Broken Assumptions: How Good Strategy Rebuilds a Business Model
When an assumption behind a business model breaks, most companies respond with bad strategy. This post argues that broken assumptions are the best possible starting point for good strategy, provided they are treated as a diagnosis. It applies Richard Rumelt’s kernel of good strategy – diagnosis, guiding policy, coherent actions – to the four elements of the business model, which bring tacit assumptions back to the table. It continues the earlier post on the assumptions we stopped seeing. There is no case study here, only a method, illustrated with short examples from Intel and Ørsted. It ends with six questions for your next strategy meeting.
Why broken assumptions produce bad strategy first
In my last post, Your Business Model Runs on Assumptions You Stopped Seeing, I followed Peter Drucker’s The Theory of the Business. Every organization runs on assumptions about its environment, its mission and its core competencies. Over time, these assumptions stop being hypotheses and become certainties. Nobody checks them anymore.
That post promised generative strategy next, and it will follow. One step comes first: the moment an assumption breaks. Now what?
The first reaction is rarely strategy. In Good Strategy Bad Strategy (2011), Richard Rumelt names four hallmarks of bad strategy. After an assumption breaks, all four appear with remarkable reliability.
Fluff. Rumelt’s definition is hard to improve: “Fluff is a form of gibberish masquerading as strategic concepts or arguments.” It usually comes first. The board presentation speaks of resilience and transformation.
Failure to face the challenge. If you cannot define the challenge, you can neither evaluate a strategy nor improve it. After a break, the challenge is reframed as a temporary disturbance: prices will normalize, customers will return, the regulator will come to its senses.
Mistaking goals for strategy. Most of these goals are statements of desire. They contain no plan for overcoming an obstacle. “Back to 2021 margins by 2028” describes a destination and says nothing about the road.
Bad strategic objectives. Objectives are bad when they miss the critical issues or cannot be achieved. After a break, the list grows until every division has found its own project in it, and the broken assumption appears nowhere on it.
A fifth reaction deserves its own name, because it looks like competence. It is the efficiency trap: the company optimizes the business model whose foundation has just disappeared. Cost programs, lean initiatives, procurement savings. They all deliver. None of them asks whether the model still carries. This is excellence pointed at the wrong target.
I have used Rumelt on this blog before, as a test to expose bad strategy – in Why Bad Strategy is a Social Contagion and in the HVO100 post. Here I use his kernel as a procedure. The reason is simple. For Rumelt, a strategy is a response to a challenge, a way through a difficulty. A broken assumption is a challenge in exactly this sense. It is where good strategy begins.
Step 1: Find the broken assumption in your business model
Rumelt’s kernel starts with a diagnosis. A good diagnosis explains the nature of the challenge and simplifies an overwhelming reality by naming the aspects that matter most. Applied to a business model, the diagnostic question becomes precise: which element carried the assumption that broke?
The four-element model from Das Richtige gründen gives the map: the value proposition (the benefit you promise a real customer), the value-creation architecture (offer, production, distribution, core capability, partners), the revenue model (costs and revenues) and the spirit of the company (team and values).
The map does more than locate a broken assumption. It brings tacit assumptions back to the table. Rumelt tells you to diagnose; he does not tell you where to look. The business model does, because every element is a bundle of assumptions. The previous post sets out a five-step method for surfacing them, starting with one sentence: our model works as long as … Making this tacit logic explicit, so that it can be examined and redesigned, was the purpose of the business model concept in my 2001 dissertation.
The value proposition is where the most common tacit assumption hides. The dissertation already separated the value proposition from the product, and I have argued on this blog that the offer is only one building block to fulfill the value proposition. The product is the means; the benefit is what the customer buys. This separation changes who you think your competitors are. Your real competitors are defined by the customer’s job to be done, not by your offer. A company that defines competition through its own product carries an assumption it never states: that the customer’s alternatives look like what it sells. When that assumption breaks, the new competitor arrives from outside the industry, and a diagnosis that searches among look-alike products will not find it.
