MECE breakdown of a problem space
intermediateClaude OpusPersonal ProductivityDecision Makingframeworkmethodologydecompositiondecision-makingstructured-thinking
Use case
Use this when a question is sprawling and the team is jumping to solutions without a shared map. The MECE structure forces Claude to prove the buckets do not overlap and do not leak — which is where the typical 'breakdown' prompt fails.
The prompt
You are decomposing a problem the way a senior strategy partner would: into buckets that are Mutually Exclusive (no overlap) and Collectively Exhaustive (nothing missing). The discipline is in the proof, not the labels. <context> Problem statement:{{problem}}Decision the breakdown will inform:{{decision}}Audience for the analysis:{{audience}}Known constraints or scope boundaries:{{constraints}}</context> <task> Step 1 — Frame. Restate the problem as a single question that a 3 to 6 bucket breakdown could answer. If the problem as written is two questions, split it and pick the one most relevant to{{decision}}. Step 2 — Propose two candidate breakdowns. Generate two distinct top-level decompositions (3 to 6 buckets each). Use different cutting axes (for example, customer segment vs. revenue mechanism vs. lifecycle stage). Label the axis used for each. Step 3 — MECE check each candidate. For each breakdown, run two tests: - Mutual exclusivity: name a concrete case that could plausibly fall in two buckets. If you find one, the breakdown fails. - Collective exhaustiveness: name a concrete case that falls in none of the buckets. If you find one, the breakdown fails. Show the test cases and the verdict explicitly. Step 4 — Recommend. Pick the breakdown that better serves{{decision}}. Justify in two sentences. Then expand the chosen breakdown one level deeper (3 to 5 sub-buckets per top-level bucket). Step 5 — Where to dig first. Identify the single bucket where investigating further is most likely to change{{decision}}. Explain why. </task> <output_format> ## Reframed question ## Candidate A and B (with axis) ## MECE tests (table) ## Chosen breakdown (two levels) ## Highest-leverage bucket </output_format> <constraints> - If you cannot find a breakdown that passes both MECE tests, say so plainly and propose a "near-MECE" version with explicit caveats. Do not paper over overlaps. - Do not invent data. The output is a structure, not a conclusion. - Bucket labels should be nouns or noun phrases that name a category, not verbs that imply an action. </constraints>
Variables
{{{{problem}}}}Replace with your {{problem}}{{{{decision}}}}Replace with your {{decision}}{{{{audience}}}}Replace with your {{audience}}{{{{constraints}}}}Replace with your {{constraints}}
Example
Input
problem: Why is our net revenue retention slipping from 118% to 104% over the last three quarters? decision: Where to focus the retention task force in Q4 audience: CEO and head of CS constraints: B2B SaaS, ~600 customers, ACV $25k median
Output
## Reframed question Which sources of NRR change account for the 14-point drop, and which are most addressable in one quarter? ## Candidates - A (mechanism axis): logo churn, downsell, expansion shortfall, price compression, mix shift - B (segment axis): SMB, mid-market, enterprise, partner-sourced, self-serve ## MECE tests | Candidate | Overlap test | Exhaust. test | Verdict | |---|---|---|---| | A | A customer that downgrades and then churns hits two buckets — fix by attributing to first event in window | An NRR change from currency FX fits none — add "non-operational" bucket | Pass with caveats | | B | Self-serve SMB sits in two buckets | All customers covered | Fails MX | ## Chosen breakdown — A - Logo churn → reason: product fit, support, price, sponsor change, M&A - Downsell → seat reduction, plan downgrade, scope cut at renewal - Expansion shortfall → fewer upsells, smaller upsells, longer cycles - Price compression → discount creep, multi-year locks, free seat policy - Non-operational → FX, accounting policy ## Highest-leverage bucket Downsell. It typically lags churn by one quarter and is the earliest reversible signal — if Q3 downsell is up sharply, Q4 churn is forecastable and CS can intervene now.
Tips for best results
- 1The MECE test step is what makes this stronger than the typical decomposition prompt — without forcing concrete overlap and gap test cases, breakdowns look clean on paper but leak in practice.
- 2Two candidate breakdowns is non-negotiable. The first one Claude proposes is almost always biased toward the user's framing; the second exposes the cutting-axis assumption.
- 3If both candidates fail the exhaustiveness test, the problem is not yet a question. Go back to step 1.
- 4Keep top-level buckets to 3 to 5. Above 6, the audience cannot hold them in working memory and the breakdown stops being useful.
- 5Pair with first-principles-decomposition when the buckets all feel inherited — MECE structures the surface, first principles tests whether the surface is real.
Related prompts
First-principles decomposition of a problem
advancedStrip a problem down to its load-bearing assumptions, then rebuild from atoms. Forces Claude to separate what is true from what is convention.
Personal Productivityframeworkmethodologyfirst-principles
Claude tree of thoughts — branch, evaluate, prune
advancedExplore a decision by branching into multiple reasoning paths, scoring each branch on the same criteria, then committing to the strongest one with explicit trade-offs.
Personal Productivityframeworkmethodologyreasoning
Fishbone (Ishikawa) analysis for a multi-cause problem
intermediateMap a multi-cause problem across the standard fishbone categories, weight each cause by likelihood and impact, then commit to the highest-leverage interventions.
Operationsframeworkmethodologyroot-cause
Need help implementing this prompt in your workflow?
Book a call