Strategy & Operations Consulting

Diagnosing the Real Problem

1. Build MECE issue tree in Excel or Lucidchart decomposing the stated issue into core drivers (e.g., Sales = Price × Volume; Volume = #Transactions × Avg Units/Transaction; branches: acquisition, retention, basket size).

   • Input: Restated problem statement from user query.
   • Test MECE using 'one-home rule' (assign each sub-issue to exactly one branch) and Excel checklist (columns: Issue, Parent Branch, Coverage %, Overlap Flag).
   • Output: Visual tree diagram with 100% coverage and zero overlap flags, because incomplete decomposition misses root drivers while overlaps cause double-counting.
   • Success criterion: Checklist shows full mapping of restated problem; iterate until zero flags.

2. Generate 3-5 testable hypotheses per major branch (e.g., "Repeat purchase rate dropped 20-30% due to loyalty program glitches").

   • Input: Tree branches from Step 1.
   • Format: Bullet list with hypothesis statement, predicted gap size (%), and test method.
   • Limit to quantifiable predictions, because vague hypotheses resist validation.

3. Prioritize hypotheses using 2x2 matrix (rows: branches; columns: Impact [% of total gap], Feasibility [1-10 score], Priority quadrant).

   • Input: Hypotheses from Step 2; estimate impact as % of gap explained, feasibility on data access (<2 weeks via standard sources).
   • Output: Ranked list selecting top 1-2 per quadrant (see Decision Rules for thresholds), because this focuses effort on high-leverage tests.

4. Generate and populate Hypothesis Validation Scorecard for top hypothesis using provided data.

   • Input: User-supplied internal metrics (e.g., CRM extracts), customer feedback summaries (e.g., 10-15 responses), external benchmarks (e.g., industry reports).
   • Use Google Sheets template below; score each source strength and compute convergence.
   • Output: Completed scorecard (template below), because triangulation from 3+ sources confirms causality over correlation.

Hypothesis Validation Scorecard

Hypothesis: [State full hypothesis, e.g., "Loyalty glitches caused 25% repeat drop"]

| Component | Quantitative (Strength: High>1k sample/Med>100/Low) | Interviews (Strength: High≥70% consensus/Med≥50%/Low) | Benchmarks (Strength: High direct/Med proxy/Low) | Convergence (≥2/3 High/Med?) | Residual Gap (%) |
|-----------|-----------------------------------------------------|-------------------------------------------------------|-----------------------------------------------|------------------------------|------------------|
| [e.g., Delay frequency] | [Evidence + Strength] | [Evidence + Strength] | [Evidence + Strength] | [Yes/No] | [Unexplained %] |
| ... (4-6 rows) | | | | **Overall: [Confidence %]** | **Total: [%]** |

• Example: For retail sales drop, scorecard populated with CRM cohort curves (High, n=5k), interview quotes (High, 80% consensus on glitches), benchmarks (Med proxy) yields 85% confidence.

Scoping an Engagement

1. Structure one-page scoping memo using Objectives-First Framework.

   • Input: Client objective and constraints from query.
   • Sections: North Star Objective, 3-5 Key Questions, Deliverables (with owners), Timeline (e.g., 8 weeks), Success Metrics (e.g., 20% uplift target).
   • Output: Markdown or PDF memo, because it aligns stakeholders pre-kickoff and excludes scope creep.

2. Simulate co-creation by generating Miro workshop agenda (90-min outline: 20-min objective alignment, 30-min question brainstorming, 40-min memo drafting).

   • Include prompts for participant input, because virtual agendas enable remote execution.

Prioritizing Initiatives

1. Plot initiatives on 2x2 Impact-Effort Matrix in Excel (x-axis: Effort [resource hours]; y-axis: Impact [NPV estimate]).

   • Input: List of 10-50 initiatives with rough NPV and hours from user.
   • Quadrants: Quick Wins (high impact/low effort), etc.
   • Output: Grid with top 3-5 prioritized (see Decision Rules), because 80/20 rule surfaces 20% effort for 80% value.

2. Run sensitivity analysis: Vary NPV inputs ±20% and re-plot.

   • Output: Updated matrix with stability bands, because it flags fragile priorities.

Use these thresholds to advance or pivot during workflows.

Example: Acquisition drop hypothesis with 60% impact and 9/10 feasibility (CRM Day 1 data) advances to test.

• Never propose solutions until top hypothesis validated via scorecard with ≥70% confidence using 3+ sources, because untested fixes waste 80% of effort on symptoms.
• Never advance from MECE tree build without zero checklist flags and 100% coverage, because gaps or overlaps distort prioritization.
• Never scope an engagement without upfront measurable success metrics (e.g., % uplift targets), because undefined outcomes enable scope creep and value erosion.

• Don't allow overlapping branches in MECE trees (e.g., inventory inefficiency double-counted with demand forecasting). Instead, apply one-home rule and checklist to merge/reassign, because overlaps dilute evidence and recommendations.
  • Example: E-commerce tree flagged 25% pricing/supply overlap; reassignment isolated 15% margin leak.

• Excel/Lucidchart: Visualize MECE trees and 2x2 matrices.
• Google Sheets: Hypothesis Validation Scorecard (template above); prioritization sensitivity (vary NPV ±20%).
• Miro: Agendas for scoping workshops.
• NVivo-style coding template: For qualitative themes (columns: Quote, Theme, Frequency %).

Primary deliverables: MECE tree diagram, prioritized hypothesis list, populated scorecard, one-page scoping memo, Impact-Effort grid.

• Interdependent drivers (>20% causal links, e.g., forecasting affects inventory/pricing): Elevate root to standalone branch; annotate downstream influence % (e.g., 60%); expand checklist with Dependency Flag column; iterate to zero flags.

  • Signal: Causal chain detected in tree build; adapt per BCG method for 100% coverage.

• Predominantly qualitative issues (>70% soft drivers, e.g., cultural silos with no KPIs): Pivot to Cultural Immersion—generate 20-question interview guide + shadowing checklist; produce theme-coding spreadsheet (NVivo template: Quote, Theme %, Hypothesis Link); build qual hypothesis tree; output workshop agenda for MECE test.

  • Signal: <30% quantifiable gap post-initial tree; quantify via pilots after behavioral themes emerge.
  • Example: Tech firm siloed culture (40% blame theme) drove attrition diagnosis.

For detailed examples, walkthroughs, and edge cases, consult 'references/REFERENCE.md'.