Delivery disciplines

Clarify the decisions behind delivery roles.

A useful team structure explains which decisions each discipline owns and where collaboration is required.

Explore the referenceDownload PDF
On this page

Delivery responsibility map

DisciplinePrimary contribution
Product and analysis
Priorities, workflow understanding and acceptance criteria
Design
Information structure, interaction and inclusive task completion
Engineering
Maintainable implementation and technical boundaries
Quality and operations
Evidence, release confidence and recoverability

Assign decisions, not only job titles

The same person may cover several disciplines in a small team. The important point is that product priority, technical authority and acceptance are not left ambiguous.

Make handoffs explicit

A design handoff includes error states. A release handoff includes operating context. A support handoff includes the conditions that require escalation.

Define contributions rather than job titles

A delivery team needs decisions about priorities, user experience, architecture, implementation and acceptance. Make those decisions visible even where a small team combines roles. The product owner should be able to explain which business outcome matters next. Design should make the task understandable. Engineering should identify implementation constraints and maintainable ways to address them.

Agree who can accept work and what evidence they need. A developer completing a change does not automatically establish that the resulting workflow meets the business requirement. Identify the people who understand the operation and involve them early enough to correct assumptions. Avoid making one person a bottleneck for every routine decision.

Keep work small enough to inspect

Break delivery into increments that demonstrate a useful behaviour. For a customer portal, one increment might let a permitted user view the status of an existing request. This brings identity, data access and interface behaviour together without requiring every future feature. Demonstrate the increment with realistic sample records and explain any part still simulated.

Maintain a shared view of unfinished work, dependencies and decisions needed. Review blocked items explicitly instead of allowing them to remain hidden inside a progress percentage. When new information changes the scope, describe the effect on the next delivery choice. A predictable cadence is helpful, but its value depends on showing meaningful work and resolving uncertainty.

Make quality a shared responsibility

Agree the checks expected before work is accepted. These may include business-rule tests, access checks, accessibility review and deployment verification. Assign responsibility for preparing representative examples and for maintaining them when the workflow changes. Quality should not become a final queue owned by a team that lacked context during implementation.

Use reviews to identify recurring sources of rework. If requirements repeatedly change late, improve how the team tests assumptions. If releases repeatedly fail in one environment, inspect configuration and deployment controls. Keep the discussion tied to observable work rather than individual blame. The result should be a clearer delivery system with responsibilities people can actually fulfil.

No. Match capability to the work and keep decision ownership clear, even when roles are combined.