The Ansoff matrix: how to compare growth options, with an example

Who should use it, and when not to

Use the matrix in annual or quarterly planning when the question is "where does next year's growth come from?", when someone proposes a new product or a new city and the team wants to compare it with the alternatives, or when growth is slowing and nobody knows whether the current market is used up. It is quick, and it forces a team to look at the plain options before the exciting ones.

It is not a forecast of how the world will change; for that, use scenario planning. It also says nothing about whether an industry is profitable or how to stand out within it. Pair it with other tools for those questions.

The core concepts

Two questions, four routes. Is the product one we already sell? Is the market one we already serve? The answers give the four quadrants in the short answer above.

"Market" means a customer group, not an industry. A market is a set of buyers in a place reached through a channel: independent cafés in Leeds supplied by van, or home buyers in the north of England reached online. Define it too broadly ("coffee") and every idea looks like penetration.

"New" means new to the company. A product can be old to the world and still new to you, with all the learning that brings.

Risk rises with distance. Each step away from the current quadrant adds things the company has to learn: a product, a buyer, a place, a channel. Penetration builds on what is known; diversification changes both product and market at once. Diversification can be related (sharing customers, skills, or channels with the current business) or unrelated.

Quadrants have blurry borders. John Dawes (2018) points out that a product new to the firm often brings new buyers too, so the line between product development and diversification isn't always clear. Classify an option by what is actually new, and judge its risk by the evidence, not only by its box.

A worked mini example

An illustration with made-up facts. Northfold Roasters in Leeds earned £640k last year: 70% from wholesale beans to 48 independent cafés in Leeds and York, 30% from about 900 online subscribers. It wants 25% more revenue next year (about £160k) with no outside money. Facts: about 4% of subscribers leave each month; the account manager estimates 150 or more local independent cafés it doesn't supply; 11 cafés asked for milk and syrups; offices have asked about coffee machines; a Manchester group of six cafés approached it; someone suggested opening its own café. The roaster runs at about 60% of capacity.

The shortlist is the two penetration options plus Manchester as the one new market. Whether they reach £160k depends on the average café's spend and the reasons subscribers leave, which the tests are built to find out.

How to run it

1. Define the base. List the current products and the current markets, each market as a customer group, a place, and a channel.
2. Check penetration headroom. Estimate how much more current products could sell to current markets: unreached buyers, more frequent buying, lost customers won back.
3. Generate options in all four quadrants. Write one to three real options per quadrant, or say why a quadrant has none.
4. Write what is new to the company. For each option, list the product, buyer, place, channel, rule, or skill the company would have to learn.
5. Rate evidence and risk. Note the evidence of demand and the main risk, and rate risk Low, Medium, or High, starting from the quadrant and adjusting for the facts.
6. Shortlist and say what you won't do. Pick two or three options by contribution to the goal against risk and resources, and name the options set aside.
7. Test each shortlisted option. Design a small, cheap test of its riskiest question, with a pass line written before it starts.

Common mistakes

• A market drawn as an industry. "Coffee" hides the difference between cafés in Leeds and offices in Manchester.
• Skipping penetration. Teams jump to new products or places before learning how much the current market still holds.
• Filling every box for symmetry. An option no one would fund doesn't belong on the grid.
• Rating risk by the box alone. A product customers already asked for can be safer than a market nobody has checked.
• Stopping at the matrix. The output is a shortlist, a decision, and tests, not a filled-in grid.

If the question is how to differ within a crowded market, the ERRC grid and strategy canvas fit better.

Where it comes from

H. Igor Ansoff published the product-market grid in "Strategies for Diversification," Harvard Business Review 35(5), September–October 1957, pp. 113–124, and developed it in his book Corporate Strategy (McGraw-Hill, 1965). This page is an independent summary in our own words, not affiliated with or endorsed by Harvard Business Review. More frameworks are in the strategy book guides.

Questions

What are the four strategies in the Ansoff matrix?

Market penetration (current products, current markets), market development (current products, new markets), product development (new products, current markets), and diversification (new products, new markets).

Which Ansoff strategy is the riskiest?

Diversification, because both the product and the market are new to the company. Market penetration is the least risky, since it builds on what the company already sells and knows.

Who created the Ansoff matrix?

H. Igor Ansoff, in the article "Strategies for Diversification" in Harvard Business Review (September–October 1957). He developed it further in his 1965 book Corporate Strategy.