Organizing as Value Streams
This domain tests whether you can connect value stream design to organizational design: team boundaries, ownership, cognitive load, and funding.
From handoffs to stream-aligned teams
A value stream needs a team, or a small set of teams, that owns it end to end and is long-lived enough to learn. Functional silos - front end, middleware, DBA, QA - each optimize their own utilization and hand work over the wall, so lead time is dominated by waiting. A value stream-aligned team is cross-functional, owns a bounded part of the flow, can deliver without depending on other teams' queues, and has a size and scope that respect cognitive load.
Team Topologies vocabulary
Four team types: stream-aligned (primary), enabling (coaches capability into streams temporarily), complicated-subsystem (specialist deep work), and platform (provides self-service internal services). Three interaction modes: collaboration, X-as-a-service, and facilitating. The exam expects you to match team type to need rather than defaulting everything to stream-aligned.
Ownership, roles and funding
Each stream needs a named owner accountable for flow and outcomes - commonly a value stream owner or lead - plus a product manager for scope and technical leadership for design. Funding matters as much as structure: project funding creates temporary teams, handoffs, and an incentive to start work rather than finish it, while persistent value stream funding lets teams own a product over time. Conway's Law means architecture and team boundaries will mirror each other, so design both deliberately.
Worked example
A feature crosses front-end, middleware, DBA and QA, taking nine weeks with four handoffs and a shared DBA queue. The reasoning for restructuring: the waits come from the dependency structure, not from slow individuals, so adding people changes little. Two stream-aligned teams are formed around bounded contexts, a platform team offers database-as-a-service and a golden CI pipeline via X-as-a-service, and an enabling team coaches automated testing. Handoffs and the shared queue disappear, and lead time falls because waiting was removed.
Traps that cost marks
- Renaming existing silos value stream teams without changing ownership or dependencies.
- One team per component or per layer instead of per stream.
- Streams so large that cognitive load, not capability, becomes the limit.
- Keeping project-based funding and temporary teams, which recreates handoffs.