Supply Chain Mapping for Wardley Maps

You are conducting the supply chain identification phase of a Wardley mapping engagement. Your goal is to decompose agreed user needs into the capabilities and components that deliver them, producing a complete dependency graph.

Prerequisites

Check that the project directory contains:

• •needs/needs.agreed.md

Check that the client workspace contains:

• •resources/index.md and research sub-reports in resources/

If needs.agreed.md is missing, tell the user to complete wm-needs first. The .agreed.md suffix indicates the client has signed off.

The project path is clients/{org}/projects/{project-slug}/.

Step 1: Re-read research through the lens of needs

Read needs/needs.agreed.md to understand the agreed user classes and their needs.

Then re-read resources/index.md and relevant sub-reports in resources/, specifically looking for:

• •Capabilities the organisation uses to deliver each need
• •Technologies and platforms underlying those capabilities
• •External dependencies (suppliers, partners, data sources)
• •Infrastructure underpinning everything

You are building a mental model of "how does this organisation actually deliver what users need?"

Step 2: Decompose each need

For each agreed need, work top-down:

1. •What capabilities directly deliver this need?
2. •For each capability, what does it depend on?
3. •Continue recursively until you reach:
   • •Commodities (electricity, internet, standard hardware)
   • •External services with no further visible decomposition
   • •Fundamental resources (raw materials, data, people)

Write each chain to chain/chains/{need-slug}.md using the template in chain-template.md.

Formatting rules for dependency trees:

Need Name
├── Capability A
│   ├── Sub-capability A1
│   │   └── Infrastructure Component
│   └── Sub-capability A2
├── Capability B
│   └── Sub-capability B1
│       ├── External Service
│       └── Data Source
└── Capability C (commodity)

Guidelines:

• •Use the tree to represent visibility. Components near the top are visible to the user. Components deeper are increasingly invisible. This maps directly to the Y-axis (visibility) of the eventual Wardley map.
• •Do NOT attempt to position components on the evolution axis. That is stage 4 (wm-evolve). The horizontal axis has no meaning at this stage. Dependency trees are the honest representation — they encode what we know (depth/visibility) without pretending to know what we don't (evolution).
• •Flag shared components. When a component appears in multiple chains, note it with "(shared)" and list which other needs also depend on it. Shared components are structurally important.
• •Stop at sensible boundaries. Don't decompose "electricity" into "power generation -> transmission -> distribution." Stop where the organisation's agency ends, unless infrastructure is specifically in scope.
• •Use the organisation's language where possible. If they call it a "Fleet Management System" not "Asset Tracking Platform," use their term. The client needs to recognise their own supply chain.

Run chains in parallel where possible — each need's chain is independent.

Step 3: Identify shared components

After all chains are written, look across them for:

• •Shared components: same capability serving multiple needs. These are structurally important — they sit at critical junctions and often become strategic leverage points.
• •Implicit dependencies: components that are assumed but not explicit in any single chain (e.g. "identity management" underlying multiple systems).
• •Potential duplication: cases where two chains use different names for what might be the same component. Consolidate these.

Step 4: Synthesise

Write chain/supply-chain.md using the consolidated template in chain-template.md. Include:

1. •Merged dependency graph — all chains combined, deduplicating shared components
2. •Shared components table — each shared component with the needs it serves and its depth from the user
3. •Component inventory — a flat table of every unique component with its parents, children, and notes
4. •Observations — structural insights:
   • •Bottlenecks or single points of failure
   • •Surprisingly deep dependency chains
   • •Missing capabilities (things that should exist but weren't found)
   • •Clusters of related components
5. •Open questions — for client review

Step 5: Present to client

Present the consolidated supply chain to the client. Ask:

1. •"Does this dependency structure match how your organisation actually works?"
2. •"Are there capabilities missing from these chains?"
3. •"Are there capabilities listed that don't actually exist or are wrong?"
4. •"Do the shared components look right? Are there others?"
5. •Address any open questions.

The client will likely:

• •Correct component names to match internal terminology
• •Add capabilities you couldn't see from public research
• •Identify dependencies between components that weren't obvious
• •Flag components that are planned but don't exist yet

Step 6: Iterate and agree

Based on client feedback:

1. •Update individual chains in chain/chains/ as needed
2. •Rewrite chain/supply-chain.md
3. •Ask the client to confirm: "Is this supply chain now accurate and complete enough to proceed to evolution assessment?"

When the client agrees:

1. •Copy to chain/supply-chain.agreed.md
2. •Append to decisions.md:
   markdown

   ## {Date} — Stage 3: Supply chain agreed
   
   **Agreed**: Supply chain document signed off by client.
   **Components**: {total count} unique components identified
   **Shared components**: {list of key shared components}
   **Scope notes**: {any caveats}
   

Important notes

• •Output is markdown only. No OWM files at this stage. The evolution axis is unknown, so OWM would require guessing half the coordinates. Markdown dependency trees are the honest representation.
• •The .agreed.md file is a gate. The next skill (wm-evolve) will refuse to proceed without it.
• •Component count matters. A typical map has 15-30 components. If you have fewer than 10, you may be too coarse. If you have more than 50, you may need to split into submaps. Discuss with the client.

Completion

When supply-chain.agreed.md is written, tell the user the next step is wm-evolve to position components on the evolution axis and produce the first real Wardley map in OWM format.