and so on. The infrastructure team faced very high operational efforts. As a reminder from system thinking, if you plan to optimize only separate parts, it does not improve the performance of the whole.

Visualizing the Landscape with a Wardley Map

Improving the performance of the whole requires to tackle the situation from a broader perspective and also understand the environment and organization it's operating and competing in. This brings us to a Wardley Map representing the context specific landscape of an organization. A Wardley Map is part of Wardley Mapping. Wardley Mapping is a strategic framework invented by Simon Wardley, and it helps to design and evolve effective business strategies based on situational awareness and movement following a strategy cycle. The strategy cycle is representation of change and how we need to react to the change according to Simon Wardley.

It starts with a purpose, the why of our business, in our case, providing high quality education for junior students everywhere and help teachers to engage with their students online. The next section of our strategy cycle is the landscape an organization is operating and competing in, which is visualized with a Wardley Map. Let's create a Wardley Map of our online school example. A Wardley Map is in general composed of a y axis and an x axis, and the y axis represents the value chain. To derive a value chain, it starts with identifying your users first, and your users can be customers, shareholders, staff, business partners, internal users, and so on. We are going to focus on the teachers and the junior students as the users of the online school.

Next is identifying the user needs, and the user needs are addressing the problems the users would like to get solved. They are the subject area for which we build a product. The users and the user needs, they represent the anchor of the map that all subsequent components relate to. The teachers would like to create course content for the students, they would like to plan a class supporting the students during their studies, and also evaluating their progress. The students have the user needs of studying courses, requesting and receiving help during their studies, and receiving evaluation feedback. Both teachers and students, they need to sign up and sign in as well. Next, we are identifying the components that fulfill the user needs either directly or indirectly by facilitating other components in that value chain, and determining the dependencies and the positions in the value chain. At the top, we have components that are most visible to the users, so where the users are touching the system. At the bottom, it gets more invisible to the user. The value to the user and the visibility, they are closely correlated. The more visible a component is, the more value it provides to the user.

The component that the teachers and students are interacting directly with is the online school component. It's located at the top of the value chain. At this point, we are dealing with a monolithic big ball of mud as one single component that we are going to decompose later into modular components when we talk about the second perspectives of domain-driven design. The online school component, this depends on infrastructure components such as data storage, search engine, message broker, SMTP server, and compute platform running on top of a virtual machine component. They are less visible to the users and placed further down the value chain. Each component of this value chain is going to be plotted along an evolution axis going from left to right. At the left, we have Genesis with brand new things. Then custom built. Then product and rental such as off-the-shelf products, or open source software solutions. Then things on the right like commodity and utility. For determining the stage of evolution per component, let's look at a couple of climatic patterns that are influencing our landscape.

That's where we come to the next section of our strategy cycle, the climate. Climatic patterns describe external forces that are impacting our landscape over which we have no control. However, understanding climatic patterns is important when anticipating change, according to Simon Wardley. Understanding climatic patterns gives us an idea where the landscape can change and where to invest in the future. One pattern is that the landscape is never static but very dynamic. Everything evolves from left to right through the forces of supply and demand competition. For example, cloud hosted services are reflecting this climatic pattern. What was decades ago non-existent, evolved through Genesis, custom built, and became product and rental, and now commodity and utility. As the components evolve, their characteristic change from an unchartered domain, undefined market, uncertain, unpredictable, constantly changing, poorly understood chaotic domain, and while it evolves and becomes more stable and mature. While we are then facing an industrialized, mature, known, commonly understood, ordered market on the right.

We can use these characteristics to determine its stage of evolution per component. The online school component is reflecting a volatile component that is changing frequently, and that is providing competitive advantage. It shall go into the custom built evolution stage. The more stable, mature, widespread infrastructure components on the other hand can go in product and rental or commodity and utility evolution stage.

