Skip to content

JPeetz/agent-skills / skills/business-strategy/enterprise-consulting-methods

enterprise-consulting-methods

Verified

Use this skill when you need McKinsey/BCG-grade strategy work — problem decomposition, hypothesis testing, storyline building, decision memos, stakeholder mapping, synthesis, meeting prep, workshop design, and native PowerPoint chart production. Covers 15 consulting methodology patterns from Situation-Complication-Problem-Recommendation (SCPR) through MECE issue trees, hypothesis trees, Minto Pyramid writing, and the 4-agent deck pipeline. All chart outputs are native python-pptx objects, editable inside PowerPoint.

Enterprise Consulting Methods

A comprehensive meta-skill package containing 15 battle-tested consulting methodology patterns from a former McKinsey consultant. Each encodes a specific consulting move so you don't reinvent the structure every time. All deck/chart outputs produce native python-pptx objects — editable inside PowerPoint, right-click → Edit Data.


Overview

CategorySkillWhat It Does
FramingSCPR FrameworkStructure arguments using Situation-Complication-Problem-Recommendation
Issue Tree BuilderBreak down problems into MECE components
Hypothesis TreeCommit to a Day-1 answer and design tests that could kill it
AnalysisSynthesisTurn raw inputs into 3 insights with evidence + so-what
PrioritizationScore features using RICE, Impact/Effort, Value/Complexity
AI Use-Case ScorerScore AI use cases on Value × Feasibility × Safety
Data InsightsAnalyze datasets and answer key questions in plain English
WritingDecision Memo Builder1-page memo forcing a yes/no decision
Top-Down MemoAnswer-first writing using the Minto Pyramid Principle
Storyline BuilderBuild slide narratives where each line = a slide title
Meetings & PeopleMeeting Prep KitPre-read + agenda + talking points + objections
Stakeholder MapMap the room onto Power/Interest
Workshop DesignerDesign sessions that end with committed artifacts
DecksMcKinsey ChartsThree workhorse charts as native python-pptx objects
QualityMcKinsey CriticReview decks/docs like a McKinsey engagement manager

When to Use

Use this meta-skill whenever you need to apply structured consulting thinking to a business problem. Common triggers:

  • Strategic question → recommendation: Issue Tree → Hypothesis Tree → run tests → Decision Memo
  • Customer research → memo: Synthesis → Top-Down Memo (or Decision Memo) → McKinsey Critic
  • Board deck: Issue Tree → Storyline Builder → Charts → Deck Pipeline → McKinsey Critic
  • Big decision with politics: Stakeholder Map → Decision Memo → Meeting Prep Kit (for each 1:1)
  • Offsite or alignment workshop: Workshop Designer → Meeting Prep Kit (pre-1:1s) → Top-Down Memo (post-send)
  • Any high-stakes presentation: Use the 4-agent Deck Pipeline (Strategist → Builder → Critic → Fixer)
  • Analysis paralysis: Hypothesis Tree to commit to a Day-1 answer before boiling the ocean
  • Exec communication: Top-Down Memo for emails/briefs, Decision Memo for yes/no asks

How to Use

Quick Start — Full Deck Pipeline

Build me a McKinsey-style deck on [topic] for [audience]. Run the full top-down workflow:
1. Use issue-tree-builder to decompose into MECE branches
2. Use storyline-builder to turn it into 8 claim-shaped slide titles
3. Use mckinsey-charts for the data slides (bar+callout, waterfall)
4. Use deck-pipeline (strategist → builder → critic → fixer) to produce the final .pptx

Individual Method Calls

Each of the 15 methods below can be invoked individually. Pick the pattern that matches your current stage of work:

  • Framing the problem: SCPR Framework → Issue Tree → Hypothesis Tree
  • Understanding the data: Synthesis → Data Insights
  • Deciding what to do: Prioritization → Decision Memo
  • Communicating the answer: Top-Down Memo → Storyline Builder → Charts → Pipeline
  • Getting buy-in: Stakeholder Map → Meeting Prep Kit → Workshop Designer
  • Quality gate: McKinsey Critic on any draft before it ships

