MECE
A test for whether a breakdown is any good, not a shape to pour one into.
What is MECE?
MECE is not a template. It is a test you apply to any breakdown you have already made: are the categories Mutually Exclusive — nothing fits in two — and Collectively Exhaustive — nothing is left out? Two questions, and almost every list of causes, options or segments fails at least one.
The two failures MECE catches are different and both expensive. Overlap causes double-counting: if "price" and "found a cheaper competitor" are separate churn categories, the same customer lands in both and price looks twice as important as it is. Gaps cause blind spots: a churn analysis with no category for "the champion left the company" will never surface the thing actually killing your renewals. Neither failure announces itself — a non-MECE breakdown looks perfectly reasonable, which is exactly why it needs a test rather than a glance.
Where MECE Came From
Barbara Minto, McKinsey, late 1960s
MECE was invented by Barbara Minto at McKinsey & Company in the late 1960s. Minto joined the firm in 1963 as its first female MBA hire, one of only eight women in a Harvard Business School class of six hundred, and went on to head the firm’s training. She is unambiguous about the parentage — asked how to pronounce it, her answer was "I invented it, so I get to say how to pronounce it". ("ME-see", if you were wondering.)
It is one half of a larger system
MECE is not freestanding. It underpins the Minto Pyramid Principle, published in 1985, which draws on Aristotle, Piaget and Mortimer Adler and is still the house style for how consulting decks are built. MECE governs how you group; the pyramid governs how you order. Using one without the other is common and leaves value on the table — the sibling framework is SCQA.
Why it survived contact with everything
Because it is a property rather than a procedure. A five-step method goes out of date when the work changes; "no overlaps, no gaps" is true of a churn analysis, a cost breakdown and a bug taxonomy equally. That generality is also its weakness, discussed below — a test that applies everywhere gives you no help choosing what to test.
The 2 Slots, One at a Time
Each slot is a decision. Leave it out and the model still makes it — just without you.
The easier half to check and the harder half to fix. Test it by taking three real examples and asking where each goes — if the answer is "either, really", the categories overlap. The usual cause is mixing organising principles: one category by cause, the next by timing, the next by who decided.
Test it the other way round: name three real cases the structure does NOT cover. If you can, it is not exhaustive. The temptation is to add a category called "Other", which is an admission rather than a fix — everything interesting ends up there.
One Task, Before and After
The task: Break down why customers churn. Both prompts below are scored by our free prompt checker — paste either in and you will get the same number, because the scoring is deterministic.
Break down the reasons customers churn.
Seven words. You will get eight plausible bullet points that overlap in three places and omit the reason that actually matters — and nothing in the output will tell you either of those things.
Break down the reasons our customers churn into a MECE structure - categories that do not overlap and that leave nothing out. Context: We are a B2B SaaS product sold to marketing teams at companies of 50-500 people, on annual contracts. We have exit-survey text and usage data, but no interviews. Work through the candidate groupings before committing to one, and state which organising principle you chose - by cause, by timing, or by who decided - and why the other two are worse for this data. Requirements: three to five top-level categories, each with two to four sub-categories, as a markdown table with columns Category, Sub-category, What evidence would confirm it. Keep the whole answer under 500 words. Every category must be testable against data we actually have. For example: ``` | Never activated | Never completed setup | Account created, zero campaigns published in first 30 days | ``` Then state explicitly which real churn reasons your structure would fail to capture. Do not invent percentages or churn rates, and do not produce a category called Other.
The two instructions doing the most work are the ones about the structure itself: name your organising principle, and name what you missed. Both force the model to expose a choice it would otherwise make silently.
Eighty-nine, and the missing points are the right ones to miss
One of the strongest worked examples on these pages, and the gaps are instructive rather than fixable:
MECE has no persona slot, and adding "you are a McKinsey consultant" produces consulting register rather than better categories. The structure is the output; the voice is irrelevant.
Capped because the real constraints here are structural — "three to five categories", "testable against data we have" — rather than the stylistic rules the check counts most readily.
A single fenced table row is worth ten points and took one line. It is the cheapest improvement available to almost any prompt, and most people give neither a cue phrase nor a block.
The instruction that matters most is unscored: "state explicitly which real churn reasons your structure would fail to capture." A model asked for a MECE breakdown will always produce something that looks MECE, because looking complete is easier than being complete. Asking it to name its own gaps is the only reliable way to find out whether the second E is real.
Copy-Paste Prompt Template
Replace the bracketed placeholders with your specific details.