In the previous post I used a three-column grid: Assumption, Implied logic, New reality. For the diagnosis, it needs a fourth column: Element that carried it. Then comes the part that most diagnoses skip. A business model is a system of interdependencies, so a break in one element travels into the others. Take cheap energy. The assumption sits in the revenue model, on the cost side. But over decades it has shaped the architecture: where the plants stand, which partners supply them, which processes run in them. A diagnosis that stops at the cost line prescribes a cost program. Follow the break through the system, and the architecture itself comes into question.
The hardest assumptions to find sit in the fourth element. In the early 1980s, Japanese manufacturers turned memory chips into a commodity and broke the assumption that Intel could win in memory. As the Harvard historian Richard Tedlow recounts in Fortune, Intel’s profits fell from $198 million in 1984 to less than $2 million in 1985. The revenue model had signalled the break for a long time. What held the diagnosis back was identity. Intel saw itself as the memory company, and its top executives could not believe they were being outcompeted in a market they had created.
Andy Grove describes in Only the Paranoid Survive (1996) how he and Gordon Moore had been agonizing over the dilemma for weeks when he asked a question: if the board threw them out and brought in a new CEO, what would that person do? Moore’s answer came without hesitation: “He would get us out of memories.” The question works because it removes identity from the diagnosis. An outsider carries no history, so the outsider sees what the numbers already say.
The broken assumption that hurts most is usually the one that defines who you are.
Step 2: Every guiding policy is a new assumption
The second part of Rumelt’s kernel is the guiding policy: the overall approach for dealing with the obstacles the diagnosis has identified. One observation about guiding policies is rarely made explicit. Every guiding policy rests on a new assumption about the world. The only question is whether you choose it consciously and write it down.
The Danish energy company Ørsted illustrates this better than any other example I know. According to its own account, 85 percent of its heat and power production was based on fossil fuels in 2008. The company, then still called DONG Energy, formulated its guiding policy as a single ratio: 85/15, reversed. As Martin Neubert, then CEO of Ørsted’s offshore wind business, explained in a McKinsey interview, the goal was to turn a generation mix of 85 percent conventional and 15 percent renewable into the opposite. Inside that policy sat an assumption: offshore wind would become the profitable core of the energy business.
The assumption held. The company planned the reversal over decades and, by its own account, achieved it in about ten years. By 2019 it had become the world’s largest producer of offshore wind energy.
Then the new assumption broke as well. In August 2025, Ørsted announced a rights issue of DKK 60 billion, backed by the Danish state as majority shareholder. Citing a material adverse development in the US offshore wind market, it had cancelled the planned partial sale of its Sunrise Wind project off New York. On the day of the announcement, the shares fell by almost 30 percent.
This does not make 85/15 a bad strategy. It was one of the clearest guiding policies any incumbent in Europe has formulated. It shows something else: every guiding policy carries an expiry date that nobody writes on it. Macroeconomics, supply chains and US regulation, all three named by the company itself, were part of the new ground on which Ørsted built, and that ground moved.
Write down the assumption inside your guiding policy, together with the signal that would tell you it has broken.
Step 3: Coherent actions across all four elements
The third part of the kernel is coherent action. Rumelt insists that the actions of a strategy must be coordinated and must reinforce each other. For a business model, this has a concrete consequence. Because the four elements are interdependent, broken assumptions in one element require actions in all four.
The coherence test is simple to state and uncomfortable to apply. Take each proposed action and ask two questions. Which element does it change? Does it strengthen what the other actions do? A cost program in the revenue model, paired with an unchanged value proposition and an unchanged team, fails the test. It keeps the old model alive at a lower cost.
Ørsted’s transformation passed this test at the time. The company completed the divestment of its oil and gas production business in 2017 and with it dropped the reference to oil and gas from its name, changing from DONG to Ørsted. The name change was an action on the fourth element. It told employees, customers and investors which company they were now dealing with.
The fourth element is where coherence most often fails. The people who were made powerful by the old assumption are the last to accept that it has broken. The employees below them often see the break earlier and understand it more deeply, precisely because the past did not make them powerful. They need to be allowed to say it. Psychological safety is the precondition for an honest diagnosis and for actions that the organization will actually carry out.
