Ansoff Matrix: How to Choose a Growth Strategy Grounded in Customer Research (2026)

The Ansoff Matrix is a growth-strategy framework that maps four ways to grow — market penetration, product development, market development, and diversification — across two axes: products (existing vs. new) and markets (existing vs. new). Its core insight is that risk climbs as you move away from what you already know, so the safest growth reuses your existing product in your existing market, and the riskiest ("diversification") builds a new product for a new market at the same time. The matrix tells you which directions you could grow; only customer research tells you whether each direction has real demand behind it.

Created by H. Igor Ansoff — the mathematician and strategist often called the "father of strategic management" — the framework first appeared in his 1957 Harvard Business Review article "Strategies for Diversification" and was expanded in his 1965 book Corporate Strategy. Sixty-plus years later it remains one of the most widely taught tools for structuring a growth conversation. This guide explains each quadrant, the risk logic behind it, its limitations, and — most importantly — how to replace the guesswork inside the matrix with evidence from real customers.

The four quadrants

The Ansoff Matrix is a 2x2 grid. One axis is your product (existing or new); the other is your market (existing or new). The four cells are four distinct growth plays, in ascending order of risk:

1. Market Penetration (existing product to existing market) — lowest risk. Sell more of what you already have to the customers you already serve. Tactics include increasing usage frequency, winning share from competitors, improving retention, and raising purchase volume. Because both the product and the market are known, execution risk is lowest. This quadrant is where disciplined teams look first.

2. Product Development (new product to existing market) — moderate risk. Build new products or features for customers you already understand. The market is known, but the product is unproven, so the central question becomes: does this audience actually want what we are about to build? This is where concept testing and discovery interviews earn their keep.

3. Market Development (existing product to new market) — moderate risk. Take a proven product into new segments, geographies, verticals, or use cases. The product works; the uncertainty is whether a new audience has the same need, buys the same way, and values the same benefits.

4. Diversification (new product to new market) — highest risk. Build something new and sell it to people you do not yet serve. You are stacking two unknowns at once, which is why commentators nickname this the "suicide cell." Diversification can deliver the biggest upside — an entirely new engine of growth — but it also carries the highest odds of failure and demands the most rigorous validation.

Ansoff's own label for the tool was the "product-market growth matrix"; the familiar four-box visual and the "Ansoff Matrix" name are later popularizations built on his 1957 foundation.

Why risk climbs across the grid

The matrix is really a map of what you do not know. Every move away from your current product-market position introduces a new source of uncertainty — a new build, a new buyer, or both. Market penetration asks you to execute better at something you already do. Diversification asks you to be right about a product and a market simultaneously, with no track record in either.

The stakes are not academic. In CB Insights' 2026 analysis of 431 venture-backed companies that shut down, 43% failed because of poor product-market fit — the single most common root cause once you look past "ran out of cash," which the report itself calls the final symptom rather than the disease (CB Insights, The Top Reasons Startups Fail, 2026). An earlier CB Insights study of startup post-mortems put "no market need" at the top at 42% (CB Insights, 2014). The pattern is consistent across a decade: most growth bets die not because the product could not be built, but because too few people wanted it.

It is worth retiring one popular scare statistic here. The oft-repeated claim that "95% of new products fail," frequently attributed to Harvard's Clayton Christensen, is an urban legend — Christensen denied ever making it, and a peer-reviewed review by Castellion and Markham (Journal of Product Innovation Management, 2013) found empirical failure rates closer to 40%. The honest takeaway is not that growth is hopeless; it is that the failure rate is high enough that validating demand before you commit is the difference between the two outcomes.

The trap inside every quadrant: assumed demand

The Ansoff Matrix does one job brilliantly: it forces a structured conversation about where to grow. What it cannot do is tell you whether the demand you are counting on actually exists. Every box hides an assumption:

• Penetration assumes non-buyers can be converted and light users can be grown.
• Product Development assumes your existing customers want the new thing.
• Market Development assumes the new segment shares the same need and buying behavior.
• Diversification assumes a brand-new audience has a problem your brand-new product solves.

