Ansoff Matrix Agent Skill

Ansoff Matrix (Evidence-Backed)

Purpose

Map a company's growth options across the Ansoff Matrix with evidence per quadrant: **use or gather

evidence → four quadrants with signals → risk-rated sequence → next-step options.** The four

quadrants — market penetration, market development, product development, diversification — organize

one question: *where does the next tranche of growth come from, and at what risk?* This is a research

instrument, not wishful whiteboarding: every candidate move must answer "what documented signal says

this demand exists?" And the close is a sequence, because growth options compound — penetration funds

development, and diversification bets the funding.

Input

Works best with: the company or product line seeking growth, its current core (who is served,

with what, at what scale — the matrix's axes are defined relative to it), and the growth outcome and

horizon on the table.

Also useful: constraints (capital, capability, risk appetite), and any research in session — a

landscape scan, five-forces read, or company-intel output lets the

matrix organize evidence instead of gathering it.

Input supplied inline with the invocation — text after the skill name, a pasted context dump, or an

appended ARGUMENTS: line — counts as answers already given. Use it against the question budget;

don't re-ask.

Arriving empty-handed? That works too. The skill opens with at most 3 questions (core, outcome

and horizon, constraints) and proceeds on labeled assumptions if they go unanswered.

Example invocation: `Ansoff growth options for our field-service product line — core: dispatch

software for mid-market HVAC firms, US. Outcome: +40% ARR in 24 months. Constraint: no acquisitions.`

Key Concepts

• Governing protocol: honors the autonomous-investigation

contract — question budget of 3, search-plan gate, Fact/Inference/Assumption labels, Just Enough

Mode (2-3 moves per quadrant), stable schema, 4-option Final Step.

• The framework (Ansoff, 1957): growth options plotted on two axes — existing vs. new *products*,

existing vs. new *markets*. Penetration (existing/existing) is the lowest-risk quadrant;

diversification (new/new) the highest, because it abandons both anchors of proven demand at once.

• The risk gradient is law. Penetration < market development ≈ product development <

diversification. A diversification move rated "low risk" needs extraordinary evidence — and the

gradient teaches why diversification proposals deserve the heaviest evidence burden and usually

arrive with the lightest.

• Signals, not wishes. Candidate moves come from documented signals: underserved-segment data,

expressed demand (voice-of-customer-miner themes),

competitor precedent, capability evidence. An empty diversification quadrant is an acceptable

answer; an invented one is not.

• Coaching vs. investigation — same map, different jobs:

organic-growth-advisor is the Interactive sibling that

*diagnoses your growth constraint* through questions (its Growth Path Matrix shares Ansoff's

axes); this skill *researches the evidence* for each quadrant's options. Diagnose there, evidence

here — they pair deliberately.

• When NOT to use: feature-level prioritization (this is portfolio altitude — use

feature-investment-advisor for a single build

decision); no growth mandate or capacity — an options map without an owner is a poster.

• Do-not-invent list: market sizes, adoption data, competitor results, demand claims. Where

sizing matters, flag it for tam-sam-som-calculator rather

than guessing.

Application

• Check session for existing evidence (landscape scan, five forces, company intel, VoC). Present

→ the matrix organizes it; search only gaps.

• Credit inline context, then ask only the unanswered questions (max 3):
• Which company or product line, and what is its current core?
• What growth outcome and horizon is on the table?
• Any constraints — capital, capability, risk appetite?
• If researching fresh, show the 3-bullet search plan — what you'll search per quadrant

(segment data, expressed demand, competitor precedents, capability signals), source types,

fact/inference separation. Continue unless revised.

• Populate the quadrants and emit the schema below exactly.

