Finding good stream boundaries with Independent Service Heuristics — Team Topologies

“Could this thing be run as a cloud-hosted (SaaS) service or product?”



Decide where to focus the attention.

Essentially we just need to choose an area of the business that needs to be represented in software.

This could be a user journey, a “product”, a possible business domain, a software service, an entire software application, a set of tasks for a single user persona, a possible value stream, etc.

Footprints Tours offers ‘alternative’ walking tours of cities exploring their social and cultural history.

They provide both guided and self-guided tours and have implemented a monolith website and mobile application to serve all of their customer needs.

The first step would be to capture some possible fracture planes.

Once a candidate domain, service, application, or value stream has been identified, the next step is to go through a series of questions to identify whether we have found a good candidate for being a separate stream of change.

• Is this thing independent enough?
• Would consumers understand or value it?
• Would it simplify execution?

• Would it be a viable business (or "micro-business") or service?
• Would it be a compelling offering?
• Could a marketing campaign be convincing?

• Would it be a viable business (or "micro-business") or service?
• What would a subscription payment include?
• Is there a clearly defined customer base or segment?

• Are the full costs of running this thing transparent or possible to discover? Consider infrastructure costs, data storage costs, data transfer costs, licence costs, etc.
• Is the thing too interconnected with other things in the organization? Or fairly separate?
• Does the organisation track this separately?

• Is the thing dependent on lots of data from multiple sources? Or fairly independent?
• Are the sources internal (under our control) or external?
• Is the input data clean or messy?
• Is the input data provided in a self-service way? Can the team consume the input data "as a service"?

• Is the thing meeting specific user needs?
• Do we know (or can we easily articulate) these user types and their needs?

• Would the cognitive load (breadth of topics/context switching) be bounded to help the team focus and succeed?
• Would significant infrastructure or other platform abstractions be needed?

• Is this thing logically independent from other things?
• Could the team "self-serve" dependencies in a non-blocking manner from a platform?

• Is the scope big enough to provide an impact? Would the scope be engaging for talented people?
• Is there sufficient value to customers and the organization that the value would be clearly recognized?

• Does this thing provide discrete value in a well-defined sphere of execution? 
• Can the team define their own roadmap based on what they discover is best for the product and its users or is the team always driven by the requirements and priorities of other teams?