Your Business Model Runs on Assumptions You Stopped Seeing

Every business model rests on assumptions about the world it operates in. The trouble starts when those business model assumptions stop being recognisable as assumptions and quietly turn into certainties. This post begins with Drucker’s account of that mechanism, then shows two sets of assumptions failing under measurement in a single year: the geopolitical order that supplied us with energy, and the climate that shaped our buildings, water and winters. It sets out an inventory of more than forty assumptions Swiss business and society still run on, reads the same list a second time as a list of openings rather than losses, and ends with a method for surfacing your own. It offers no solutions. It offers a way to find them, and names the gap that the next post will take up: we have no established method for generating the models that replace the broken ones.

Why business model assumptions turn invisible

Assumptions are not a failure of rigour. No organisation can act without them. Every strategy presupposes a view of how the environment develops, what customers want solved, which capabilities stay valuable, and how value is created and captured.

They turn dangerous at one specific moment: when they stop being recognisable as assumptions. Then hypotheses harden into certainties, former conditions of success become unspoken rules, and contrary evidence reads as noise.

Peter Drucker described this in The Theory of the Business in 1994. Every organisation rests on assumptions about its environment, its mission and its core competencies. A valid theory requires them to fit reality, to fit one another, to be known throughout the organisation, and to be tested constantly. It is not graven on tablets of stone, he wrote. It is a hypothesis about things in constant flux. His sharpest observation explains our present situation: a successful organisation takes its theory for granted and substitutes culture for discipline.

There is a social half to this that Drucker leaves out. A business model is also a social construct: a dominant group inside the firm defines what business you are in, and that definition quietly fixes which options are thinkable at all.

A certainty is an assumption that was right long enough to stop being visible.

Which makes the work ahead smaller than it sounds. Three states, not three words. An assumption is stated and checked. A certainty is an assumption that succeeded its way into invisibility. An explicit assumption is one that has been retrieved, dated and made checkable again. These sentences were spoken out loud once. Somebody stated them, tested them against the world, and acted. They did not start out tacit. They became tacit. The task is retrieval, not invention.

One assumption sits above all the others

Everything that follows rests on a condition: that an assumption can be checked against reality at all. In Earth for All, its 2022 report, the Club of Rome names the loss of a shared basis of fact as a central obstacle to collective action – misinformation drives polarisation and distrust, and agreeing on basic facts stops being possible.

That is the one failure that makes the others unrepairable, and it has a smaller cousin inside every firm: the shared mental model that decides which evidence counts as evidence. A broken assumption you can measure is a design task. A broken assumption you can no longer settle is a fight.

This post assumes the distinction still holds. That assumption is itself under pressure.

Here are some of ours. Power comes out of the socket. Because we are rich, we will always be supplied. Copper will always be there to buy. Money is cheap. We build houses for winter. There is snow in the mountains. Europe is at peace. America is a reliable partner. China is a sales market and a place to produce. Our precision will always be needed. A degree protects you from automation. Education happens in schools. Our pensions are secure. Our customers come to us. We know our competitors. We act on facts.

Which of these are true? Better question: how much longer?

For two blocks of them, the answer is already in. Both arrived in the same year.

Two geopolitical assumptions broke before anything else

Europe ran on two of them. That fuel is a price problem rather than an availability problem. And that China is a place to sell and to produce, not a rival at the top of the value chain. Both broke, and they broke differently.

Availability: war decides who gets supplied

US and Israeli operations against Iran began on 28 February 2026. Iran responded by asserting control over the Strait of Hormuz and attacking shipping that ignored its routing demands, as the Congressional Research Service documents. A ceasefire in April and a memorandum on 17 June did not hold. The conflict resumed in July.

The strait carried about a quarter of the world’s maritime oil trade in 2025. Days after the attacks began, Iranian forces declared it closed. Cross-strait traffic largely halted, and hundreds of vessels with thousands of mariners were left trapped inside the Persian Gulf.

The IEA’s executive director called it the largest oil supply shock in history. In March and April 2026 the agency organised the largest coordinated emergency stock release ever undertaken: 400 million barrels from 32 member countries, announced on 11 March. Global oil prices rose 50 per cent between February and May.