Ansoff himself warned against the opposite failure mode — endless deliberation. He popularized the phrase "paralysis by analysis" in Corporate Strategy (1965) to describe teams so buried in strategic analysis that they never decide. The goal, then, is not more analysis; it is better evidence, faster — enough signal to move with confidence without stalling. That is exactly where modern, AI-native research changes the economics.

How Koji grounds each quadrant in real customer evidence

Traditionally, validating an Ansoff move meant weeks of recruiting, scheduling, and manually running interviews — the kind of delay that pushes teams to skip research and "just decide." Koji collapses that timeline. As an AI-native research platform, Koji runs AI-moderated interviews (voice or text) with dozens of customers or prospects in parallel, automatically probes every answer with intelligent follow-ups, and delivers thematic analysis and a shareable report in hours instead of weeks. Teams using AI-assisted research routinely compress time-to-insight from weeks to days.

Here is how that maps onto the matrix:

• Market Penetration: Interview churned and low-frequency users to learn what blocks more usage. Koji's structured questions — Koji supports six types: open_ended, scale, single_choice, multiple_choice, ranking, and yes_no — let you quantify barriers (a scale question on satisfaction, a single_choice on the top reason for low usage) while the AI probes the why behind each rating.
• Product Development: Run concept tests with your existing base. Use ranking questions to force trade-offs between candidate features, and let Koji's follow-up probing surface the reasoning a static survey would miss.
• Market Development: Screen and interview the new segment before you invest. A yes_no plus scale battery reveals whether the need transfers; open-ended probing reveals whether they buy the same way.
• Diversification: This is the highest-risk box, so it deserves the deepest validation. Koji lets you run discovery interviews with true noncustomers cheaply enough to kill a bad idea before it becomes a budget line.

Because Koji democratizes this — you do not need a dedicated research team or a PhD in methods — a product manager can validate an Ansoff bet the same week it is proposed. While traditional survey tools like SurveyMonkey capture what people clicked, an AI-native platform captures why, which is the exact signal the matrix is missing.

A worked example

Imagine a B2B scheduling app with strong retention among clinics. Leadership debates two Ansoff moves: build an analytics module for existing clinics (product development) or take the current app into law firms (market development).

Instead of arguing from opinion, the team launches two Koji studies in a single afternoon. The product-development study asks 25 existing clinic admins to rank five possible modules and rate how painful their current reporting is; analytics ranks third, behind a billing integration nobody had prioritized. The market-development study screens 20 legal-office managers: a yes_no on "do you struggle with client scheduling" comes back 70% yes, but open-ended probing reveals their real bottleneck is conflict-checking, which the app does not do.

Within days, the team has redirected the roadmap toward a billing integration (a development play with proven demand) and shelved the law-firm expansion until the product gap is addressed — two decisions that would otherwise have cost a quarter of engineering to learn the hard way.

Limitations to keep in mind

The Ansoff Matrix is a starting point, not a verdict. Its 2x2 simplicity is also its weakness: "new vs. existing" is really a spectrum (adjacent segments, partial product overlaps), and the grid says nothing about whether you have the capabilities, capital, or timing to execute. Diversification is not always riskier in practice — a strong adjacency can be safer than defending a declining core. Use the matrix to generate options and structure the debate, then pair it with capability analysis, a portfolio view, and — above all — primary customer evidence before you commit resources.

Related Resources

• Structured Questions Guide — the six question types that quantify demand inside any Ansoff quadrant
• SWOT Analysis with Customer Research — pair internal strengths with external growth options
• Porter's Five Forces for Market Research — assess industry attractiveness before you enter a new market
• TAM SAM SOM for Product Researchers — size the opportunity behind a market-development move
• Go-to-Market Research Guide — de-risk the launch once you have picked a quadrant
• How to Find Early Adopters — recruit the first customers for a diversification bet
• Concept Testing Methodology — validate a new product before you build it