The McKinsey 7S Analyst skill for Claude

The problem the skill solves is one anyone who has ever watched a reorganisation up close will recognise. One thing gets changed, usually the thing you can draw on paper: a new org chart, a new strategy at the kick-off, a new system. Six months later nothing has changed except the confusion. The reason is stated in the file itself: change programmes often fail because managers adjust one element without changing the other elements along with it. That creates misalignment, and the model gives it a name. The new structure does not fit the existing systems, or the leadership style clashes with the new strategy.

It is written for entrepreneurs, managers and advisers who need to get to the bottom of an organisation rather than describe it. The frontmatter names three moments: you want to assess organisational alignment, you want to plan a change programme, or you want to structure the integration after a merger or acquisition. It also contains trigger phrases that are closer to everyday reality. The file itself is written in Dutch, so those phrases are Dutch too: "we groeien maar het loopt vast" (we're growing but it's getting stuck) and "cultuur past niet bij strategie" (culture does not fit strategy). Say something like that to Claude while the skill is installed, and it picks it up automatically. It belongs to the skill library NL that we make freely available from our AI and automation service, with no account and no sales email afterwards.

What the skill is not: a judgement on people. Roles, staffing, style and skills are properties of the organisation here, not an assessment of individuals. And it is not a strategy generator. You do not get an answer to the question of whether serving enterprise customers is a good idea, you get an answer to the question of what is not right in your organisation the moment you take that idea seriously.

The seven elements are divided into two groups. The hard elements are tangible and directly influenceable: Strategy, the plan by which the organisation achieves and keeps its competitive advantage; Structure, the organisational structure with hierarchy, reporting lines and management; and Systems, the processes, workflows and systems that carry the day-to-day operation. These are the three a management team adjusts fastest, because they can be set down in a document.

The soft elements are intangible and culture-driven: Shared Values, the core with mission, vision, culture and shared beliefs, and that sits central in the model; Skills, the unique competencies and abilities within the organisation; Style, the leadership style and the management culture; and Staff, the people, roles, talent management and staffing. These four cannot be rewritten in a quarter, and that is exactly where the average reorganisation falls apart.

Why it works, the file puts like this: 7S forces holistic thinking and makes hidden bottlenecks explicitly visible. Hidden is the key word here. Nobody in the example company would say there is a problem with the leadership style. What people say is that big quotes keep sitting there unactioned. Only once you hold style up against strategy does it become clear why: someone who has to wait for the director on every decision cannot sit at the table with a customer who wants to make a decision. The difference between an isolated observation and a diagnosis looks like this in diff form.

Where the VRIO framework NL looks inward to determine which resource is a defensible advantage, 7S looks inward to determine whether the organisation is even capable of cashing in that advantage. They complement each other: VRIO tells you what you have, 7S tells you whether you are built for it.

Then come the triggers, listed literally in the file, and because the file itself is written in Dutch, so are most of them: analyseer onze organisatie (analyse our organisation), 7S framework, mckinsey 7s, organisatie alignment (organisational alignment), strategy structure systems, we groeien maar het loopt vast (we're growing but it's getting stuck), cultuur past niet bij strategie (culture does not fit strategy), and post-merger integratie (post-merger integration). The last three are interesting, because those are not model names but complaints. So it also starts if somebody does not know the framework at all and only describes the problem.

The main text consists of four parts. First theory and background, with the origin of the model and the core principle about alignment. Then the seven elements, divided into three hard and four soft, each with a one-line description. Then why it works, with the explanation about change programmes falling apart over one isolated intervention. And finally the output structure, and that is the engine of the skill.

That output structure is a contract of five parts, introduced with the words that Claude always delivers them. One: the 7S Scan, the current state per element, one paragraph per S. Two: the Alignment Matrix, stating which elements are and are not aligned with each other. Three: the Top 3 Misalignments, the three biggest bottlenecks with an explanation. Four: the Action plan, ranked steps to restore alignment. Five: quick wins, two to three immediately actionable steps. After that comes the example prompt with the situation you see in FIG.01, and the source list with two titles.

What is not in it is at least as important. There is no separate chapter with refusals, and there is no numbered step plan. The boundary of this skill sits entirely in that fixed output contract: the numbers are fixed, one paragraph per element, exactly three bottlenecks, two to three quick wins. It also states no turnaround time, no promise about results and no claim about how often the model works. If you want to learn yourself how to build an instruction file like this, the full explanation is in writing a SKILL.md NL.

The model became widely known through the 1982 book In Search of Excellence, which according to the file sold more than three million copies. That is a fact about the book, not about your result, and that is exactly how we mean it. It does explain why 7S still turns up in almost every advisory project: it is one of the few models that stays understandable both at a board table and in a team meeting.

Waterman, Peters and Phillips, 1980, Structure Is Not Organization. The article in Business Horizons in which the model was introduced. The message: anyone who wants to change an organisation by changing its structure changes the least important part the fastest.

Peters and Waterman, 1982, In Search of Excellence. The book that brought the model to a broad audience, with company examples that filled in the idea of alignment between hard and soft elements.

What has been added in the meantime is not in the skill, and you are entitled to know that. The model is descriptive, not predictive: it helps you see where the friction sits, it calculates nothing and it gives no scores. If you want to bring in the outside as well as the internal diagnosis, other models are more obvious. The PESTEL Analyst NL maps the macro environment, the Porter Five Forces Analyst NL the forces in your industry, and the SWOT analysis skill is the broad first stocktake that 7S then goes deeper into internally.

Even without installing the skill, you can use the model: write your organisation out under seven headings, and then draw lines with a pen between the headings that contradict each other. What is left is your agenda for the coming quarter.

After that, using it is simple: describe your organisation under the seven headings, the way the example prompt in the file does. If you cannot fill in an element well, write down what you do know; a half-filled element still produces a usable tension with the rest. Because the file falls under the MIT licence, you are also allowed to adapt it: add your own examples from your sector, extend the trigger words, or extend the output structure with a sixth part you need. If you are new to skills, start with what are Claude skills NL and the broader explanation of working with AI in the knowledge base.

There are also situations where you are better off leaving it be. For questions about the market itself, 7S is the wrong instrument, because it looks exclusively inward: trends, regulation and macro developments are the domain of the PESTEL Analyst NL, and competitive pressure that of the Porter Five Forces Analyst NL. If you want to know which of your strengths is really defensible, the VRIO Analyst NL is the more logical choice. And if the diagnosis is already made and it is now about execution, a planning or change approach belongs there: OGSM or OKR for translating it into goals, and the Kotter Change Coach for the programme around it.

The most important limit is the quality of your own description. It only reads what you write down, and describing an organisation the way you would like to see it produces an analysis of a company that does not exist. So write down what actually happens on Monday morning, even if that sounds less tidy. The temptation is especially strong with Style and Shared Values: almost nobody writes down that the leadership style is directive, while the example prompt in the skill does exactly that. Ask a second person to fill in the same seven headings and compare the two versions; the difference between the two is often already a misalignment in itself.

And finally, realise that a diagnosis is not a change. A model points out where the friction sits; people do the fixing. Anyone who confuses the two ends up with a nice diagram and an unchanged company.

An organisation that is internally aligned only performs once enough work is also coming in to perform on. In the matrix above that often stays invisible, while findability is just as much a system that is or is not aligned with your strategy: anyone who wants to serve enterprise customers but is only found for small jobs has a misalignment that no org chart solves. The free SEO scan shows in a few seconds where your site stands. And if you want to talk further about what else AI can do for your business, from stand-alone skills to full automation, we simply do that in a conversation.