Meanwhile the hybrid campaign in Europe has become routine: sabotage, cable damage, arson, cyberattacks and airspace violations, increasing in number and severity. In August 2026 an explosives-laden drone was found beside a Ukraine-bound cargo aircraft at Leipzig/Halle.

Money buys you a place in the queue, not a delivery. Purchasing power never guaranteed physical availability. It described a political order, and that order is gone.

China: from sales market to technology leader

The second assumption broke without a shot being fired.

The IEA’s Global EV Outlook 2026 reports that China made more than 80 per cent of the world’s battery cells in 2025, and an even higher share of the materials that go into them. Chinese carmakers supplied 60 per cent of all electric cars sold worldwide. Outside China, Europe and the United States, 55 per cent of electric cars sold were imported from China, against less than 5 per cent five years earlier.

This is not cheap labour arriving in Europe. It is the high-value end of the chain: cells, chemistry, power electronics, software, and increasingly trucks. The lead was not won in the car industry at all: the battery collected its learning curve in laptops and phones while Europe was still debating technology neutrality.

The customer became the competitor, and the competitor now sits upstream of you. Whoever assumed they would keep the clever part and outsource the rest is discovering that the clever part was the cell.

The climate scenarios stopped being a forecast in 2026

The next block failed in the same year. Here the evidence is not a market but a measurement series.

The hottest summer since 1864

June to August 2026 was the hottest Swiss summer since records began in 1864, beating the 2003 benchmark. Many low-lying stations counted roughly two weeks more heat days than in their previous record summer. The ten days from 19 to 28 June were the hottest ten-day stretch at several stations in 125 to 160 years of measurement.

The driest period since 1901

April to August 2026 was the driest and hottest such period since 1901, at 61 per cent of normal precipitation nationwide. Close to 40 per cent of Swiss gauging stations set new summer low-water records and almost 60 per cent of water-temperature stations set new highs. Three of the large lakes fell below their previous seasonal lows, and the Limmat regulation rules had to be set aside to secure minimum residual flow. The Rhine at Basel tracked a modelled extreme scenario built for 2025 to 2040.

The consequences were operational, not atmospheric: a nationwide fire ban from the end of July, alpine pastures supplied by helicopter, restricted Rhine shipping. All of it is documented in the MeteoSwiss Klimabulletin Sommer 2026, prepared with the federal environment office and the forest research institute.

An ordinary summer of the 3-degree world

The number that matters for anyone planning a building or an infrastructure is at the end of that bulletin, and it comes straight out of the national climate scenarios Klima CH2025, published on 4 November 2025 by MeteoSwiss, ETH Zurich and the C2SM. Those scenarios describe Switzerland’s climate at defined global warming levels: the 1.5-degree world, the 2-degree world, the 3-degree world, each measured against 1871–1900. Under current global policy we are heading for the third of them. In that world, the average Swiss summer would run about 3.6 °C above the 1991–2020 reference. Summer 2026 came in at 3.4 °C.

In temperature terms this was not an outlier. It was an ordinary summer of the world we are heading into, and in that world half of all summers would be hotter still. On rainfall, 2026 was worse than that future average: about 280 mm nationwide, against roughly 350 mm projected for the average future summer and 420 mm in the recent normal.

One more figure, because it kills a specific assumption on its own. From June to August the rural station at Zürich/Affoltern recorded not a single tropical night. The city station at Zürich/Kaserne recorded thirty-eight.

Nine months separated the publication of those scenarios from the summer that matched them. They changed no building code in advance. That is not a climate story. That is a story about what organisations do with a tested hypothesis they would rather not act on.

Two broken assumptions, one price at the pump

The neat way to read the last two sections is as two separate problems. The pump price says otherwise, because it contains both at once.

In mid-September 2026 German petrol and diesel hit all-time records, and the ADAC pointed out that crude does not explain them: in late April, Brent cost about the same and petrol was thirty cents cheaper.

The difference is a river. The Rhine gauge at Kaub fell to single digits in August 2026, the first such reading since records began and far below the previous low of 2018. Freight rates for tanker barges from Rotterdam to Karlsruhe more than tripled within two months.

Switzerland pays the same bill. Fuel is up by roughly a quarter since January, heating oil by about a third since June, and the TCS names the three drivers plainly: Rotterdam product prices, the dollar, and Rhine freight rates.