Applying Doctrinal Principles

For each evolution stage, we can apply appropriate methods, that's where we come to the Wardley Doctrine, the next section of the strategy cycle. Doctrine describes universal principles that each industry can apply, regardless of their context. Applying doctrinal principles enables an organization to be able to respond to changes quickly and to be able to absorb changes gracefully. One of the doctrinal principle is using appropriate methods per evolution stage. It means to build components in Genesis, and custom built in-house, preferably using Agile methods. Or using or buying off-the-shelf products, or using open source software for components in product and rental evolution stage with preferably lean methods. Or outsourcing components and commodity to utility suppliers, or using preferably Six Sigma methods.

Coming back to our online school, we are going to develop the online school component in-house. For the infrastructure components such as search engine, data storage, message broker, and so on, we are using open source solutions at the current state. The virtual machine component is provided by a server hosting provider as an off-the-shelf product. With a Wardley Map at hand, we are already applying doctrinal principles. One of the doctrinal principles is that we know who our users are. We know their user needs and we can focus on them as another doctrinal principle. We know the details in terms of what is necessary to fulfill the user needs and generating value by solving the user needs. It also depicts what to build in-house and where to use off-the-shelf products, or open source software, and where to outsource to utility suppliers. We can share and discuss this map with others to challenge assumptions as another doctrinal principle in order to create a better map and a better understanding, and using a common language to be able to communicate effectively. We are going to use this map as a foundation for future discussions while evolving our legacy system.

Optimizing for Flow of Change from Team Perspective

Let's go back to the team perspective. To optimize for flow of change, from a team perspective, we need to avoid functional silos with handover. Instead, we need to aim for autonomous cross-functional teams that are designing, developing, testing, deploying, and operating the systems they are responsible for. We need to avoid handover so that work is never handed off to another team. We need to use small long-lived teams as the norm. The teams need to own the system or subsystem they are responsible for so they need to have end-to-end responsibility to achieve fast flow. We need to reduce the team's cognitive load. If the team's cognitive load is largely exceeded, it becomes a delivery bottleneck leading to delays, quality issues, and so on. While the communication within a team is highly desired, we have to restrict a high communication bandwidth between the teams to enable fast flow.

Four Team Types of Team Topologies

That's where team topologies can help us with, with their well-defined team types and their well-defined interaction modes. We have the autonomous, cross-functional, stream-aligned teams that are aligned to a continuous stream of work, focusing on fast flow of changes. To be able to focus on a fast flow of changes and to be able to produce a steady flow of feature deliveries, they need the help from the other teams. For example, they need support from the platform teams that are responsible for platforms that usually abstract away infrastructure and networking or cross-cutting capabilities. They provide internal self-service services and tools for using that platform that then the stream-aligned teams can easily consume. Or they need the help from the enabling team that help the stream-aligned teams to acquire missing capabilities. Or the complicated subsystem teams as an optional team type that are supporting the stream-aligned teams on particularly complicated subsystems that require very specialized knowledge. They all aim for increasing the autonomy and reducing the team cognitive load of the stream-aligned teams to enable fast flow of change.

Three Interaction Modes

To arrange the teams into the team types, it's not enough to become effective. How these teams are interacting with each other and when to change and evolve team interaction is also very relevant for high effectiveness. With the interaction mode of collaborations, teams are working very closely together, aiming for a common goal. This interaction mode is suitable for rapid discovery and innovation, for example, when exploring new technologies. X-as-a-Service suits well when one team needs to use a code library or a component, an API or a platform that can be effectively provided by another team as a service, in a workspace where predictable delivery is needed. Facilitating, this is the interaction mode that comes into play when one team would benefit from active help of another team. This interaction mode is typical for enabling teams. The combination of this well-defined team types and this well-defined interaction modes promotes organizational effectiveness. While we are applying team topologies, it also helps automatically to apply Wardley's Doctrinal principles.

Architecture for Flow

Let's come back to our previously created Wardley Map of the online school. Optimizing for fast flow of change requires to know where are the most important changes in a system [inaudible 00:16:41] the streams of changes? The type of streams can differ in every organization, ranging from task, role, activity, geography, and customer segment oriented stream types. In the current online school example, we are focusing on activity streams represented by the user needs of a Wardley Map, and the user needs of creating course content, planning classes, evaluating students, and so on, they are good candidates for activity oriented streams of changes. They need to be focused on when optimizing for fast flow. Let's address next the problem domain. The user and the user needs of our Wardley Map are representing the anchor of the map, but are also representing the problem domain in the domain-driven design context. Understanding the problem domain and partitioning it into smaller parts, the subdomain could be then the next step. That's where I would like to bring in then the next perspective of software architecture and design with domain-driven design.