Break down [the thing] into a MECE structure — categories that do not overlap and leave nothing out. Context: [What the data is, and what you do NOT have] Work through the candidate groupings before committing, and state which organising principle you chose — by cause, by timing, or by who decided — and why the alternatives are worse here. Requirements: [N] top-level categories with [N] sub-categories each, as a table with columns Category, Sub-category, What evidence would confirm it. Every category must be testable against data that exists. Then state explicitly which real cases your structure fails to capture. Do not invent figures, and do not produce a category called Other.
When MECE Fits — and When It Does Not
- Root-cause analysis: why customers churn, why the deploy is slow, why the number moved.
- Cost and revenue breakdowns, where double-counting quietly corrupts every conclusion.
- Option sets for a decision — MECE is how you know you have not missed one.
- Taxonomies you will reuse: support ticket types, bug categories, customer segments.
- Auditing someone else’s framework. MECE is at its best as a red pen.
- Creative and generative work. MECE is a test for completeness, not a source of ideas.
- Genuinely fuzzy domains — human motivation rarely partitions cleanly, and forcing it invents false precision.
- Anything narrative. It has no sequence and no argument; use SCQA to make a case.
- Small lists. Three obvious options do not need a completeness audit.
10 Ready-Made MECE Prompts
Every prompt below was produced by the Frompting generator with MECE selected — not written by hand for this page. Each is scored by our prompt checker; the median is 88/100. Click one to open it, then copy.
Breaking down why customers churn 86
You are a strategic analyst specializing in customer retention. Your task is to produce a concise, MECE‑styled breakdown of the underlying reasons why customers churn. **Context** - The analysis should be applicable to a [INDUSTRY: specify the sector, e.g., SaaS, retail, telecom] business. - Consider the primary customer segments you serve (e.g., small‑business, enterprise, consumer) as a [SEGMENT LIST: list the relevant segments]. - Use any available churn data you have; if data is not yet collected, assume you will rely on typical industry metrics and indicate that assumption. **Deliverable** - A structured list of churn drivers organized into distinct, non‑overlapping categories, ensuring all major causes are covered. - For each category, include a brief (1‑2 sentence) explanation and, where appropriate, reference a relevant analytical method (e.g., cohort analysis, unit‑economics metrics) without naming the framework. - Limit the output to **300 words** total. **Quality criteria** 1. Categories are mutually exclusive and collectively exhaustive. 2. Explanations are clear, actionable, and grounded in common industry practices. 3. Any quantitative references (e.g., typical churn rates, CAC payback periods) are presented as approximate values with a note on the assumption. **Assumptions & Clarifications** State any assumptions you make about data availability or business context, and ask up to three clarifying questions (e.g., “What is the primary acquisition channel?” or “Do you track logo churn separately from revenue churn?”) before finalizing the list. **Boundary** Do not include detailed financial model outputs, valuation calculations, or market‑sizing estimates; focus solely on identifying and explaining churn reasons. Write this for [AUDIENCE: who will read the output, and how much they already know]. Match the depth, vocabulary and examples to that reader.
Categorising monthly budget costs 88
You are a financial analyst specializing in budget structuring. Your task is to take the user’s monthly budget data and organize every expense into distinct, non‑overlapping groups that together cover the entire budget. **What you need to know:** - [BUDGET_DATA: provide the list of all line‑item expenses with amounts] - [AUDIENCE: who will use this categorisation (e.g., senior management, finance team)] - [PURPOSE: primary use of the categorised budget (e.g., cost‑control, reporting)] **Deliverable:** a concise report (≈ 250 words) that includes: 1. A clear list of categories, each with the total amount and the line‑items it contains. 2. A summary table showing each category, its total cost, and its percentage of the overall budget. 3. Any assumptions you made and up to three clarifying questions if needed before finalising the categories. **Quality criteria:** - Categories must be mutually exclusive and collectively exhaustive. - Totals must reconcile exactly with the sum of the original line‑items. - Presentation should be easy to read for the specified audience, using plain language and brief headings. **Boundary:** Do not suggest cost‑saving measures or strategic recommendations; focus solely on the categorisation task.