The price on a Swiss forecourt now contains a strait and a river level. One is politics, the other is physics, and they cancel the same assumption: that energy is a market where price is the only variable and somebody else handles the rest.

And the cheap price was never the true one

It is tempting to read the carbon price as the political part of the bill and the rest as the market. That has it backwards.

Fossil fuels have never carried their full cost. The German Environment Agency recommends valuing greenhouse gases emitted in 2026 at 350 euros per tonne of CO2 at a one per cent time preference rate, and at 1,000 euros per tonne if the welfare of present and future generations is weighted equally. Prevailing carbon prices sit roughly an order of magnitude below that.

The political intervention is therefore not the levy. It is the absence of one. The damage is real; it is simply invoiced to other people in other years, and the summer of 2026 was one of the instalments.

This is also why a cleaner fuel is not an answer. A drop-in substitute keeps the subscription intact; it only changes what you are subscribing to.

For a business model this is a timing question rather than a moral one. If your economics work because a cost sits outside your accounts, you are carrying an unpriced liability. Every step towards recognising it looks like a political shock from the inside and like a correction from the outside.

An inventory of broken business model assumptions

Switzerland as the test case, deliberately. Small country, first place in the Global Innovation Index for years, an international reputation for stability and precision. If the everyday certainties are expiring here, they are expiring everywhere. Substitute your own country as you read.

Where hard sources exist they are linked. Everywhere else the third column states an observable development and should be read as exactly that.

Fourteen fields, forty-six assumptions: geopolitics and dependency, energy and supply, climate, water and buildings, tourism and the Alps, industry and world trade, cost and efficiency, innovation and industrial capability, work and knowledge, education and science, medicine, society and pensions, money and property, the business model itself – and the one assumption underneath all of them.

