Ansoff Matrix — Weighing Growth Directions Systematically

A two-by-two matrix that orders growth options along two axes — existing or new products, existing or new markets.

Purpose

The Ansoff Matrix structures the question of where future growth should come from into four basic directions: market penetration (existing products in existing markets), market development (existing products in new markets), product development (new products in existing markets), and diversification (new product, new market). Its value lies in disciplining the discussion: growth ideas are not judged one by one but as positions on a risk gradient — the further from the familiar on both axes, the greater the uncertainty.

The matrix is intended as a frame for portfolio decisions about growth initiatives: it forces the distribution of bets across the four fields into the open, to be reconciled with risk appetite and capabilities. It goes back to a 1957 article by Igor Ansoff, making it one of the oldest strategy tools still in use. It owes its longevity to its simplicity — which is also its most criticized property.

Procedure in five steps

Define the baseline
First, determine precisely what counts as an existing product and an existing market — by segment, region, and use case. This boundary work is not a formality: whether a neighboring country is a new market or an extension of the current one changes the risk assessment considerably.
Generate options per field
Concrete growth options are developed for all four fields, not just the comfortable ones. Thinking through inconvenient fields such as diversification has diagnostic value even if nothing there is ever executed. Each option is sketched with revenue potential, time horizon, and required capabilities.
Assess risk and capability gaps
Each option is examined for its distance from the core business: what market knowledge and which technical and operational capabilities are missing? The matrix's risk gradient serves as a heuristic but is made concrete per option — market development with a strong local partner can carry less risk than aggressive penetration against a dominant incumbent.
Balance the portfolio
The options are then viewed as a whole: how do resources spread across the four fields, and does that spread match earnings pressure and risk-bearing capacity? A pure penetration portfolio signals a growth ceiling; a diversification-heavy one signals overreach. The balance is the real decision, not any single option.
Commit and revisit
For the chosen options, milestones and kill criteria are defined, especially for ventures in the riskier fields. The matrix is redrawn at regular intervals, because the fields shift: a newly entered market becomes an existing one and resets the baseline for the next round.

Strengths and limits

The matrix communicates within minutes where an organization seeks growth and what risk profile that choice implies.
It exposes lopsided portfolios — such as the common fixation on penetration while the core market stagnates.
The risk gradient across the fields provides a first, robust heuristic for allocating resources across growth bets.
Its simplicity allows use at any level, from the corporate whole down to a single product line.

The matrix evaluates nothing — it sorts options but says neither which is attractive nor how to execute it.
Competitors, industry dynamics, and timing do not appear in the grid, although they decide whether a path succeeds.
The existing/new dichotomy is often impossible to draw cleanly in platform and ecosystem markets, inviting false clarity.
The blanket risk gradient can mislead: staying in the core business is sometimes the riskiest option of all in disruptive periods.

With AI and agents

The step that LLMs change most is option generation with first-pass assessment. For market development, target markets can be scanned systematically — regulation, competitive density, price levels, distribution structures — and condensed into comparable short dossiers where a costly country study used to be needed for each. For product development, agents can mine customer feedback, support data, and patent landscapes to substantiate gaps in the offering instead of guessing at them. The matrix fills up with researched candidates rather than with whichever ideas happen to be in the room.

Risk assessment per field also gains resolution. The classical heuristic — diversification is riskiest — can be replaced per option with evidence: documented market entries by comparable firms, typical failure causes, realistic timelines to profitability. Agents can also take over monitoring: instead of refilling the matrix annually, indicators per growth path — demand signals, competitor entries, regulatory movement — are watched continuously and deviations flagged, so that kill criteria actually bite.

What cannot be delegated is the portfolio decision itself. How much risk an organization can carry depends on its balance sheet, culture, and ownership — factors a model can describe but not weigh. And with AI support the matrix invites a new failure mode: because convincing dossiers can be produced quickly for every field, all paths look feasible at once. The art of leaving things out — the actual core of any growth strategy — becomes harder with better research, not easier.

Relation to scenarios

In relation to scenario thinking, the Ansoff Matrix works in both directions. Scenarios change the attractiveness of the fields: a future of fragmenting trade blocs devalues market-development paths, while one of technological upheaval in the core business turns diversification from optional into mandatory. It therefore pays to fill in the matrix separately per scenario and identify which growth options hold up across several futures — those robust paths deserve priority. Conversely, the risky fields of the matrix mark exactly the ventures for which a scenario check before investment is indispensable.

See the scenario analysis method page

How ready is your organization for AI and agents? — Take the AI Readiness Check