Domain-Driven Design (DDD)

Domain-driven design is about designing software based on domain models, first proposed by Eric Evans. It comes with a core statement, in order to build better software, we have to align its software design with the business domain with the business needs and the business strategy. While applying domain-driven design, it also helps to apply doctrinal principles of Wardley Mapping.

DDD and Wardley Map

Domain-driven design comes with patterns and practices of strategic and tactical design. In the problem space, strategic design, we're analyzing the problem domain and distilling the problem domain into smaller parts, the subdomains. Not all subdomains are equal, some are more valuable to the business than others. We have different types of subdomains such as core, supporting, and generic. The core domain that is the center part of our problem domain providing competitive advantage, and it should be hard for competitors to copy or imitate. It's supposed to be complex, and it tends to change often. That's where we have to strategically invest most and innovate on. That's the subdomain we need to build in-house and supposed to go into Genesis or custom built evolution stage.

The supporting subdomain helps to support the core domain. It does not provide any competitive advantage. It's quite simple. They do not change often. If possible, we should look out for buying off-the-shelf products or using open source software that goes then in the product rental evolution stage. If this is not possible, and if we have to custom build the supporting subdomain, we should not invest heavily in that part of the system. The generic subdomain, these are subdomains that many business systems have, for example, authentication or payment, so they aren't core, they provide no competitive advantage, but businesses cannot work without them. They are generally complex but already solved by someone else. Buying off-the-shelf products, or using open source software or outsourcing to commodity suppliers should be applied for the generic subdomains. There is no need for innovation here. In general, these different subdomain tabs can help us to prioritize the development effort that we should invest most in the core domain providing competitive advantage.

In regards to our online school, the user needs of creating course content, planning a class, learning support, and studying courses, they fall into the result of the core domain. They provide competitive advantage, leading to a high level of differentiation. They embody complexity, and they tend to change often and are not widespread in every other business system. It is the core domain that we will want to strategically invest in most, and that's where we will have the most opportunity for innovation. On the other hand, the evaluation of student progress and learning does not necessarily provide a competitive advantage, but it supports the teachers experience and is necessary for the organization to succeed. It tends to be less complex and change less frequently, so we should not invest too much in that subdomain since it does not provide competitive advantage.

The user needs of signing up and signing in, they embody user needs of a generic subdomain, and many business system needs the subdomain of registration and authentication. They are widespread, standard, and expected with a high level of ubiquity and a high level of stability. They are not core, do not provide competitive advantage, but without registration and without authentication, the online school solution would not work securely. They are generally complex, but already solved by someone else. There exists several solutions on the market already, so there is no need for innovation and strategic investment here.

Bounded Contexts

Next, we can switch to the solution space of strategic design. That's where we are going to decompose our monolithic big ball of mud into smaller modular components, designing then the bounded context. A subdomain, in the ideal case, can contain one or in some cases also more than one bounded context. A bounded context defines where a single domain model can be applied and forms a unit of purpose, mastering, and autonomy, and a domain model itself that represents the domain logic and business rules that are relevant to that area of the system. A bounded context can provide different types of boundaries. It can form linguistic and semantic boundaries so that the languages' terms of the domain model are only consistent in silos of its bounded context. It also serves as an ownership boundary. A bounded context should be owned by one team only. However, a single team can own multiple bounded contexts. A bounded context serves also as a physical boundary and can then be implemented as separate solutions. Not all bounded contexts need to share the same architectural and business logic implementation patterns, so they can differ from context to context.

