Stakeholder Mapping

Stakeholder Mapping

> Part of the discovery-phase skill pack · discovery group · reads discovery-context.md (run profile-builder first if missing).

Two-tier map for outsourcing reality: client stakeholders + agency stakeholders. The agency tier is the one most often forgotten, and it's where engagement-killers (sales conflict, delivery capacity, partner dependencies) live.

Step 1 — Read discovery context

Read discovery-context.md (sections 1. Client → Decision-makers, 3. Engagement → Runner role) and problem-canvas.md if it exists (anchors stakeholders against the actual problem).

If discovery-context.md is missing, ask the BA inline: "(a) name 1–3 client decision-makers + role; (b) engagement mode (pre-sale / paid discovery / delivery / staff aug)?" — tag any unverified stakeholder field as [ASSUMED]. Never block; recommend profile-builder for high-stakes work.

Step 2 — List client-side stakeholders

Identify 3-6 people on the client side. Cover these roles (combine if one person fills multiple):

• Decision-maker — whose approval ships this
• Budget owner — who pays
• Delivery owner (client side) — who'll integrate the result into their org
• Impacted teams — whose work is touched (CS, sales, ops, security, legal)
• Domain experts — compliance, security, accessibility — needed for sign-off
• A sceptic — the person most likely to block. Always include. If everyone is supportive, the map is incomplete.

Step 3 — List agency-side stakeholders

Often skipped. Typical roles:

• Engagement lead / account manager — owns the client relationship
• Delivery owner (agency side) — would lead build if engagement converts
• Discipline leads — engineering, design, QA — capacity gates
• Sales / partnerships — if pre-sale, has commercial pressure
• Domain SME — if the domain (healthcare, finance) requires specific expertise
• Finance / ops — for fixed-fee viability

Step 4 — Per-stakeholder fields

For each stakeholder (both tiers):

| Field | Purpose | |---|---| | Name + role + tier (client/agency) | | | Their stake | What they win/lose from the outcome | | Their success metric | How they personally know this is worth it | | Their concerns | In their own words, ideally a quote | | Input needed (and when) | Which discovery phase needs them | | Decision rights | Approve / veto / advise / inform |

Step 5 — Surface conflicts

After both tiers are filled, scan the success-metric column. Conflicts often look like:

• Client sales wants "fast time-to-value" vs client security wants "SSO integration first"
• Agency sales wants "convert to delivery" vs agency engineering wants "scope this down or we'll miss deadlines"
• Client decision-maker wants "build it" vs client end-user manager wants "fix the workflow first"

Name conflicts on the map. Do not try to resolve them here. Resolution happens during proposal/SoW or in stakeholder workshops.

Step 6 — 15-minute async check-ins

Recommend (or send, if BA has the access) a 3-bullet async message to each high-stakes stakeholder:

1. The problem-canvas one-liner.
2. "Here's what I think your success criterion is — right?"
3. "Anything you'd kill this engagement over? Say it now, not in 4 weeks."

Capture responses in the map.

Step 7 — Set review cadence

Pick one based on engagement timeline:

• Pre-sale sprint (≤2 wks): no recurring review, ad-hoc sync per major artifact
• Discovery sprint (3-4 wks): weekly async update, doc-based
• Delivery phase: bi-weekly stakeholder readout
• Staff aug: mirror client's existing rituals

Output

./discovery/stakeholder-map.md per ./template.md.

Append to _log.md: [stakeholder-mapping | YYYY-MM-DD] client_tier: ; agency_tier: ; conflicts: ; cadence: .

Anti-patterns

• Skipping the agency tier. It's where the most fixable conflicts hide.
• No sceptic on the map. Find one. If you really can't, your discovery is happening inside an echo chamber.
• "We'll loop them in later." Stakeholders who learn about a project on week 4 veto it on week 4. Cheaper to absorb the veto on week 1.
• Static map. Update after major artifacts (insights, scope, prototype). Especially before any deliverable skill runs.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: stanislavnianko
• Source: stanislavnianko/product-discovery-claude-skills
• License: MIT

Install and usage instructions live in the source repository linked above.