The AI BCG Matrix Strategist skill for Claude

No software, no subscription, no link to your accounting: one text file of 1,009 words that tells Claude how to apply Bruce Henderson's Growth Share model to your portfolio, what input it needs per product, what guideline it uses when classifying, and in what fixed format the answer gets delivered. You download the zip at the top of this page, put it in your Claude environment, and from that moment Claude analyses your portfolio according to this logic. What a skill actually is and why this kind of file works, you can read in what Claude skills are NL.

The problem the skill solves is probably one you recognise straight away. Every product in your business gets roughly the same attention, because everything feels important. Marketing budget follows the product that shouts the loudest, not the product that is worth the most. There is an old service running somewhere that nobody dares to cancel. And the new idea you are excited about gets a few extra hours tacked on, instead of a real choice: go all in or stop. The BCG Matrix has existed since 1968 to solve exactly this, and this skill turns filling it in into a conversation of a few questions instead of an afternoon with a whiteboard.

It is meant for anyone selling more than one thing: business owners with multiple services, agencies with different propositions, teams managing a product line, companies with separate business units. The questions that trigger it are stated literally in the file: where should I invest, which products earn the most, should I keep this product. Anyone who asks Claude something like that while the skill is loaded gets a matrix, not a general answer.

One thing you should know upfront: the skill analyses, it does not research. It works exclusively with what you supply or estimate, and it says so itself. The quality of the classification stands or falls on the honesty of your estimates of market growth and market share. Rate yourself structurally too generously, and you get a flattered matrix, which buys you nothing.

Those two axes produce four positions, and each position has its own logic. Stars are leaders in a growing market: they demand a lot of investment but also deliver a lot, and if you feed them well they later become Cash Cows. Cash Cows dominate a mature market: plenty of cash flow, little investment needed, and that cash flow should flow towards your Stars and your best Question Marks. Question Marks sit in a growing market where you are still small: high risk, and the file is blunt about it: invest or divest, do not do it by halves. Dogs have no growth and no position: divest, or deliberately keep small in a niche.

Is a model from 1968 still usable? The file itself takes a position on that, through its source list: alongside Henderson it includes BCG Classics Revisited: The Growth Share Matrix from 2014, in which BCG held its own model up to the light again. The framework itself is not outdated, because the underlying question never changes: which part of your portfolio deserves your scarce money and attention. What does change is the pace at which quadrants shift, and that is exactly why it helps that the analysis no longer costs a consultancy project.

The difference between talking about your portfolio and actually filling in the matrix shows up fastest in the language. This is what the skill does with the sentences that come up in every strategy session:

Then comes the method in four steps. Step one is the input. For each product or business unit the skill asks for a name, an estimate of market growth in per cent a year or simply low, medium or high, an estimate of the relative market share, and optionally revenue and profit contribution. If you have no exact figures, it may work with estimates, but those get flagged as an estimate in the output.

Step two is the classification, with an explicit guideline: market growth above 10 per cent a year counts as high, and a relative market share above 1, so bigger than your biggest competitor, counts as high. Step three links each quadrant to an action: invest and protect for Stars, milk and pass through for Cash Cows, choose deliberately for Question Marks, divest or niche for Dogs.

Step four zooms out to the whole portfolio: if there is a shortage of Stars, the future runs dry; if you have too many Question Marks, you burn cash without choosing; if there are too many Dogs, capacity stays locked up in mediocre products.

The output format is fixed, and that is deliberate. For each item you get the quadrant, the market growth, the relative market share, an action in one sentence and a short reason. Then follows portfolio advice of two or three sentences about the balance and the priority for the next 12 months, and the answer closes with sharp follow-up questions, as in the example: which new product in a growing market could become your next Question Mark? That fixed format lets you place analyses from different moments side by side and see what has shifted.

The tone is prescribed too. It has to write like a seasoned strategy consultant: clear, not vague, no jargon without explanation, do not dodge decisions, and it is allowed to be blunt about Dogs. That last point is a detail that says a lot: the file knows the Dog diagnosis is the most painful one, and explicitly gives Claude permission to make it anyway. At the back are three sources and three things the skill will not do on principle, and those limits get their own section further on.

Why the model works lies in what it leaves out. Almost every portfolio discussion derails the same way: ten criteria get juggled at once, from customer satisfaction to how enjoyable the work is, and by the end of the afternoon everything is important. Henderson reduced the question to two axes that both come down to money. Market growth determines how much investment a position demands, relative market share determines how much margin a position delivers. All the romance falls outside the matrix, and that is exactly the point.

The second insight is the cycle from FIG.02: a portfolio is a system in which cash flows from mature positions fund the growth positions. Anyone with only Cash Cows earns well today and has a problem in five years. Anyone with only Stars and Question Marks grows fast until the money runs out. The balance questions in step four of the skill come straight out of this systems thinking, and the portfolio advice from the example shows how that plays out: one Star without a Cash Cow is not a luxury position but a risk.

What the matrix does not do is explain the environment. Why a market is growing or shrinking, which forces are putting pressure on your margin, what legislation or technology is going to do: other frameworks exist for that, and the file points to them itself. For the forces within your market there is the Porter Five Forces skill NL, for the world outside your market the PESTEL Analyst skill NL. And anyone who wants to decide which direction the growth should come from after the portfolio choice reaches for the Ansoff Matrix Strategist skill too: those two matrices are each other's logical sequel.

After that, using it is simple: list your products, with what you know per product about market growth and your position relative to your biggest competitor, and ask for a BCG matrix. If anything is missing, it asks for it itself. A more detailed guide for every environment sits in installing Claude skills NL, and anyone who wants to learn to write a SKILL.md themselves after using skills finds the complete recipe in writing a SKILL.md NL.

There are also situations where you are better off leaving it alone. If you have only one product and no plans for more, there is little to divide and a positioning question is more useful than a portfolio question. If you want backing with real market data, you have to gather that yourself first, because the skill deliberately fetches nothing and deliberately runs no calculations. And the most important limit sits in the input: the matrix is only as honest as your estimates. That is partly caught by flagging assumptions as an estimate, but no framework corrects a structurally too rosy self image.

The last honest limit: the matrix tells you which product to feed, milk or let go, but not how to do that in practice. The execution, from shifting marketing budget to sunsetting a product properly, stays human work. What the skill does do is make sure that execution is based on something: an explicit classification with reasons attached, instead of on the mood of the day.

A sharp portfolio choice decides where your marketing budget goes, but that budget still has to earn its return somewhere. For most businesses that starts with being found by the people looking for their Star. The free SEO scan shows in a few seconds where your site stands. In the guide which strategy model when NL you see how this model relates to the others. And if you want to talk through what AI can do for your business beyond this, from separate skills to full automation, we simply do that in a conversation.