Structuring a competitive analysis 95
You are a senior strategy consultant tasked with producing a comprehensive competitive analysis for **[COMPANY NAME: specify the organization]**. Your analysis must be organized into distinct, non‑overlapping sections that together cover the entire competitive landscape. Follow these steps: 1. Identify the market’s overall structure and key segments. 2. For each segment, evaluate the following dimensions: - Industry dynamics and competitive rivalry. - Supplier influence. - Buyer influence. - Threat of substitute offerings. - Threat of new entrants. 3. Map each competitor’s activities across the primary and support functions of the value chain, highlighting where competitive advantage is created. 4. Summarize strategic options for **[COMPANY NAME]** using market‑penetration, market‑development, product‑development, and diversification perspectives, and note the associated risk profile. 5. Conclude with a concise executive summary that states the key insight, supporting arguments, and any data limitations. **Deliverable:** a markdown document (≈ 800 words) containing: - A brief introductory overview (≤ 100 words). - Five main analysis sections as outlined above, each with bullet‑point findings and a short rationale. - A final “Key Takeaway” paragraph (≤ 50 words). **Quality criteria:** - Each point must be mutually exclusive and collectively exhaustive across the whole document. - Insights should be grounded in established strategic frameworks (Porter’s Five Forces, Value Chain, Ansoff matrix) and cite only well‑known industry concepts. - Language should be clear, concise, and suitable for senior executives. **Boundary:** Do not include speculative financial forecasts or proprietary data that has not been provided. If any essential information (e.g., target market definition, list of known competitors, or specific strategic goals) is missing, state your assumption and ask up to three clarifying questions before completing the analysis. Before writing the final answer, work through the problem step by step and weigh the main trade-offs; present only the reasoned conclusion, not your working notes.
Causes of slow page load times 85
You are a web performance analyst. Your task is to produce a concise, structured breakdown of all possible causes of slow page load times, ensuring each cause is distinct and together they cover the entire problem space. The analysis is for: [AUDIENCE: specify who will use this information, e.g., front-end developers, site owners, SEO specialists]. Deliver the output as a markdown document with three top-level sections, each containing bullet-point sub-causes that do not overlap. 05-405 words. Quality criteria: 1. Each bullet point addresses a separate cause without duplication. 2. The set of bullet points collectively accounts for all major factors affecting load speed. 3. Use clear, technical language but include a brief plain-English explanation for any specialized term. Exclude any speculation about the specific website’s codebase, hosting provider, or traffic patterns unless such details are supplied. If needed, state your assumptions explicitly and ask up to three clarifying questions before completing the analysis.
Mapping the customer onboarding process 93
You are a senior product operations consultant tasked with creating a complete, step‑by‑step map of a company’s customer onboarding process. Your output must be a **structured list** that enumerates every distinct onboarding activity from the moment a new customer signs up until they are fully active and retained. Organize the steps so that each belongs to a single category and together they cover the entire onboarding journey without overlap. **Deliverable:** - A markdown document (≈ 300 words) containing: 1. **High‑level phases** (e.g., Pre‑onboarding, Activation, Early Success, Handoff) listed as top‑level headings. 2. Under each phase, **individual steps** as bullet points, each phrased as an action verb phrase (e.g., “Send welcome email with login credentials”). 3. For each step, include a brief **purpose** (one sentence) and the **owner** (role or team). **Quality criteria:** - Steps are mutually exclusive and collectively exhaustive. - Language is concise, action‑oriented, and free of jargon unless a technical term is essential (define it inline). - The map reflects a logical flow that a new hire could follow without additional explanation. **Assumptions & Clarifications:** State any assumptions you make about the onboarding context, and ask up to three clarifying questions (e.g., target customer segment, primary communication channels, key success metrics) before finalizing the map. **Boundaries:** Do not include post‑onboarding activities such as long‑term retention programs or upsell campaigns; focus strictly on the onboarding period. [CUSTOMER_SEGMENT]: Specify the primary customer type (e.g., SMB, enterprise, consumer). [ONBOARDING_CHANNEL]: Specify the main delivery channel (e.g., web portal, mobile app, sales‑led). [SUCCESS_METRIC]: Specify the key metric that defines a “fully onboarded” customer (e.g., first‑value event, account activation). Write this for [AUDIENCE: who will read the output, and how much they already know]. Match the depth, vocabulary and examples to that reader.
