Ansoff Growth Matrix

What this tool does

CM8-254 puts your growth options on the Ansoff matrix and makes you cost them. Every idea becomes a row: its quadrant, what it could earn in a mature year, what it will cost, how long until the first invoice, its risk, its capability fit, its owner and the decision taken. The tool ranks them, draws the grid and prints a growth plan. Its value is that four uncomfortable questions — how much of this is new, what does it cost, what could it earn, and can we do it — get answered in writing, side by side, before anybody spends money.

The four ways to grow

Two questions place every initiative: is the product new to us? and is the market new to us? The four quadrants that follow are ordered by how much is unfamiliar — which is the same as ordered by risk.

• Market penetration — existing products, existing markets. More of what you already make, to the people who already buy it: a second-source position at an account that knows you, the customers who quietly stopped ordering, a shorter lead time to catch rush work, a second shift. Nothing is new, so this is the lowest-risk quadrant by a wide margin and usually the cheapest — the investment is often time rather than capital.
• Market development — existing products, new markets. The product already works; the unknown is who else wants it — a neighbouring region, a new sector, a different channel. Moderate risk: production risk is near zero, but you are buying market knowledge you do not have, including whether they pay. Pilot it, costed end to end, before committing.
• Product development — new products, existing markets. The customers already trust you; the unknown is whether you can build the thing — finishing or assembly so you quote a finished item, or a service contract on a machine you sell. Moderate risk the other way round: demand is visible on their purchase orders, delivery is not. It fails when the product costs more to make than anyone will pay.
• Diversification — new products, new markets. Both unknowns at once: buying a business in an adjacent trade, or opening a line for customers you have never met. The highest-risk quadrant, because nothing you know reduces the uncertainty — not your process knowledge, not your relationships, not your reputation. Occasionally right, especially when your market is shrinking. Never the default.

The discipline of exhausting penetration first

This is the most valuable idea in the tool. Most firms chase diversification while leaving obvious share on the table with the customers they already have. The pattern is consistent: a growth conversation opens and within ten minutes the room is discussing a new sector, a new product line or an acquisition — because those are interesting, and because selling more to existing customers feels like an admission. Meanwhile the same business is turning away rush work on lead time, has lapsed accounts nobody has called, is second-source on parts it could be first-source on, and has never asked its three largest customers what else they buy that it could make. Each of those is revenue with no new product, no new market, no new approval and no new capability — a fraction of the risk and a fraction of the cost of the exciting option on the table.

The discipline is simple and unpopular: before funding anything in the other three quadrants, write down what is left in penetration and why you are not taking it. If the honest answer is that you have not tried, the plan is not ready. Two tests make it concrete. The share test: for each of your five largest customers, what proportion of their spend on things you could make do you hold? Not knowing is itself the first piece of work. The decline test: who ordered two years ago and not this year, and why? Both usually surface more addressable revenue than the new-sector idea, and both start next week without capital. This is an argument about sequence, not about the merits of the other quadrants: penetration returns more per unit of money and risk, and the cash it throws off is what pays for the riskier moves.

The two scores, stated honestly

The return multiple answers one narrow question: for every unit of money invested, how many units of annual revenue does this initiative claim to produce once mature? It shows an em dash wherever investment is blank or zero. The risk-adjusted score divides that multiple by the risk rating, so a cheap familiar move outranks an expensive unfamiliar one. It is a rough sort key, not a valuation:

• It uses revenue, not profit. A 4× multiple at 8% margin is worse business than a 2× multiple at 30%, and the tool cannot know your margins.
• It does not discount for time. Revenue arriving in thirty months counts the same as revenue arriving in six — which is what the months column is for.
• It does not discount for the probability of failure. Dividing by the risk rating is a crude proxy, not a probability, and it treats a judgement as a measurement — a rating set by the people who want the initiative approved, which is why the confidence column exists.

Deliberately absent: any payback-period figure. Deriving one needs a margin assumption, a ramp-up curve and a discount rate the tool does not have.

Capability fit — the question that kills more initiatives than money

Growth plans fail more often on capability than on funding. Money can usually be found; the approval, the skill, the relationship or the management attention often cannot, and none appear in a return multiple. Rate the fit against what you have today. Strong means the skills, plant, approvals and relationships already exist. Partial means named, closeable gaps — a shift leader to recruit, an audit to pass. Weak means a capability you do not have in any form: you are not growing, you are learning a trade while trying to sell its output. The tool will not accept an initiative marked weak and approved without a note, because somebody has to have written down how the gap closes and by when.

The diversification share — an overreach warning

The balance tile shows what proportion of the plan's revenue potential sits in the diversification quadrant, rejected initiatives excluded. It is the classic overreach signal, measured in the plan's own currency: if a third of the growth you are counting on comes from the quadrant where you know least, a third of your plan rests on your least reliable estimates. The threshold defaults to 25% and can be changed in the settings; the point is to fix a limit while the plan is being written. When it flags, the answer is not automatically to cut the initiative — it is to ask whether penetration is exhausted, whether that revenue figure has evidence behind it, and whether the plan survives the initiative returning zero.

Sequencing, and how many you can actually run

Sequence beats selection. Five initiatives run at once by the same four people deliver less than three run in order, because growth work competes with the day job and loses. The usual order: exhaust penetration, fund it from working capital, then one adjacent-quadrant move at a time, and diversification only once those run without daily attention.

The capacity setting asks how many you can genuinely run at once, and the headline tile compares it against how many are marked running. Most small firms should set two or three: the constraint is rarely money, it is the hours the few people capable of driving change have left. One relation worth naming once — the BCG growth–share matrix is about the portfolio you have today, Ansoff about the growth you might add tomorrow.

Review cadence

Revisit the plan quarterly and rebuild it annually. Three questions per running initiative: has the revenue estimate changed now we know more, has the investment estimate changed now we have started spending, and is the risk rating still honest? The second answer is almost always that it costs more than anyone thought — record it rather than absorbing it quietly, because that pattern across several initiatives is the most useful thing you will learn about your own estimating. Rebuild from scratch when something structural changes: a major customer won or lost, a competitor arriving, a change of owner.

Saving and printing

Everything is written to this browser's local storage as you type, and that storage belongs to one browser on one computer. Treat Export .json as the real save, restored anywhere with Import .json; Export CSV gives you the register for spreadsheet work; Reset asks twice, then erases everything, with no undo. Print Report produces the full pack from whatever the filter shows, and states that filter in the scope line.

FAQ

What counts as a new market? New to you, not new to the world. A different sector, region, customer size or channel is a new market if you do not sell there and do not know how buying works. If you already have customers in it, it is an existing market even if you only have two.

An initiative sits between two quadrants. Which do I pick? The riskier one. If it genuinely splits into two separable moves, make it two rows.

Why is revenue annual at maturity? So every row compares on the same basis. First-year revenue punishes anything with a long ramp; lifetime revenue rewards whoever assumes the longest life. The months column carries the timing this strips out.

Should rejected initiatives stay on the register? Yes. They are excluded from the totals and the quadrant summary, but recording what was rejected, and why, stops the same discussion happening again next quarter.

Accuracy & disclaimer

Everything here comes from estimates made before the work starts. Revenue potential is a belief about what an initiative could earn, investment required a belief about what it will cost, and the risk rating a judgement written as a number so it can be sorted. None is a measurement, and initiatives are routinely wrong on all three in the same direction — optimistic on revenue, light on cost, generous on risk. The matrix sequences options; it does not validate them. It cannot tell a costed business case from a hopeful one, and it is not financial, investment, accounting or business advice of any kind.