Assumption	Implied logic	New reality
Geopolitics and dependency
Europe is at peace.	Military force is not a planning case for European business or infrastructure.	It is one. Drones close airports, Baltic cables get cut, explosive packages travel in air freight. Your buildings, grids and logistics are the target surface.
America is a stable democracy and a reliable partner.	Alliances, trade rules and security guarantees change only at the margin.	They now change by the election cycle. Any plan that needs a reliable America past the next vote is a plan with an unpriced option in it.
Security is provided; we can concentrate on business.	Sea lanes and trade rules are maintained by somebody else and are simply there.	Hormuz shows what that assumption is worth: traffic down 95 per cent, and no European state in a position to reopen it.
Weapons systems are bought for decades.	Threats change more slowly than procurement cycles.	Drones and electronic warfare iterate in months. A twenty-year procurement now buys obsolescence on a schedule.
Energy, where geopolitics and climate cancel the same assumption
Energy is a market question; we buy where it is cheapest.	Supply is commercial, and price is the only variable to manage.	You never saw the real price. Fossil fuels have always sold below cost, with the damage billed to other people and later years. The forecourt price is the subsidised one.
Because we are rich, we will always be supplied.	Ability to pay guarantees physical availability.	The largest oil supply disruption on record happened anyway, and money did not reopen the strait. Wealth buys you a place in the queue, not a delivery.
Power comes out of the socket.	Electricity is always there at the capacity needed.	Demand is rising with electrification while generation turns weather-dependent, and the grid itself is now a target for sabotage.
Sixty per cent of our electricity is hydropower.	A high historical share guarantees stable domestic supply.	In 2026 close to 40 per cent of gauging stations set summer low-water records and the Rhine ran at an extreme-scenario level. Water in the average year does not help you in the dry one.
Thermal plants run and provide baseload.	Fuel, cooling and economics are secure.	Fuel costs are political, carbon costs are rising, and cooling water was short in the hottest summer on record. Baseload has quietly become residual capacity.
Buildings designed for a climate that no longer exists
We build houses for winter.	Cold is what the people inside have to be protected from.	A house completed in 2026 will spend most of its life in a climate closer to the 3-degree world than to the one it was designed for. From May into September it is the heat that decides whether the people inside sleep, work and stay healthy.
Thick walls keep a house cool in summer.	Thermal mass is enough protection against overheating.	Mass only works if the heat can leave at night. In the summer of 2026 it often could not, and the walls then store the problem instead of solving it.
Night-time cooling makes air conditioning unnecessary.	The air outside gets cold enough each night to flush the building.	In the city it no longer does. And the official stations will not tell you: Fluntern sits on a wooded slope, Affoltern on the open edge of town. The series you base your design on was not recorded at your address.
We know our climate, we have long measurement series.	The record describes the conditions our buildings and infrastructure actually face.	The network was sited for meteorology, not for habitation. Between the reference station and the flat under the roof lie several degrees that nobody measured.
Air conditioning is the only way to cool a building.	Cooling is a machine you install once the heat becomes unbearable.	Most of the work happens before the machine: shading, load reduction, night purging, surface and district cooling, and free cooling from boreholes or groundwater – cooling without a compressor, which is where the electricity goes. The compressor is the last resort, not the starting point, and the one that adds load to a grid already peaking in the heat.
We are the water tower of Europe. (Swiss self-image as the reservoir feeding four major European rivers.)	Water is permanently abundant and available on demand.	In August 2026 three of the large lakes fell below their previous seasonal lows and the Limmat regulation rules had to be set aside. The tower ran short in its own country.
Alpine water shortages are a transport problem.	Scarcity is temporary; herds and grazing patterns need not change.	Helicopters flew water up and animals came down early. When the exception repeats every few years it is not an emergency, it is the operating model.
Tourism and the Alps
We have pleasant summers, so tourists come.	The climate is a stable locational advantage, so hotels never needed to invest in cooling.	The 2026 summer product included a nationwide fire ban and local water restrictions. And guests from India have complained that a Swiss hotel room is hotter than what they left at home. A climate that spared you the investment stops being an advantage the moment it stops sparing you.
We have snow in the mountains.	Natural snow stays available at sufficient altitude and duration.	The snow line is rising and the season is shortening. Artificial snow needs cold, water, energy and money, and the first of those is the one you cannot buy.
Industry and world trade
China is a sales market and a place to produce.	We produce cheaply there, grow there, and keep control of our knowledge.	China made over 80 per cent of the world’s battery cells in 2025 and built nearly three-quarters of the world’s electric cars. The customer has become the competitor, and in trucks too.
Raw materials like copper will always be available.	Efficiency means saving energy. Material is a purchasing matter, and a market always delivers.	We have it backwards. The sun delivers more energy to this planet in an hour than humanity uses in a year, and it arrives again tomorrow. Copper, lithium and rare earths do not, and they sit in few hands. We economise on the unlimited resource and squander the limited one. Whoever keeps the material has it; whoever sells it has to import it again.
Our supply chains are secure.	Thin inventories and a few efficient suppliers hold up under stress.	A closed strait and a dry river were enough. You cannot supply a chemical plant or a steelworks by lorry, and the rail corridor was under renovation.
Cost leadership as a broken assumption
Low cost decides whether a company survives.	The lowest-cost producer outlasts everyone.	Being cheapest makes you interchangeable. What decides now is whether the customer would miss you.
Cost cutting will get us to the future.	The future comes from spending less on the present.	Cuts improve this year’s result by removing next year’s capability. No amount of saving answers the question of whose problem you intend to solve.
