⌘ The Architecture Reference

Path strategy · The Architect's Path · Intermediate

Analyzing Strategy

The analysis metapatterns — MECE, Logic Trees, and Hypothesis — plus the world-context frameworks PESTEL, Scenario Planning, Futures Funnel, and Backcasting that ground a strategy in reality.

Path strategy Intermediate ⏱ 5 min read Complete

🗂️ Analogy

A good filing system makes every document obvious to find and impossible to misplace. Borges mocked the opposite: a “Celestial Emporium” taxonomy of animals that includes “those belonging to the Emperor,” “those drawn with a fine camel-hair brush,” and “those that from a distance resemble flies.” Analysis is the discipline of building lists that aren’t absurd.

Before you create or communicate a strategy, you analyze. Hewitt’s analysis layer is a set of metapatterns for critical thinking that apply at every scope, plus the world-context frameworks that keep a strategy grounded.

The three metapatterns

graph TD
P["Strategic problem"] --> M["MECE list<br/>(no overlap, no gaps)"]
M --> LT["Logic Tree"]
LT --> DG["Diagnostic tree<br/>(why → root cause)"]
LT --> SL["Solution tree<br/>(how → courses of action)"]
M --> H["Hypothesis<br/>(P → Q, a stake in the ground)"]
H --> R["Focused analysis"]
  • MECE (“mee-see,” from McKinsey) — Mutually Exclusive, Collectively Exhaustive. The most useful tool in the box, because strategy is “a list of lists.” Profit = Revenue − Cost is MECE; North/South/West is not. Follow the Rule of Three: pick an abstraction level that keeps lists at three or five items, because odd numbers are memorable (The Good, the Bad, and the Ugly). Executives have “Executive ADD” — give the MECE summary, not the long list. Present database options as a ranked Good / Better / Best (itself MECE), and show your “long math” so the conclusion is trusted.
  • Logic Tree (Issue Tree) — a MECE decomposition of a problem. A diagnostic tree works backward to the root cause (why); a solution tree maps courses of action (how). Built like the Five Whys / fishbone diagram. The Jefferson Memorial example: eroding stone → harsh chemicals → pigeons → spiders → midges → lights attracting midges at dusk → solution: turn the lights on one hour later — free and instant.
  • Hypothesis — a claim about why something might be true (form P → Q) that puts “a stake in the ground” to focus analysis. McKinsey consultants form one within hours. A hypothesis is a list of sub-hypotheses, each based on an insight, each based on data points.

⚠️ Problems vs. opportunities

Focusing only on problems can, at best, maintain the status quo. Pursue opportunities — gains customers don’t yet know they want. The iPhone (2007) and the five-colored iMac (1998) addressed gains, not pains.

Reasoning under uncertainty

Hewitt’s Hypothesis pattern adds discipline drawn from logic and probability:

  • Distinguish raw facts from interpretive insights, and beware tautologies (“all bachelors are unmarried” is trivially true).
  • Use inductive reasoning carefully (beware Russell’s turkey, fed daily until Thanksgiving) and deductive reasoning from principles.
  • Estimate probability with ranges or High/Medium/Low traffic lights, never false precision — the Fed predicted 0 of 144 crises from 2005–2014. Hold several hypotheses at once, think probabilistically, and update often.
  • Ease and Impact Scoring plots work on a scatterplot of ease (x) vs. impact (y): top-right green (do first), top-left yellow (second), bottom-right (third), bottom-left red (last).
graph TD
TR["High impact, easy<br/>do first (green)"]
TL["High impact, hard<br/>do second (yellow)"]
BR["Low impact, easy<br/>do third"]
BL["Low impact, hard<br/>do last (red)"]
TR --> TL --> BR --> BL

🔑 Signal and noise

Apply the 80/20 (Pareto) rule: separate signal (what matters) from noise (random patterns mistaken for signal). A few fundamentals get you 80% of the way — make a “good enough” conclusion quickly, then refine.

Grounding in the world: four nested patterns

The broadest creation scope feeds four patterns in sequence — PESTEL → Scenario Planning → Futures Funnel → Backcasting — so the strategy is grounded, not a shopping list:

  • PESTEL (Aguilar, 1967) — the broad climates: Political, Economic, Social, Technological, Environmental, Legal. It is MECE and precedes resource allocation — “it tells you the weather so you pack an umbrella.” Caveat: “as an idea, it’s really not much more than the acronym.”
  • Scenario Planning — an organized way of asking “What if—?”, born at RAND in the 1950s. It counters the default “Do Nothing” strategy and its optimism bias.
  • Futures Funnel — a single Venn-like slide nesting four sets: Possible (widest), Plausible, Probable (focus here), and Preferred (the value judgment — what you want).
  • Backcasting — the inverse of forecasting: state the “Beautiful Future” as if it already happened (“no customer finds a bug before we do”), then work backward through each necessary prior state (its antecedents) to the status quo, tagging each with a probability.

💡 Sound analysis is iterative and fractal

Analysis is self-similar at any scale — applied informally in your head for small problems, formally with evidence for big ones. Always analyze the problem and the solution separately.

See also

When to use it — and when not

✅ Reach for it when

  • When you need to structure a messy problem into clean, decision-ready lists and trees.
  • When forming a fast hypothesis to focus analysis before you have all the data.
  • When scanning the broad climate (PESTEL) so the strategy isn't a shopping list of shiny objects.

⛔ Think twice when

  • When a list mixes levels of abstraction or leaves gaps — a non-MECE list misleads.
  • When you demand false precision (an exact probability) instead of ranges or High/Medium/Low.
  • When chasing only problems; pursue opportunities (gains) too, not just pains.
path-strategyWhat Is Technology Strategy?

Strategy is the art of creating power — technologists win funding by speaking the language of business, using proven patterns to analyze, create, and communicate a fundable course of action.

path-strategyCreating Strategy

The creation patterns build a strategy's substance in concentric scopes — industry frameworks (SWOT, Porter's Five Forces, Ansoff) and corporate portfolio tools (BCG Growth-Share, APM, Value Chain).

path-strategyCommunicating Strategy

Winning approval is its own discipline — the 30-second answer, Ars Rhetorica, dramatic structure, the meeting before the meeting, and the Ask Deck that turns analysis into funded action.

Check your understanding

Score: 0 / 4

1. What does MECE stand for, and why is it the most useful tool in the box?

Mutually exclusive = no overlap; collectively exhaustive = nothing left out. Profit = Revenue − Cost is MECE; North/South/West is not. A good list also makes clear who the audience is and why they care.

2. What are the two kinds of Logic (Issue) Tree?

The diagnostic tree works backward to the root cause (like the Five Whys / fishbone diagram); the solution tree imagines the ideal end state and regresses toward the present.

3. What does PESTEL stand for?

Aguilar's PESTEL frames the broad climates affecting a business; it 'tells you the weather so you pack an umbrella' and is itself MECE, but has no method inside it beyond the acronym.

4. How does Backcasting differ from forecasting?

Backcasting creates a concrete 'Beautiful Future' vision, then hypothesizes each necessary prior state (its antecedents), recursing back to the status quo and tagging each with a probability.

Comments

Sign in with GitHub to join the discussion.