Output schema (do not reorder)

~~~markdown

Ansoff Growth Options: [Company / Product Line]

As-of date: | Current core: | Growth outcome sought:

1. Market Penetration (existing product, existing market — lowest risk)

• [Candidate move] — signal: [evidence, URL, label] — risk: [low/med/high, why]
• [2-3 moves]

2. Market Development (existing product, new market)

• [Candidate move: segment, geography, or channel] — signal: [evidence of underserved demand, URL, label] — risk: [rating, why]
• [2-3 moves]

3. Product Development (new product, existing market)

• [Candidate move] — signal: [expressed demand, VoC theme, competitor precedent, URL, label] — risk: [rating, why]
• [2-3 moves]

4. Diversification (new product, new market — highest risk)

• [Candidate move] — signal: [the extraordinary evidence this quadrant requires, URL, label] — risk: [rating, why]
• [1-2 moves; an empty quadrant is an acceptable answer]

5. Recommended Sequence (the "so what")

• First: [move] — because [evidence strength + funding logic]
• Then: [move] — funded/de-risked by the first
• Not yet: [the tempting move and why the evidence says wait]
• The assumption that breaks this sequence: [one line]

Assumptions to Validate

• [Assumption 1] / [Assumption 2] / [Assumption 3]

~~~

A copy/paste fill-in version of this schema, with quality checks, lives in template.md.

Final Step (offer exactly 4 options)

• Size the top move with TAM/SAM/SOM (tam-sam-som-calculator) (Recommended)
• Pressure-test the sequence with a premortem
• Deep-dive the diversification quadrant's evidence
• Convert the first move into an opportunity solution tree (opportunity-solution-tree)

Accept 1, 2, 3, 4, 1 and 2, Verbose Mode, or a custom path.

Examples

A quadrant entry earning its place (fictional):

> ## 2. Market Development

> - Adjacent trade: plumbing contractors, same size band — signal: plumbing firms appear

> unprompted in 14% of our category's review-site mentions asking "does this work for plumbing?" —

> Fact ([review threads, URLs]); the two incumbents serving plumbing both gate scheduling

> behind enterprise tiers — Fact ([pricing pages]) — risk: medium — demand signal is real

> but second-hand; sales motion transfers, integrations don't fully.

The sequence close doing its job:

> - First: win-back campaign into the churned-but-reachable base (penetration) — strongest

> evidence, funds everything else, 1-quarter payback

> - Then: plumbing-contractor entry (market development) — de-risked by the penetration win's

> cash and case studies

> - Not yet: the IoT hardware bundle (diversification) — one analyst mention and founder

> enthusiasm is not extraordinary evidence

> - The assumption that breaks this sequence: churned customers left for fixable reasons; if

> win-loss shows they left the *category*, penetration is a dead first move and development leads.

See examples/sample.md for a complete worked matrix (fictional FSM-software

market) with an honestly empty diversification quadrant and a sequence whose breaking assumption is

named. examples/sample-industrial.md shows the opposite lesson: a

populated diversification quadrant whose entry fails the evidence bar in writing.

Common Pitfalls

• The brainstorm grid. Four quadrants of unsourced ambition. Every move answers "what signal

says this demand exists?" or it doesn't ship — that single rule converts Ansoff from wall art

into an instrument.

• Risk-gradient denial. A diversification move rated low-risk on enthusiasm. The gradient is

the framework's whole teaching: new product *and* new market means both anchors are gone.

• Quadrant stuffing. Filling diversification because empty feels lazy. An honestly empty

quadrant is a finding; a padded one is a liability with a deadline.

• Options without sequence. A menu with no first move, no funding logic, no breaking

assumption. Growth options compound — order is the strategy.

• Sizing by vibe. Attaching invented market sizes to moves. The do-not-invent list routes

sizing to the TAM/SAM/SOM calculator, where the math shows its work.

References

• organic-growth-advisor (Interactive) — the coaching

sibling: diagnoses *which* growth path fits your constraint; this skill evidences the options

• autonomous-investigation (Workflow) — the governing protocol
• intelligence-collection-disciplines (Component) — signal sources per quadrant
• porters-five-forces (Workflow) — the profit-pool read that feeds this analysis
• tam-sam-som-calculator (Component) — sizes the moves
• voice-of-customer-miner (Workflow) — expressed-demand signals for product development
• opportunity-solution-tree (Interactive) — structures the first move's execution
• H. Igor Ansoff, "Strategies for Diversification" (Harvard Business Review, 1957)
• Adapted from market-intelligence/ansoff-matrix-prompt.md in the

https://github.com/deanpeters/product-manager-prompts repo.