Lean works, and our supply chains are optimised.	A system tuned for minimal inventory stays capable under disruption.	It is tuned for the normal case, and the normal case has stopped showing up. Every buffer you removed was an option you sold.
Innovation and industrial capability
We lead in precision parts, and precision will always be needed.	Precision is a capability, and capabilities do not expire, so demand is secure.	Much of that precision ends up in combustion engines. An electric drivetrain has no injection system, no valves, no camshaft. Demand does not move to a cheaper supplier – the part stops existing. And most suppliers cannot see it, because they sell to a system integrator and book the work as a precision part, not as a powertrain.
Being excellent at our part is enough.	Whoever makes the best component captures a fair share of the value.	Sensors, software and control shift where the customer’s value sits. You can be the best component in somebody else’s business model and still be the one who is interchangeable.
Switzerland is innovative because of all the ETH spin-offs. (ETH Zurich, the federal technical university.)	Many foundings mean much market-effective innovation.	Founding is not scaling. A country that starts a thousand companies and grows none of them has a statistic, not an industry.
We are highly innovative.	Strong rankings guarantee future competitiveness.	Switzerland still tops the Global Innovation Index, for the fifteenth year running. In the same edition China entered the top ten for the first time. A ranking measures inputs and outputs, not whether your own model survives the decade.
We have a great many patents.	A high count proves innovative strength.	The count says nothing; the direction says everything. In logistics, the incumbent was patenting the business it already had while the challenger was patenting the one it intended to build. Read your portfolio as a map of where your innovation budget actually went.
Work and knowledge
A university degree guarantees a job.	Automation replaces manual and unskilled work. Whoever has studied is protected by qualification.	It is the other way round this time. AI is taking the analysing, drafting and summarising that graduates were hired to do, while the work that needs hands in a room is the hardest to automate.
Junior roles are how we train tomorrow’s senior people.	There will always be entry-level work to learn on.	The tasks juniors learned on are the first to be automated. Cut the bottom rung this year and in ten years there is nobody to promote.
Education and science
Education happens in schools and universities.	Knowledge is scarce, so it has to be taught in a room, at a time, by an institution that then certifies it.	Learning has left the building. An AI tutor is available at any hour, at any level, in any language, and it does not have thirty other pupils. What stays inside the walls is judgement, research, and learning alongside other people – none of which is the lecture.
Scientific findings are accepted by society.	Scientific quality generates its own authority.	The climate scenarios were public years before the records confirmed them, and they moved almost nothing. Being right is not the same as being acted on.
Why prevention has no revenue model
Medical progress means cutting-edge medicine.	Progress comes from better diagnostics, more demanding surgery and more capable devices, for people who are already ill.	For the big chronic diseases the technology to see them coming years in advance already exists. What is missing is a revenue model that pays when nobody falls ill. As long as we bill per case, prevented illness is lost turnover, and the system quietly prefers treatment. The same mismatch shows up in the obesity market: the model sells episodes while the customer’s job is permanent.
Society, pensions and skills
Our pensions are secure. (AHV, the pay-as-you-go state pension.)	Contributions, working lives and political consensus carry the promise indefinitely.	The 2025 to 2055 scenarios show fewer contributors per recipient and rising care costs. The promise was made under a demography that has ended.
We fix the skills shortage by bringing people in.	Somewhere there is a lasting surplus of suitable people.	Everyone is ageing and everyone is recruiting. You are not hiring from a pool, you are bidding in an auction.
Money and property
Capital is cheap.	Money costs almost nothing, so any project that returns something at all is worth financing.	An entire generation of managers has never calculated at other rates. When money costs something again, projects that only ever worked because capital was free stop working, and nobody in the room has seen it happen before.
Property is a safe asset.	Land and buildings hold their value almost regardless of what stands on them.	Location, energy condition and insurability now separate assets sharply. A building that cannot be cooled, cannot be heated affordably or cannot be insured is not the same asset as the one next to it.
What can be insured stays insurable.	Risk is priceable, so somebody will always write the policy.	Insurers are withdrawing from individual regions and perils as losses become predictable rather than random. An uninsurable asset is not a cheaper asset; it is a different business model, and often none at all.
Growth is the normal case.	Plans, careers and valuations assume a market that gets bigger.	Hardly any European model is built for a shrinking home market. Growth has to come from a larger share, a better offer or a different customer, and all three are design questions rather than forecasts.
The business model itself
Our customers come to us.	Demand finds its own way, and the relationship is direct.	An AI agent now sits between you and the customer and decides who gets shown. Being good is no longer the same as being found – and reach you rent is not a defensible position.
We sell to people.	On the other side of the transaction is a human being with habits, loyalty and a reluctance to switch.	Increasingly the buyer is an agent acting for the customer. It has no habits and no loyalty, it re-compares every time, and every soft bond your model relied on counts for nothing.
We know our competitors.	The threat looks like us and sells what we sell.	The one that takes your business comes from an industry you do not monitor and earns its money somewhere else entirely.
Scale protects us.	Volume lowers unit cost and raises the barrier to entry.	Scale protects you from someone carrying the same fixed costs. It protects you from nobody who found a way not to carry them.
Our brand carries us.	Decades of reputation keep customers choosing us.	Brands work while customers do the choosing. Once an agent draws up the shortlist, your reputation has to be legible to a machine that reads specifications, not stories. A brand is icing on a business model, and icing does not survive on its own.
And the one underneath all of them
We act on facts.	Once the analysis is right, the decision follows.	Every assumption in this table broke in public, with the evidence available in advance. Knowing was never the bottleneck. Deciding was.