Workflow: Full Strategic Engagement

This is the canonical top-down workflow for turning a fuzzy strategic question into an actionable recommendation:

Phase 1: FRAME
  SCPR Framework → Issues Tree → Hypothesis Tree
  (Structure the problem, commit to a Day-1 answer)

Phase 2: ANALYZE
  Synthesis / Data Insights → Prioritization
  (Process raw data, rank what matters)

Phase 3: RECOMMEND
  Decision Memo Builder → Top-Down Memo
  (Force a yes/no, communicate answer-first)

Phase 4: COMMUNICATE
  Storyline Builder → McKinsey Charts → Deck Pipeline
  (Build narrative, create charts, produce deck)

Phase 5: STRESS-TEST
  Stakeholder Map → Meeting Prep Kit → McKinsey Critic
  (Map politics, prep meetings, get EM-level review)

Phase 6: EXECUTE
  Workshop Designer
  (Align the team, leave with committed artifacts)

Methodology Reference

1. SCPR Framework (Situation-Complication-Problem-Recommendation)

Purpose: Structure arguments for executive summaries, memos, and strategic recommendations.

Structure:

  • Situation: Current state of the market/business (baseline context)
  • Complication: Recent shift or change creating urgency (new dynamics, competitive threats)
  • Problem: Crisp answerable question to solve
  • Recommendation: Proposed actions that are MECE, specific, and time-bound

Core Principles: MECE at recommendation level, clarity, actionability. The Complication must be distinct from the Problem — Complication is "what changed," Problem is "what question to solve."

2. Issue Tree Builder (MECE Decomposition)

Purpose: Break down complex problems into MECE components for structured analysis.

Structure: Top-down, 2 levels max (Governing Question → 3 Branches → 2-3 Sub-Branches each).

  • Level 0: Governing Question (specific, answerable)
  • Level 1: Major branches — decompose to fundamentals (not actions)
  • Level 2: Testable hypotheses

Common Decompositions:

  • Revenue = Price × Quantity × Pipeline
  • Profit = Revenue - Cost
  • Growth = New customers + Expansion + New products
  • Market entry = Attractiveness + Competitive position + Capability

Prioritization: After building the tree, rank branches by Impact × Effort. Start with High Impact + Low Effort.

3. Hypothesis Tree (Day-1 Hypothesis)

Purpose: Commit to a best-guess answer before analysis, so the work targets what would change your mind.

Structure: Top = Day-1 answer → 3 sub-claims (if all true, top is proven) → Tests that could kill each sub-claim.

Key Rules:

  • You MUST commit to a Day-1 answer (even a weak guess beats no guess)
  • 3 sub-hypotheses max
  • Every leaf is a test that could kill the branch
  • State conviction (High/Med/Low) — test low-conviction first
  • Disconfirmation > confirmation: design tests to kill the hypothesis

Issue Tree vs. Hypothesis Tree:

Issue TreeHypothesis Tree
TopThe questionThe answer (your guess)
BranchesSub-questions (MECE)Sub-claims proving the top
BottomAreas to investigateTests that would kill the claim
Use whenYou don't know what mattersYou have a strong prior

4. Synthesis (Insight Extraction)

Purpose: Turn raw inputs (interviews, surveys, research) into 3 actionable insights with evidence and implication.

Process:

  1. Cluster — let themes emerge from data (induction, not deduction)
  2. Force-rank to 3 by frequency, severity, surprise
  3. Name each insight as a claim (falsifiable, specific)
  4. Stack evidence (2-3 bullets with direct quotes, exact numbers, named sources)
  5. Write the so-what (one-line implication: "Because this is true, the reader should ___")
  6. Anti-summary check: would the reader know what to do next?

The test: "Did the output change how the reader will act?" If no, you summarized — rewrite.

5. Prioritization Frameworks

Purpose: Score and rank features, initiatives, and tasks using proven frameworks.