In this example, we might have derived several bounded contexts with the appropriate methods, so content creation, class management, course studies, and learning support, they fulfill the core domain related user needs. They are strategically important and require most development effort. They go into the custom built evolution stage and are going to be built in-house. Student evaluation notification handling belongs to the supporting subdomains. There might exist solutions on the market already, but the team decided that a higher level of specialization is needed to build them in-house. However, the development investment should not be too high since they do not provide competitive advantage. The Identity and Access Management bounded context belongs to a generic subdomain. There exists several solutions on the market already, and it should go either into product and rental or commodity and utility evolution stage.

Next is finding suitable team boundaries. As I said earlier, the bounded contexts serve as well-defined ownership boundaries forming a unit of purpose, mastery, and autonomy. Bounded contexts are indicating what teams to split the system into smaller parts. On the other hand, they also work as suitable team boundaries for the stream-aligned teams. To optimize cognitive load, we have to limit the number and size and complexities of software system a single team has to work with. This brings us to the evolution stages of Wardley Map. According to the characteristics of the evolution stages, the further left bounded context or a component in general is located on the Wardley Map, the higher the level of uncertainty is, which requires then practices that are focusing on exploration and discovery.

Bounded context in Genesis or components in Genesis tend to change far more frequently than components in commodity. For components residing in commodity, we can draw on best practices solutions providing a clear path to action. Each evolution stage addresses different practices for their related paths to action, ranging from best, good, emerging, and novel practices. The more unclear the path to action is, the higher the level of cognitive load is. That means that a single team could take ownership of more or larger bounded context and components that are residing in commodity than residing in Genesis. Another aspect to optimize for team cognitive load is to establish clear responsibility boundaries. For clear responsibility boundaries, we need to assign each bounded context to a single team and not sharing bounded context across teams, because that would then diffuse the ownership of teams. However, one team can own several bounded contexts.

Architecture for Flow

The next step is identifying services that are needed to support a reliable flow of changes. That brings the focus on the infrastructure and platform related components of our Wardley Map located in product and rental, and commodity and utility evolution stage. In order to be able to focus on a fast flow of change, the stream-aligned teams are relying on a platform team providing an easily consumable platform as a service. This platform shall reflect thinnest viable platform that is just big enough to fulfill the user needs of its consumers and shall not become bigger than needed. The thinnest viable platform could start with documentation, standards, best practices, templates, and can evolve later into a digital platform with self-service APIs and tools. We also need to mind the dependencies and communication bandwidth between teams. Are there tightly coupled or blocking dependencies, or a high amount of ongoing communication and coordination efforts between teams? They need to be eliminated. Identifying capability gaps and helping teams to acquire missing capability is then the responsibility of the enabling teams.

The previous consideration might result in this illustrated team constellation. In general, most teams in an organization shall be cross-functional, autonomous, stream-aligned teams. In our example, the four core domain related bounded contexts residing in the custom built evolution stage, they are going to be split among three stream-aligned teams. The supporting and the generic subdomain related bounded context of this example are going to be handled then by another stream-aligned team. The infrastructure components then will be taken care of by one or more platform teams.

Before we are going into the transition and implementation phase, we need to mind potential efficiency gaps. We want to avoid building or evolving the system based on inefficiencies, because inefficiency would also impede the fast flow of change. If the evolution stage of the components we are using in our organization internally differ from the ones that are available in the market, that might indicate an efficiency gap. For example, if we're using custom built components, but the components in the market are available as commodity, we are lagging behind and encounter potential inefficiencies. The bigger the gap, the less efficient that organization might be. We need to close this efficiency gap. One approach to close the efficiency gap could be done by migrating components from on-premises infrastructure to the cloud. That's what we are going to address.

Implementing flow Optimization

At this moment, we have not transitioned anything yet, but we did prior architectural and design considerations. Now the question arise, how do we transition and implement flow optimization? To close the previously mentioned efficiency gap, we want to migrate the components from on-premises infrastructure to cloud hosted services. The cloud migration could start with a team first approach. The cloud migration is going to be supported through dynamic reteaming and evolution of team topologies. Since we are addressing infrastructure components, we are starting the transition with forming the platform team first from members of the previous backend and the previous infrastructure team. This new platform team is working on the site with the freedom to learn and explore and discover and assess cloud migration strategies.