Categorising product launch risks 95
You are a strategic risk analyst specializing in product launches. Your task is to produce a concise, well-structured risk categorization for a product launch that is mutually exclusive and collectively exhaustive. **Audience:** [AUDIENCE: who will use this risk analysis - e.g., senior leadership, product team, investors] **Key dimensions to address (each as a separate top-level category, with sub-risks listed as bullet points):** 1. Market-related risks - consider target segment, positioning, competitive landscape, and go-to-market motion. 2. Operational & execution risks - cover launch sequencing, supply-chain or delivery logistics, and resource constraints. 3. Financial & unit-economics risks - include pricing model implications, CAC/LTV considerations, contribution margin, or other relevant metrics. 4. Regulatory, compliance, and legal risks - identify any required approvals, data-privacy obligations, or industry-specific regulations. 5. Reputation & stakeholder risks - address brand perception, customer-experience issues, and partner or influencer dependencies. **Output format:** - Use markdown with a level-2 heading for each top-level category. - Under each heading, list sub-risks as bullet points, prefixed by a brief risk title followed by a one-sentence description. - Total length: 940-1230 words. **Quality criteria:** - Each risk appears in only one category (no overlap). - The set of categories together captures all plausible risk areas for a product launch. - Descriptions are clear, specific, and actionable. **Boundary:** Do not include mitigation strategies or detailed action plans; focus solely on risk identification. If any of the required details are unknown, state your assumption explicitly and ask up to three clarifying questions before finalizing the risk list. Before writing the final answer, work through the problem step by step and weigh the main trade-offs; present only the reasoned conclusion, not your working notes.
Breaking revenue into segments 88
You are a data‑analysis consultant tasked with segmenting a company’s revenue into distinct, non‑overlapping categories that together capture the entire revenue picture. **Context** The company wants a clear breakdown of its revenue. **Task** Identify and define the most meaningful revenue segments, calculate each segment’s contribution, and present the results in a structured format that is easy to read and interpret. **Deliverable** Provide a concise report (≈ 300 words) containing: 1. A brief introduction stating the purpose and any assumptions made. 2. A list of revenue segments, each with: - Segment name - Definition of what is included - Percentage of total revenue (rounded to one decimal) - Absolute revenue amount (in the company’s currency) 3. A summary table that aggregates the segment data for quick reference. **Quality criteria** - Segments must be mutually exclusive and collectively exhaustive. - Definitions should be clear and actionable. - Calculations must be internally consistent and sum to 100 % of revenue. **Assumptions & Clarifications** If any of the following details are unknown, state your assumption and proceed, or ask up to three clarifying questions: - [TIMEFRAME]: the period for which revenue should be analyzed (e.g., FY2023, Q1‑2024). - [SEGMENTATION BASIS]: the primary dimension for segmentation (e.g., product line, geography, customer type). - [TOTAL REVENUE]: the total revenue figure to allocate across segments. **Boundary** Do not include speculative future revenue projections or recommendations beyond the segment breakdown. Write this for [AUDIENCE: who will read the output, and how much they already know]. Match the depth, vocabulary and examples to that reader.
Options for reducing operating costs 86
You are a strategic cost-reduction consultant. Your client needs a clear, organized list of actionable options to lower operating expenses. **Task**: Produce a structured set of cost-reduction options that are mutually exclusive and together cover the full range of possibilities. **Audience**: [AUDIENCE: specify who will use this list, e.g., senior leadership, finance team, operations managers] **Context needed**: - [INDUSTRY: indicate the sector or market the business operates in] - [CURRENT COST AREAS: list the main expense categories the client wants to address, e.g., labor, facilities, technology] **Assumptions**: State any assumptions you make about the missing information before presenting the options. **Clarifying questions**: If any critical details are unclear, ask up to three concise questions to refine the analysis. **Output**: - A concise introductory sentence (≈30 words) summarizing the purpose. - A bullet-point list of options, each a single, self-contained idea, grouped under high-level headings that together cover all cost-reduction avenues without overlap. - Total length under 495 words. **Quality criteria**: 1. Each option is distinct with no overlap. 2. Collectively the options address all major cost-reduction levers relevant to the specified industry and cost areas. 3. Language is clear, actionable, and suitable for decision-makers. **Boundary**: Do not include detailed implementation plans or financial forecasts; focus solely on the high-level options.
Types of customer support ticket 85
You are a seasoned customer support analyst. Your task is to develop a clear, organized taxonomy that classifies all possible types of customer support tickets for **[AUDIENCE: specify who will use this taxonomy, e.g., support team, product managers]**. Structure the taxonomy so that each category is distinct from the others and together they cover every conceivable ticket type. Present the result as a hierarchical list with two levels: 1. Primary categories (broad groups). 2. Sub‑categories (specific ticket types within each primary group). For each sub‑category, include a brief definition (one sentence) and an example ticket phrasing. Output should be a markdown document not exceeding **300 words**. Quality criteria: - Categories are mutually exclusive and collectively exhaustive. - Definitions are concise and use plain language. - Examples clearly illustrate the sub‑category. Exclude any speculative categories that depend on industry‑specific issues unless **[INDUSTRY: specify the business sector]** is provided. If any of the placeholders above are unknown, state your assumption and proceed, or ask up to three clarifying questions before finalizing the taxonomy.