FrameworkWhen to UseFormula
RICEQuantitative data available(Reach × Impact × Confidence) / Effort
Impact/Effort MatrixQuick decisions, limited data2×2 grid
Value vs. ComplexityEngineering-heavy decisions2×2 grid
Weighted ScoringCustom business criteriaΣ (Score × Weight)

When data is missing: Provide baseline estimates based on industry patterns, mark clearly as estimates, use ranges (not false precision), and flag high-uncertainty items for validation.

6. AI Use-Case Scorer

Purpose: Score personal AI use cases on Value × Feasibility × Safety to tier them: Do Now / Do This Quarter / Park / Avoid.

Scoring (1–5 each axis, final = V × F × S):

  • Value: Time saved + Quality lift + Strategic fit
  • Feasibility: Tool maturity + Your skill/time
  • Safety: Quality risk + Data risk + Trust risk (inverted: higher = safer)

Tiers: ≥60 Do Now, 30–59 This Quarter, 10–29 Park, <10 Avoid.

7. Data Insights

Purpose: Analyze any dataset and answer key questions in plain English. Best for quick analysis without being a data scientist. Ask for CSV/JSON data, specify the questions, and get structured analysis back.

8. Decision Memo Builder

Purpose: 1-page memo forcing a yes/no decision with specific owner and date.

Memo DNA — four non-negotiables:

  1. Brutal brevity (~400 words max)
  2. Storyline clarity (Situation → Complication → Resolution)
  3. Decision-forcing (last line = yes/no ask with owner and date)
  4. Evidence density (every claim backed by number, quote, or source)

Structure: Title (≤10 words) → Context (3-4 sentences) → Complication (2-3 sentences) → Options (2-3, 1-2 lines each) → Recommendation (1 paragraph) → Risks & Mitigations (≤3 bullets) → The Ask (1 line).

9. Top-Down Memo (Minto Pyramid)

Purpose: Answer-first writing — lead with the conclusion, then MECE arguments, then evidence.

Structure: LEAD (one sentence answer) → ARG 1 (claim + evidence) → ARG 2 → ARG 3 → (optional) Risks/Open Questions.

The test: Can the reader stop after the first paragraph and know what you think? If they need to read to the end, you wrote a story, not a memo.

10. Storyline Builder

Purpose: Build presentation storylines where each line = one slide title, creating logical narrative flow.

Key characteristics:

  • Each line = action-oriented message (not topic label)
  • Logical flow: Problem → Context → Analysis → Solution → Roadmap
  • Reader must understand the full story from titles alone

Templates by situation:

  • Market Strategy/Pitch Deck: Market opportunity → Competitive position → Product strategy → Roadmap
  • Internal Problem-Solving: Problem framing → Root cause → Solution options → Prioritization → Next steps
  • Project Roadmap: Approach → Phases → Activities → Timeline → Success criteria

11. Meeting Prep Kit

Purpose: Personal prep tool for meetings where YOU are driving an outcome.

Output sections:

  • Pre-read (3 bullets, ≤30 words each) — minimum context needed
  • Time-boxed agenda (adds up to meeting length, decision time reserved at end)
  • Talking points (top 3, in spoken language) — lead with strongest, address elephant, make ask
  • Anticipated objections (top 3 with rebuttals per person)

Design rules: Outcome-first, top 3 not all 12, speakable not slidable.

12. Stakeholder Map

Purpose: Map people onto Power/Interest grid with stance (Champion / Supporter / Neutral / Skeptic / Blocker / Unknown) and one influence move per person.

Quadrants:

  • Manage Closely (High Power, High Interest) — deciders, most time goes here
  • Keep Satisfied (High Power, Low Interest) — pre-brief so they're never surprised
  • Keep Informed (Low Power, High Interest) — give them a role
  • Monitor (Low Power, Low Interest) — quarterly update enough

Key Rules: Real names not roles. Honest stances. Coalition sequence (who to align first).

13. Workshop Designer

Purpose: Design half-day or full-day workshops that end with committed artifacts and named owners.