If none of your actions touches the team and its values, you have a cost program, not a strategy.
Where Rumelt’s kernel reaches its limit
The kernel presupposes that you can diagnose the challenge. In a complicated situation, you can: cause and effect are knowable in advance, and experts can find them. In a complex situation, in the sense of Snowden and Boone’s Cynefin framework, cause and effect only become visible in retrospect. When several assumptions break at the same time – the geopolitics of energy, the climate, the economics of AI – any diagnosis will be partial.
Rumelt himself addressed this in The Crux (2022). His advice is to concentrate on the crux of the challenge: the part that can be overcome and promises the greatest progress. That helps. It still leaves one question open. Once the old assumptions are gone, someone has to imagine which new business model could stand on the new ground. That is no longer diagnosis. That is generating. It is the subject of the next post.
Six questions to ask when assumptions break
- What must be true for each of the four elements to work? Walk through value proposition, architecture, revenue model and team, and write the answers down. Then mark the assumption that has broken, in one sentence.
- Who else gets your customer’s job done? Define your competitors through the job, then check which of them your assumptions have ignored.
- Where has the break already traveled? Follow the interdependencies from the element that carried the assumption into the other three.
- What would a new CEO do? Grove’s question removes identity from the diagnosis. Ask it out loud, and let someone without a stake in the old model answer first.
- Which new assumption does your guiding policy rest on, and what signal would show that it is breaking? Write both down before you act.
- Which of your actions changes the team and its values? If the answer is none, go back to question 1.
Broken assumptions are painful. They are also the moment when a company can see its own business model more clearly than at any time since its founding. Most waste that moment on fluff and cost programs. Use it.
Understand. Imagine bigger. Act.
Annotated bibliography
Drucker, P. F. (1994). The Theory of the Business. Harvard Business Review, 72(5), September-October 1994, 95-104. – Every organization runs on assumptions about its environment, its mission and its core competencies; crises arise when these assumptions no longer fit reality, even though the organization keeps doing things well.
Grove, A. S. (1996). Only the Paranoid Survive: How to Exploit the Crisis Points That Challenge Every Company and Career. New York: Currency Doubleday. – Intel’s exit from memory chips as the model of a “strategic inflection point”; source of the question what a new CEO would do.
Rumelt, R. P. (2011a). Good Strategy Bad Strategy: The Difference and Why It Matters. New York: Crown Business. – Good strategy consists of a kernel of diagnosis, guiding policy and coherent actions; bad strategy shows four hallmarks: fluff, failure to face the challenge, mistaking goals for strategy, and bad strategic objectives.
Rumelt, R. P. (2011b). The Perils of Bad Strategy. McKinsey Quarterly, June 2011. – Condensed version of the book, published shortly before it; defines strategy as a response to a challenge and a way through a difficulty.
Rumelt, R. P. (2022). The Crux: How Leaders Become Strategists. New York: PublicAffairs. – Strategy starts with challenges rather than goals; the strategist’s skill lies in finding the crux, the part of a challenge that is both addressable and decisive.
Snowden, D. J., & Boone, M. E. (2007). A Leader’s Framework for Decision Making. Harvard Business Review, 85(11), November 2007, 68-76. – The Cynefin framework distinguishes complicated from complex contexts; in complex contexts, cause and effect can only be seen in retrospect, and expert analysis loses its power.
Stähler, P. (2021). Das Richtige gründen: Werkzeugkasten für Unternehmer. 5th updated edition. Hamburg: Murmann Publishers. – The four-element model of the business model: value proposition, value-creation architecture, revenue model and the spirit of the company.
Ørsted / Neubert, M. (2020). Ørsted’s renewable-energy transformation. Interview with Martin Neubert, CEO of Ørsted’s offshore wind business, McKinsey & Company, 10 July 2020. – First-hand account of the 85/15 vision and how DONG Energy turned a fossil generation mix into the world’s largest offshore wind business.
