Classic · Model
The Ansoff matrix.
Igor Ansoff divided growth into four routes depending on whether you're selling something existing or something new, and whether you're doing it to existing or new customers. What makes the matrix useful isn't the four boxes in themselves, it's that the risk rises noticeably from one box to the next. Selling more of what you already have to the people who already buy is a completely different thing from building something new for someone you don't know, and the two often get mixed together under the same word, growth.
The model
Ansoff published the division in Harvard Business Review in 1957, in the article Strategies for Diversification. Four routes, in rising order of risk.
Market penetration. Existing offer to existing customers, meaning more of the same to the people who already buy. Lowest risk, because you already know both the product and the customer.
Market development. Existing offer to new customers, meaning a new industry, a new geography or a new size of company. You know the delivery but not the buyer.
Product development. New offer to existing customers. You know the buyer but you have to build something you haven’t built before.
Diversification. New offer to new customers. Highest risk, because nothing is known, and it’s the only one of the four where a failure teaches you nothing you can use in the rest of the company.
That the risk rises between the boxes is the model’s real point, and that’s why the dots in the picture get more numerous. The order of the two middle ones is disputed, though, meaning whether it’s harder to take something existing to a new market or to build something new for customers you already have. Which of them weighs heaviest in your company depends on whether you’re stronger in sales or in delivery.
How it’s used
By placing what you do today and what you’re planning in the boxes, and then looking at the distribution. If all the plans sit in the bottom right box, you don’t have a growth plan, you have several parallel start-ups.
A reasonable question is also what each box costs in management attention, and not only in money. Diversification requires someone in management to be present in something new over a long period, and that’s the resource that runs out more often than the money does.
The model connects to The core of a strategy in a straightforward way: choosing one box means deselecting three, and a growth plan that doesn’t do that is an ambition.
In the work on the business plan
This is probably where the matrix comes into its own best. A business plan tends to contain a row of initiatives that each sound reasonable on their own, and that together are hard to judge because they’re described in running text and in different chapters.
Put those same initiatives in the boxes and two things become visible at once. One is the prioritisation, meaning where the weight actually lies once everything has been drawn up side by side. The other is the level of risk, because the boxes are themselves a risk scale, and an initiative that lands in the bottom right carries a completely different uncertainty from one that lands in the top left.
That makes it useful in the room rather than in the document. An owner, a board or a management team can look at the same picture and make their own assessment of likelihood and risk without having to read twenty pages first, and the objections that come tend to be concrete: that an investment is in the wrong box, that there are too many in the hardest one, or that the easiest box is standing empty even though it’s the only one you can carry out with the staffing you have.
It also makes it useful the year after, because the same picture can be redrawn and compared.
Where it usually gets misused
When the boxes are used to describe what you’ve already made up your mind about. The matrix can always be drawn so that your own proposal looks reasonable, and it then becomes support for the decision rather than a test of it.
It also gets misused when risk is read as something to avoid. It’s perfectly reasonable to move right or downwards in the matrix, but then it should be done with your eyes open and with a timeline that holds up for it taking longer than the first box.