Strategy.

The framework

1. 01

   Test for mutual exclusivity.

   Ask: can a single element belong to more than one category? In a customer segmentation: 'large enterprise' and 'high-spending accounts' can overlap — large enterprises often are high-spending. That's a MECE failure. Fix it by choosing a single axis of segmentation: company size OR spending level, not both. Every overlap means you're counting something twice — and every analysis built on top of that will be wrong.

2. 02

   Test for collective exhaustion.

   Ask: is there any element not captured by this list? In a revenue breakdown: direct sales + channel sales + enterprise. What about self-serve? International? Upsell and expansion? A collectively exhaustive list accounts for the whole universe. The test: ask who or what has been left out — the edge cases, the exceptions, the segments you're not targeting. Missing them doesn't make them irrelevant. It means your analysis doesn't account for them.

3. 03

   Build issue trees, not lists.

   MECE's most powerful application is the issue tree: a hierarchical decomposition of a problem into its component parts. Revenue declining → revenue = volume × price → volume = new customers + retained customers → new customers = acquisition rate × qualified leads. Each branch is MECE at its level. The tree makes it impossible to miss a driver — and it tells you exactly which branch to investigate for the root cause.

4. 04

   Apply it to your communication.

   MECE structures more than analysis. A presentation is MECE if each slide covers a distinct part of the argument without repeating the previous one, and the deck as a whole covers the full case. An email is MECE if the three points don't overlap and together address the entire issue. Read the headers only. Do they overlap? Are they complete? If yes and yes — the body will follow.