Design rules:

  • One outcome (not three)
  • Diverge then converge (never both at once)
  • Breakouts < 4 people
  • Energy curve: heavy thinking morning, decisions afternoon (never right after lunch)
  • Pre-build the artifact template
  • Facilitator plays referee, not participant

Output: Workshop charter → Hour-by-hour agenda → Breakout playbook → Facilitator "what if" notes → Artifact template → Post-workshop send (24-hour SLA).

14. McKinsey Charts (Native PowerPoint)

Purpose: Generate McKinsey-style charts as native python-pptx objects (editable in PowerPoint).

Three chart types:

  1. bar_callout: Single-series column with one highlighted bar + callout textbox. Use for TAM/single-number stories.
  2. stacked_bar_over_time: Stacked column over time with one series highlighted. Use for composition/mix stories.
  3. waterfall: Start → adds → subtracts → total bridge. Use for revenue/cost bridge analysis.

Design choices (intentional, not configurable):

  • Title is a claim, not a label
  • One highlight color (everything else grey)
  • No gridlines, no chart border
  • Source line bottom-left in grey italic
  • Single font (Inter), consistent sizes

Dependencies: Python 3.9+, python-pptx. See references/charts.py for the full implementation.

15. McKinsey Critic (EM Review)

Purpose: Review decks, docs, and strategies like a McKinsey engagement manager at 2am before a client presentation.

Grading: Green (ready) / Yellow (fixable overnight) / Red (needs rework).

What it checks:

  • Narrative flow (Problem → Context → Analysis → Solution)
  • Data rigor (specific numbers, sources, fair comparisons)
  • So-what on every slide
  • Claim-based titles vs topic titles
  • Frameworks vs bullet lists
  • Financial frames where needed

Output format: Grade → Section-by-section table → Top 3 fixes (ranked by impact) → One thing that works.


Common Pitfalls

General Consulting Mistakes

  • Skipping the hypothesis tree: The #1 reason strategy work takes 3x longer than it should. Commit to an answer before you start.
  • Boiling the ocean: Don't investigate every branch of an issue tree. Prioritize by Impact × Effort and start with the highest-impact, lowest-effort branches.
  • Chronological narratives: The way you thought about something is the opposite of how the reader needs to hear it. Lead with the answer.
  • Synthesis by list: "Here are 7 themes" is clustering, not synthesis. Cut to 3 with implication.
  • Missing the so-what: Every insight/argument needs an explicit implication. If the reader doesn't know what to do next, you haven't finished.

SCPR Mistakes

  • Situation too detailed: Keep to essential context only.
  • Complication = Problem: They're different. Complication is "what changed," Problem is "what question to solve."
  • Overlapping recommendations: Ensure MECE structure.
  • No timelines: Always include "by when" in recommendations.

Issue Tree Mistakes

  • Actions instead of fundamentals: Level 1 decomposes the problem (Price × Quantity), not solutions.
  • Not MECE: Overlapping branches or missing key dimensions.
  • Too deep: 2 levels from governing question max.
  • Not testable: Terminal branches must be provable/disprovable with data.

Decision Memo Mistakes

  • Bullet-point soup with no storyline: Must flow as Situation → Complication → Resolution.
  • Decision-averse hedging: "We should consider..." — recommend or don't write the memo.
  • Missing numbers: "Significant impact" without specificity.
  • Recommendation leaked into options: Options are neutral; recommendation is separate.
  • Ask = "discuss next steps": The ask must be a specific yes/no with owner and date.

Top-Down Memo Mistakes

  • Burying the lead: Don't start with "Over the last 6 weeks, the team explored..." Lead with the answer.
  • Topic-shaped arguments: "Pricing" is a topic. "Pricing is 2.4x the standalone value" is an argument.
  • More than 4 arguments: You haven't synthesized.
  • Evidence that doesn't defend its argument: Each bullet should clearly serve the headline.