This new platform team can then determine a migration plan and its new target platform. In the example of the online school, we are starting with the replatforming cloud strategy. With replatforming, we are modifying the underlying infrastructure by replacing on-premises infrastructure components with cloud managed services, while we keep the applications architecture the same. With replatforming, we are offloading the management of the infrastructure components to the cloud provider and reducing then the operational efforts on our side and also improve performance scalability and elasticity of the infrastructure itself, and we are going to close efficiency gaps. Once the on-premises infrastructure is fully migrated to the cloud environments through replatforming, the remaining infrastructure team and the previous platform team can merge into a joint platform team. Even though the resulting team is becoming bigger, it makes sense in some cases to merge teams, when we want to bring collective intelligence together, and when the joined teams leads to a higher flexibility and to higher fluidity than having separate teams. For example, in our case, the merged team is able to share their previously gained cloud knowledge from replatforming then with the new team members, for example, by pair programming. The merged platform team can split later then in smaller teams if this is applicable.

So far, we have replatformed our infrastructure components, but we also want to modernize our monolithic big ball of mud, and eliminate its related bottlenecks. That brings us to the next level of cloud migration by refactoring the applications architecture. Refactoring typically involves breaking down the application into smaller components. In the example of the online school, we want to break the monolithic big ball of mud part by incrementally decomposing it into our previously designed bounded contexts. To start the refactoring journey, we are forming first stream-aligned teams from members of the previous frontend and backend teams.

The refactoring journey is supported by an evolution of team topologies interaction mode. At the beginning, the stream-aligned teams, they can closely collaborate with the platform team to jointly discover and assess potential cloud options for the future bounded context that the new stream-aligned team is responsible for. The stream-aligned team might decide to refactor their bounded context using serverless technologies. As soon as potential options were discovered and knowledge was shared, the interaction can go then into limited collaboration and limited facilitation mode, so going from daily to occasional on-demand interaction. At a later stage, this can then evolve to X-as-a-Service once the serverless ecosystem is more established in the organization, where then the platform team provides best practices, standards, tools, APIs, to easily consume the cloud services.

In the meantime, the remaining frontend and backend teams can merge into a preliminary joined team, which takes care of the remaining monoliths, as long as it's existing, and as one team with no handover involved. From there, they will then incrementally split into their separate stream-aligned teams. These new formed stream-aligned teams, they receive active help from the previous stream-aligned and platform team during their cloud journey. The cloud platform team can suggest possible cloud options that are applicable for their bounded contexts. They can coach them how to use a new build pipeline and a serverless ecosystem. They can also point to the platform as a service it has established before and incorporate their feedback, why they are using it. The previous stream-aligned team can share their knowledge with the new stream-aligned teams. For example, coaching them about best practices or how to implement serverless functions using hexagonal architecture, for example, so that at the end, we have refactored our monolithic big ball of mud into modular components based on serverless technologies.

What to Adopt and What to Leave Behind

At this stage, we have transformed our functional silo teams with handover into well-defined team types and interaction modes, and could restrict the formerly high communication and coordination bandwidth between teams. We could decompose our monolithic big ball of mud with a messy model and fuzzy boundaries into a system with clear bounded contexts and domain models. With bounded context, we can establish well-defined ownership boundaries. We moved from tight change coupling to a loosely coupled system and optimized the team cognitive load, so that we could eliminate delivery bottlenecks and improve the software delivery performance. In addition, we could reduce the high operational efforts by migrating on-premise infrastructure components to cloud hosted services, and close then the related efficiency gap.

Key Takeaways

At the end, the combination of Wardley Mapping, domain-driven design, and team topologies that helps to understand the environment an organization is operating and competing in. It also helps to understand the climatic patterns that are impacting the landscape. Also helps to apply doctrinal principles that enables an organization to be able to respond to changes quickly. The combination helps us to gain domain knowledge and discovering the core, and providing competitive advantage. It helps us to know what components to build in-house, where to buy or use off-the-shelf products or open source software, or where to outsource to utility suppliers. It helps us to decompose the problem domain into modular components. It helps to align the teams and evolving their interaction to the system we build and the strategy we plan.

