The GE McKinsey Matrix, Automated: A 9-Box Guide for the AI Era

What the GE McKinsey Matrix actually is

Developed by McKinsey for General Electric in the 1970s, the matrix plots business units (or product lines, or markets) on a 3×3 grid across two composite axes:

• Industry Attractiveness — market size, growth rate, profitability, competitive intensity, regulatory posture.
• Business Unit Strength — market share, brand equity, cost position, distribution reach, capability moat.

Each axis is a weighted score of 5–10 factors. Where a unit lands (Invest, Selectively Grow, Harvest) dictates capital allocation.

Why the classic scoring breaks

Three failure modes show up every time a team runs the matrix manually:

• Weight rigging. Whoever owns the unit tunes weights until their unit lands in Invest.
• Stale evidence. Attractiveness scores anchor to last year's market data; growth curves move faster than the scoring meeting.
• Consensus collapse. Cross-functional workshops converge on the middle box — nothing is a clear Harvest, everything is Selectively Grow.

Automating the 9-box with AI

The scoring problem is a structured-reasoning problem. AI models can be forced to output per-factor scores with evidence citations and confidence intervals — the exact discipline human workshops skip.

1. Define the portfolio and factor weights up front

Freeze the units, the attractiveness factors, and the strength factors before any scoring starts. Weights get signed off by the executive sponsor, not the unit owners.

2. Run per-factor scoring with citation-required prompts

For each unit × factor cell, ask the model to produce a 1–10 score with an evidence line and a confidence tag (High / Medium / Low). Low-confidence cells get flagged for human review.

3. Aggregate and place on the grid

Weighted sums per axis land each unit in one of the nine boxes. The output isn't a slide — it's a table with cell-level evidence you can audit.

4. Stress-test the placements

Run adversarial prompts: "What would move Unit X from Invest to Harvest?" If the answer is a small perturbation in one factor, the placement is fragile — flag it. OMEGA's war-game loop handles this step natively.

5. Re-run quarterly against fresh data

The matrix is a snapshot. Automating it means you can re-run it every quarter without burning six weeks of workshops. The Decision Memory keeps the prior scoring, so drift is visible.

GE McKinsey vs BCG Growth-Share

BCG's matrix uses two single dimensions (market growth, relative market share). GE McKinsey uses two composite dimensions with weighted sub-factors. BCG is faster to eyeball; GE McKinsey holds up under scrutiny. For real capital allocation decisions, use GE McKinsey — and let the model do the arithmetic.

Where OMEGA fits

OMEGA's Strategic Analyzer runs the per-factor scoring pass with evidence citations. The War-Game module stress-tests placements. The Decision Graph links each unit's scoring across quarters so you can see which bets are drifting — see the AI-enhanced strategic planning guide for the broader loop.