Stakeholder Map Mistakes

  • Mapping by title, not by power: The CTO might not care; the staff eng who's been there 8 years might be the blocker.
  • Skipping the "unknown" stance: Unknown is honest. "I think they're supportive" without evidence is wishful.
  • Over-engaging Monitor quadrant: Don't burn cycles converting people who don't care.
  • No coalition sequence: Knowing who matters isn't enough. You need the order of alignment.

Workshop Design Mistakes

  • Multi-outcome workshops: "Alignment + brainstorm + decisions" = none of the three. Pick one.
  • No artifact template: "We'll write it up" = it doesn't get written up. Pre-build.
  • Mixing divergent and convergent: "Let's brainstorm options and pick the best one in 45 min" produces neither good options nor a defensible pick.
  • Decision moment right after lunch: Energy nadir. Schedule application work there, not commitment.
  • The sponsor facilitates: They have a stake. Pick a neutral facilitator.

McKinsey Charts Mistakes

  • No python-pptx installed: The chart code requires pip install python-pptx. Run this before calling chart functions.
  • Using images instead of native charts: The whole point is editable charts. Never paste screenshots of charts — use the python-pptx functions.
  • More than 5 series stacked: Split into small multiples instead.
  • Rainbow charts: One highlight color, everything else grey. No rainbow.

Deck Pipeline Mistakes

  • Running the pipeline without a clear audience: The Strategist needs to know who the deck is for.
  • Skipping stages: Each agent in the pipeline serves a distinct purpose. Skipping the Critic means shipping an un-reviewed deck.
  • Missing the critic's top 3 fixes: The Critic identifies the highest-impact problems. Fix those and only those before polish.

Verification

For Any Consulting Output

  1. Memo/Top-Down Memo: Can the reader stop after the first paragraph and know the answer? Is every claim backed by a number, quote, or source?
  2. Issue Tree: Is the governing question specific and answerable? Are branches MECE? Are terminal branches testable with data?
  3. Hypothesis Tree: Is there a clear Day-1 answer? Can each test kill its branch? Is there an overall kill-criterion?
  4. Storyline: Can you understand the full argument from slide titles alone? Are titles claims, not topics?
  5. Synthesis: Did the output change how the reader will act? Are insights claims (not topics) with evidence and so-what?
  6. Decision Memo: ~400 words max? Last line is a yes/no ask with owner and date?
  7. Stakeholder Map: Are stances honest? Is there a coalition sequence? Is each stakeholder mapped by power (not title)?
  8. Meeting Prep: Do talking points ladder to the outcome? Are objections specific per person?
  9. Workshop: Is there one outcome? Pre-built artifact template? Facilitator "what if" playbook?
  10. McKinsey Charts: Are charts native pptx objects (right-click → Edit Data)? Is there a source line? One highlight color?
  11. Deck Pipeline: Check the Critic's grade. If Yellow/Red, apply all top 3 fixes before shipping.
  12. McKinsey Critic: Run the output through the critic before it reaches stakeholders. Fix structure before polish.

Python Charts Verification

pip install python-pptx
python references/charts.py  # Generates test_output.pptx with all 3 chart types

Open test_output.pptx in PowerPoint and right-click any chart → Edit Data to confirm it's a native chart object.


Composition Patterns

Most real work uses 2–3 of these together:

NeedStack
Customer research → memoSynthesis → Top-Down Memo → McKinsey Critic
Strategic question → recommendationIssue Tree → Hypothesis Tree → run tests → Decision Memo
Big decision with politicsStakeholder Map → Decision Memo → Meeting Prep Kit
Board deck (production)Issue Tree → Storyline Builder → Charts → Deck Pipeline → Critic
Offsite / alignment workshopWorkshop Designer → Meeting Prep Kit → Top-Down Memo
Pricing or feature analysisIssue Tree → Prioritization → Decision Memo
Weekly exec updateSynthesis (of week's data) → Top-Down Memo (the ask)
Product launch proposalSCPR Framework → Storyline Builder → Charts → Deck Pipeline
Quarterly planningWorkshop Designer → Prioritization → Top-Down Memo (post-send)
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 322kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 322kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 322kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 197kUpdated Claude CodeCodexVerifiedInstall