With this combination, we are able to identify potential efficiency gaps and trying to eliminate them, close these efficiency gaps. It helps us to eliminate bottlenecks and increase the software delivery performance so that we are then able to respond to changes quickly and absorb changes gracefully. The combination of Wardley Mapping, domain-driven design, and team topologies, that's providing a powerful, holistic toolset to design and to build and evolve adaptive socio-technical systems that are optimized for fast flow of change with a focus on improving the performance of the system as a whole.

Questions and Answers

Skelton: What feels like a good way to start with this approach? Can you provide some advice for people who are, let's say they've seen the talk, they can see the value of Wardley Mapping, domain-driven design, and team topologies together. What feels like a sensible approach to start at, let's say, an organization that has 1000 people, relatively small, and maybe like 600 engineers in total, that kind of size? Maybe they're doing something, they've got some online service or something like that, where will be a good place to start to use these techniques? Then, how would you go about doing it?

Kaiser: It can start differently. For example, what I like to start with is addressing the current state where you're at right now. What are your current challenges that you're dealing with. You can start with analyzing, with asking your teams as you're also suggesting in your team topology book, where you address the team situation in regards to the team cognitive load, and communication and coordination efforts between teams. You can survey style, for example, that you ask every team regarding the challenge that they're facing. Another aspect that you can bring in is to create a Wardley Map of your current situation, the Wardley Map that is representing the environment an organization is operating and competing in.

Every team itself can create their own Wardley Map, but they also can share their method in discovery with others. It is greatest value to the communication that you have between, because that is, for example, your view of your landscape that your organization is competing in, but sharing it with others, and communicate with others and also see blind spot that they have not seen before. It's really a great tool to know where you are. What is your current landscape where you're operating and competing in? What are your challenges of your teams regards to communication and coordination effort? Also, what is necessary? You could follow, for example, also, when you create this Wardley Map, the value chain. You can also identify, what is necessary? What components do you have to touch, and what teams are responsible for these specific components? Which teams need to communicate and coordinate to make this change effective? We can use this Wardley Map's foundation to identify bottlenecks.

Skelton: It feels like a really key aspect to empower the teams, to give them these tools, like being able to use Wardley Mapping, think about things in terms of bounded context from DDD, and obviously use the ideas from team topologies, about team interactions and limiting cognitive load and so on. To empower these teams to use these techniques as part of their toolkit, so that they can have much more informed conversations, much better conversations and share that with other teams. Say, "From our perspective, it looks like this. What does it look like from your perspective?" We're not trying to find a single perfect map or view, we're looking for multiple perspectives to give us additional insights effectively, because no one group is going to have all of the knowledge and the awareness.

Kaiser: No, then it's going to be not only one map. If you have a larger organization, then, for example, if they have platform teams that are providing the platform, they have a totally different Wardley Map than those that are end-to-end responsibility, or close to the customers, the stream-aligned team. They have different users, with their different user needs, and the platform team can create their own Wardley Map, where you as the stream-aligned teams are then the internal users, and you have then the user needs of designing and deploying and building your application. Where you then also need help, support from enabling teams regards to application security, testing, the software design, and so on.

There could be also from the enabling team a different Wardley Map as well, where you then can focus on your specific users of the Wardley Map and you can put them then together there. For example, the Wardley Map for the platform team. Platform as a service, for example, could be an aggregated component in the Wardley Map of the stream-aligned team. It also helps you to gather, for example, if you have multiple platform teams, if you lay them on top, then you can also identify whether you are using duplicated components, for example, for CRM systems. Then you can also see, it visualizes also duplications that you can reduce it then to one component, for example, if that makes sense, in your context.

Skelton: It feels like these things are definitely heading in the direction of, we're not looking to create a perfect machine. We're looking to have an ecosystem in those different perspectives. We're not going to be able to design this thing with just one perspective. That feels like an important industry trend.


See more presentations with transcripts


Recorded at:

Sep 13, 2022