Team Interaction Mode
The defined way two teams work together — collaborating, consuming a service, or facilitating — chosen deliberately and expected to change over time.
Team structure receives attention; the interactions between teams receive almost none, and they are where coordination cost actually accumulates.
The three modes:
Collaboration — high bandwidth, working closely, appropriate when an interface or an approach is still being discovered. It is expensive and should be explicitly temporary.
X-as-a-service — one team consumes another's capability through a stable interface with minimal interaction. Low bandwidth, and the correct steady state for most relationships.
Facilitating — one team helps another build capability, then withdraws.
The design instruction: collaboration should converge into X-as-a-service. A collaboration that never ends has become permanent coupling, and it is the most common failure this model names — two teams attending each other's meetings indefinitely because nobody defined the interface.
The second named failure: a platform team operating as a gatekeeper rather than as a service. It reintroduces the queue the platform was created to remove, and it usually happens because the platform is under-resourced relative to demand.
The underlying constraint is cognitive load — a team's ownership must fit what it can hold. That is what justifies most platform investment: not reuse, but reducing what each product team must understand to ship safely.