STRUCTURED ANALYSIS · PRACTICAL GUIDE

MECE Principle: Break Problems Down Without Gaps

Build issue trees whose branches have clear boundaries and collectively cover the decision you need to make.

Quick answer

Split a problem into parts that do not overlap and together cover the defined scope.

MECE stands for Mutually Exclusive, Collectively Exhaustive. It reduces duplicate analysis and hidden gaps; it is a practical quality target, not a claim that every real-world category has one perfect answer.

The two MECE tests

Mutually exclusive

Each item belongs to one branch in principle. Clear boundaries prevent double counting and duplicate ownership.

Collectively exhaustive

All branches together cover the defined question, including important cases that may not yet have data.

Define the boundary first. “Every sales problem” cannot be usefully exhausted. “Direct causes of June year-over-year sales decline in China direct e-commerce” can.

Case study: diagnosing a sales decline

June sales fell 18% year over year. A brainstorm produces traffic, churn, competitors, conversion, price, and advertising. Those ideas overlap and mix levels.

Sales = purchasing customers × average order value
A. Fewer purchasing customers
  • Fewer visitors
  • Lower visitor-to-purchase conversion
  • New and returning customers at the next level
B. Lower average order value
  • Fewer items per order
  • Lower average item price
  • Discount or product-mix changes

The formula creates stable first-level boundaries. Data can show whether customer count, order value, or both explain the result.

Overlapping split
  • New customers
  • Online channel
  • Advertising
  • Brand strength

Advertising is part of online acquisition, while new customers may come from it.

Cleaner split
  • Organic traffic
  • Paid traffic
  • Direct traffic
  • Partner traffic

Each visit receives one primary source before deeper analysis.

Common breakdown dimensions

DimensionBest forExampleRisk
FormulaCalculated outcomesProfit = revenue - costMay miss behavioral drivers
ProcessFinding bottlenecksAcquire → activate → buy → retainUnclear stage boundaries
ObjectComparing segmentsNew / returning customersOverlapping labels
TimeReviews and plansBefore / during / afterEvents may span periods
GeographyRegional performanceNorth / South / East / WestInconsistent definitions
Internal / externalInitial factor scanOperations / marketToo broad for action

Issue-tree template

CORE QUESTION
Within ______, why / how ______?

├─ A. First-level branch: ______
│  ├─ A1. Testable hypothesis: ______
│  └─ Data needed: ______
├─ B. First-level branch: ______
│  ├─ B1. Testable hypothesis: ______
│  └─ Data needed: ______
└─ C. First-level branch: ______
   ├─ C1. Testable hypothesis: ______
   └─ Data needed: ______

BOUNDARY: time ______  market ______  object ______
DECISION this analysis will support: ______

Overlap and omission checklist

  • Does the question define time, object, market, and metric?
  • Does one level use one breakdown rule?
  • Would each item belong to only one branch?
  • Is there an important counterexample outside every branch?
  • Are branches at a similar level of abstraction?
  • Is “other” small and clearly defined?
  • Can each leaf become a hypothesis, data request, or action?
Connection to the Pyramid Principle: MECE checks horizontal relationships within one level. The Pyramid Principle also requires lower-level ideas to support the statement above.