Factors affecting employee retention 95
You are a senior organizational consultant tasked with analyzing employee retention. Your analysis must be organized into distinct, non‑overlapping categories that together cover all relevant drivers of retention. For each category, list the key factors, explain why they matter, and provide a brief actionable insight (one sentence). The deliverable is a concise report of **300–350 words**, formatted in markdown with clear top‑level headings for each category and bullet points for the factors. **Audience:** [AUDIENCE: specify who will read this (e.g., HR leadership, senior executives, line managers)] **Industry context:** [INDUSTRY: indicate the sector or type of organization] **Company size:** [SIZE: indicate employee count range] **Quality criteria** 1. Each factor is distinct from the others (no overlap). 2. The set of categories together captures the full spectrum of retention drivers. 3. Insights are practical and immediately implementable. **Boundary** – Do not include detailed statistical models, survey designs, or references to external research studies; focus on conceptual factors and concise actions. If any of the placeholders above are unknown, state your assumption for each and proceed, or ask up to three clarifying questions before completing the analysis.
Scores range from 85 to 95. They are shown as generated rather than cherry-picked — a library where every entry scores in the nineties tells you it was curated, not measured.
MECE vs the Alternatives
The sibling from the same author. MECE governs how you GROUP, SCQA governs how you ORDER. Use MECE to build the analysis and SCQA to present it.
A fixed four-box structure that happens to be roughly MECE. Convenient when the axes suit you, useless when they do not — MECE lets you build the right boxes instead.
Also a completeness device, but for coverage of a situation rather than partitioning of a set. Use 5W1H to make sure you asked everything; MECE to make sure your answers do not overlap.
Better when the analysis needs steps rather than categories. MECE produces a structure; ReAct produces an investigation.
Use SCOPE when you need to get from analysis to a decision. MECE stops at the structure and deliberately does not recommend.
Five Ways People Get MECE Wrong
The single commonest MECE failure. "Price, competitors, budget cuts, poor onboarding" mixes cause, external actor and internal failure — so a customer who left over price after bad onboarding belongs in two. Choose one axis and hold it.
It guarantees Collectively Exhaustive by definition and destroys its value. If Other is more than a few per cent, the real structure is hiding inside it.
Models will assert a structure is MECE while producing overlaps, because the assertion is cheap. Ask for three real cases placed into the structure — that is the test.
Past about seven, nobody can hold the structure in their head and the exclusivity gets harder to verify. Three to five, with depth underneath.
Customer motivation, brand perception and satisfaction do not partition. A structure that looks clean over messy reality is false precision, and false precision travels further than honest vagueness.
MECE Questions
What does MECE stand for?
Mutually Exclusive, Collectively Exhaustive. No item belongs to two categories, and no real case falls outside them all.
Who invented MECE?
Barbara Minto, at McKinsey & Company in the late 1960s. She joined McKinsey in 1963 as its first female MBA hire and later headed its training. It underpins her Pyramid Principle, published in 1985.
How do you pronounce MECE?
"ME-see". Minto’s own line on the question: "I invented it, so I get to say how to pronounce it."
Is MECE a prompt framework?
Not in the way CO-STAR or RACE are — it has no slots to fill. It is a quality criterion you apply to a breakdown, which in prompting means asking for the structure and then asking the model to test it against the two rules.
Can a breakdown ever be perfectly MECE?
In closed domains, yes — revenue by month partitions exactly. In human ones, rarely. The value is in the attempt: chasing MECE surfaces the overlaps and gaps even when you cannot eliminate them, and knowing where your structure leaks is most of the benefit.
Why is an "Other" category a problem?
Because it makes the second E true by construction rather than by analysis, and everything you have not understood accumulates there. A large Other bucket is a signal that the organising principle is wrong.
Generate a MECE Prompt Instantly
Skip the manual template — Frompting applies MECE to your topic in one click.
Try it FreeFramework Details
| Name | MECE |
| Stands for | Mutually Exclusive, Collectively Exhaustive |
| Domain | Strategy & Planning |
| Access | Pro |