VSM-01] Run a Value Stream Mapping exercise for the Goald delivery workflow

[paruff]

[VSM-01] Run a Value Stream Mapping exercise for the Goald delivery workflow

Labels: process, ai-ops, p1-critical, effort-medium
Assignee: 👤 Human
Phase: Phase 1 — AI Foundation
DORA 2025 Source: Value Stream Management chapter; AI Capabilities Model companion report
Effort: Medium (half-day workshop + 1 day documentation)
Blocks: VSM-02, METRICS-02

---

Why This Is Now Critical (2025 DORA Finding)

The 2025 DORA report identifies Value Stream Management (VSM) as the practice that turns AI’s individual productivity gains into organisational advantage. Without it, AI creates local optimisations that pile up work downstream — you ship code faster but it stacks up in review, QA, or deployment queues rather than reaching users.

For Goald specifically: Copilot can generate a screen in 20 minutes. If the bottleneck is actually the review→merge→deploy pipeline, that speed is absorbed by the system and never reaches users. VSM makes the real bottleneck visible.

“VSM acts as an AI force multiplier. By visualising your flow from idea to customer, you can identify where work waits and where friction exists.” — DORA AI Capabilities Model report, Dec 2025

---

What Value Stream Mapping Means for a Small Team

VSM does not require a week-long enterprise workshop. For a solo/small-team app, it is a 2–3 hour exercise that maps:

1. Trigger — a GitHub issue is created
2. Design — human writes acceptance criteria and BDD scenario
3. Implementation — Copilot generates code
4. Review — human reviews diff against GITOPS-04 checklist
5. CI — lint + typecheck + test gates run
6. Merge — feature lands on main
7. Deploy — Expo OTA or EAS build reaches users
8. Feedback — user analytics or support contact surfaces outcome

For each step, record: average wait time, average active time, and the most common failure/rework reason.

---

Acceptance Criteria

The Map

• docs/VALUE_STREAM_MAP.md created with a text-based or Mermaid diagram of the Goald delivery flow (Issue → Code → Review → CI → Merge → Deploy → Feedback)
• For each step, document:
  • Estimated average elapsed time
  • Estimated average active (hands-on) time
  • Primary rework/failure reason observed in the last 10 PRs
• Bottleneck identified and highlighted — the single step with the highest wait-to-active ratio
• AI insertion points annotated — where Copilot adds value and where it currently creates friction

The Actions (captured as GitHub issues)

• At least one follow-up issue created per bottleneck identified
• If the bottleneck is CI speed: file an issue to optimise the CI workflow
• If the bottleneck is review quality: update GITOPS-04 checklist
• If the bottleneck is deployment: escalate DEVOPS-03 (EAS Build) priority
• VSM is reviewed monthly alongside the PROCESS-02 copilot-instructions tuning

Mermaid Template to Populate

flowchart LR
    A[Issue Created\nWait: ?h\nActive: 15min] --> B[Spec Written\nWait: ?h\nActive: ?h]
    B --> C[Copilot Implements\nWait: 0\nActive: ?min]
    C --> D[Human Review\nWait: ?h\nActive: ?min]
    D --> E[CI Gates\nWait: 0\nActive: 3min]
    E --> F[Merge\nWait: ?h\nActive: 5min]
    F --> G[OTA Deploy\nWait: 0\nActive: auto]
    G --> H[User Feedback\nWait: days\nActive: passive]

    style D fill:#FEF3C7
    style A fill:#DBEAFE

Loading

---

Files to Create

• docs/VALUE_STREAM_MAP.md
• Any follow-up issues raised as a result of bottleneck analysis