Consulting Delivery

Consulting Delivery

Use When

• Use when research must become consulting-grade diagnosis, recommendation, workplan, or transformation framing.
• Use when the deliverable needs issue trees, hypotheses, quantified options, or implementation logic.

Do Not Use When

• The task is purely academic writing or pure source collection with no client decision attached.
• The work only needs prose polish rather than consulting problem-structuring.

Consulting Intake Guidance

• Client or decision-owner question.
• Time horizon and constraints.
• Expected deliverable and implementation ambition.

Consulting Method Detail

• Read this SKILL.md first, then load only the relevant reference file for the current phase.
• Structure the problem before expanding the workplan.
• Tie analyses and recommendations back to the client decision throughout.

Quality Standards

• Decision owner and question are explicit.
• Issue tree and hypotheses are visible.
• Recommendation includes value, risk, and implementation implications.

Consulting Failure Notes

• Workstreams with no governing client question.
• Recommendations with no owner or sequence.
• Treating stakeholder politics as outside the analysis.

Consulting Deliverable Detail

• A consulting-grade problem frame, workplan, recommendation structure, or implementation path.

References

• Use the references/ files for problem framing, workplans, stakeholders, and transformation logic.

This skill turns a research corpus into a consulting-grade decision process. It is about how to structure the problem, sequence the analysis, manage the client interface, and carry recommendations into implementation.

Consulting output must be more than polished slides. Run critical-reasoning-and-argument before finalizing issue trees, hypotheses, options, recommendations, value cases, risks, or implementation logic.

When to use

• Strategy, market-entry, growth, transformation, or operating-model work
• Client-facing diagnostics or recommendations
• Internal decision memos that must survive executive scrutiny
• Research that must lead to quantified choices, implementation options, or a value case

Five rules

1. Start from the client decision. The problem statement must name the decision, owner, and time horizon.
2. Structure before analysis. Use issue trees, hypotheses, and workplans before gathering endless facts.
3. Quantify the value and the risk. A recommendation without upside, downside, and assumptions is unfinished.
4. Manage stakeholders explicitly. Good analysis fails when the buyer, blocker, and user were treated as the same person.
5. Implementation begins in the diagnosis. Recommendations must name who does what, in what sequence, and what proves progress.
6. Reason before recommendation. Every recommendation must pass critical-reasoning-and-argument: argument map, strongest objection, alternatives, evidence sufficiency, and implementation constraints.

Router

Core workflow

1. Define the governing client question.
2. Break it into an issue tree and draft initial hypotheses.
3. Design the analysis plan: what must be true, what data proves it, what would falsify it.
4. Sequence the work into a workplan with owners and decision points.
5. Translate findings into options, value case, risks, and implementation path.

Ship gate

• Decision owner, question, and time horizon named
• Issue tree or equivalent decomposition exists
• Critical-reasoning gate passed for hypotheses, options, recommendations, risks, and implementation implications
• Hypotheses are testable, not slogans
• Analyses are tied to decisions, not curiosity
• Stakeholders mapped beyond the formal sponsor
• Recommendation includes assumptions, risks, and implementation path
• Metrics or KPIs identified for tracking progress

Consulting Source Pitfalls

• Research with no client decision attached
• Workstream explosion because the problem was never decomposed
• Elegant recommendation with no owner or sequencing
• Treating stakeholder politics as outside the scope of strategy
• Confusing slide polish with problem solving

Companion skills

Inputs

Capability Contract

Analysis and planning default to read-only. Client commitments, messages, implementation changes, procurement, spending, or publication require explicit authority.

Consulting Fallback Notes

Without client access or implementation data, return a hypothesis-led workplan with assumptions and untested branches, not a transformation claim.

Decision Rules

Consulting Scenario

A cost problem is decomposed into testable drivers before interviews or recommendations are scheduled.

Companion skills

• research-orchestration — drives the research waves
• critical-reasoning-and-argument — validates hypotheses, issue trees, options, recommendations, value cases, and implementation logic
• analytic-tradecraft — for uncertainty, alternatives, and confidence
• executive-communication — for answer-first client delivery
• report-and-proposal-craft — for the long-form report or proposal shell
• data-visualization — for charts, tables, and KPI trees

Sources for this skill

• Rasiel, Ethan M., and Friga, Paul N. The McKinsey Mind. Tier 1.
• Burtonshaw-Gunn, Simon. Essential Tools for Management Consulting. Tier 1.
• Inside the Minds: Leading Consultants. Tier 1.
• Hattori, Shu. The McKinsey Edge. Tier 1.
• Rewired: The McKinsey Guide to Outcompeting in the Age of Digital and AI. Tier 1.

Workflow

1. Define the client decision, stakeholders, constraints, and measurable outcome.
2. Build an issue tree and convert branches into testable hypotheses.
3. Stop when a recommendation lacks evidence, ownership, dependencies, or feasibility.
4. Recover by returning to the failed branch, recording the gap, and revising the workplan.
5. Translate supported findings into actions, owners, risks, and decisions.

Outputs

Evidence Produced

Degraded Mode

Without client access or implementation data, return a qualified workplan, preserve unassessed branches, and do not claim transformation feasibility.

Anti-Patterns

• Jumping to a solution. Fix: frame the decision and issue tree first.
• Keeping a disproved hypothesis. Fix: close or revise the branch.
• Producing ownerless recommendations. Fix: assign an accountable role.
• Hiding evidence gaps. Fix: expose them in the trace.
• Promising implementation without context. Fix: mark feasibility unassessed.

Worked Example

A cost problem is decomposed into testable drivers, and only branches supported by verified analysis proceed to implementation recommendations.