Read the middle column from top to bottom and one sentence keeps reappearing in different costumes: somebody else takes care of that, and it will stay that way. That is the assumption behind all the assumptions. It is the one that is no longer occupied.

Read your broken assumptions as openings, not losses

Nothing above is new information. What follows is the same list with a different question put to it.

Not: what are we losing. But: what has just become permitted that was impossible before.

A broken assumption is an opening, not a loss

Take the cooling row. As long as everyone knew that cooling means installing a machine, nobody spent forty years looking for anything else. The certainty was not a fact. It was a search ban. Now somebody is searching, and shading, night purging, free cooling and district cooling turn out to be a business waiting for whoever builds it first. The same applies to every row in that table.

Every dead assumption is a place where something new has become possible. Forty broken assumptions are not a crisis report. They are the largest opening in decades. Whoever surfaces them first gets to design. Everybody else adapts.

Rubble or building ground is a choice of question

This is not optimism as a mood. In 1948 anyone who asked how bad things were in Germany was entirely right, and built nothing. The difference between rubble and building ground is not the finding. It is the question you put to it.

The migrants who crossed the Atlantic make the sharper case, because for them no reality had collapsed at all. They left behind more than property. They left habits: guild rules, inherited rank, the settled order of who may become what. Nobody refuted those rules. They were simply not packed. When Alexis de Tocqueville travelled the United States in 1831 and wrote Democracy in America, he was describing a society running on assumptions that had been left behind in a European harbour.

Our position is better than either of theirs. They rebuilt on ruins or started from nothing. We would be starting from a country that works.

The frontier had a price. We do not have to pay it

There is an honest objection to all of that, and it belongs in the text rather than in a footnote.

Unreflected departure produces victims. What was a new beginning for the settlers was dispossession for the people already there. Post-war reconstruction was also a long exercise in not looking at what had just happened. Departure without honesty about its cost is not optimism. It is recklessness with better branding.

What is different now is that we are rich. We can afford to take people with us, and that is not sentimentality but a condition of feasibility: change against your own workforce is slower than change with it, and the people doing the work often see the shift more clearly than the management that grew up with the old model. A company is not a machine with an input and an output.

The change is coming with us or without us. The only negotiable variable is the number of losers. Refusing to surface your assumptions is not a decision against change. It is a decision not to shape it, and it moves the bill to the people least able to pay it.

Honesty without optimism turns into cynicism.

Optimism without honesty turns into denial.

Action without either turns into activity for its own sake.

We could do this voluntarily

Not because we are forced, but because we can imagine something better and are capable of building it: something radically better for the people we serve, if we want it. Nothing here is destroyed. Only the assumptions are. We have the opening without the rubble, and we are hesitating anyway.

The alternative is not standing still. The alternative is doing it later, under pressure, on someone else’s terms.

How to surface your business model assumptions

Five steps. They work in a workshop and they work alone on a train.

The hard part is not the new idea. Unlearning is, which is why the method starts with retrieval rather than invention.

1. Complete the sentence. Our model works as long as … Do not ask what you want. Ask what has to be true. Three to five conditions is usually enough, and writing them down is most of the work.

2. Sort them by the four elements. Value proposition, value architecture, revenue model, company spirit. Not as a checklist, but because a broken assumption in the revenue model demands something entirely different from a broken one in the value proposition. A business model is a system of interdependencies, not a list of blocks, which is where the concept came from in the first place.

3. Interrogate each one. Why did we believe it? What evidence supports it today? What counter-signals are we already explaining away? What happens to us if it is false?

4. Invert it. Not a forecast, a way of opening the search space. Timekeeping is free: what else could a watch stand for? The night no longer cools the building: what does a house look like that does not need it to? Customers do not come to us: what do we build then?

5. Run one small real test. An idea is not a reinvention. A model exists only when customer value, capabilities, revenue logic and organisation fit together. Pick one critical assumption, define an observable signal of its failure, formulate a plausible counter-assumption, and try it on real customers.

Drucker prescribed the same discipline in a harder form: every three years, challenge every product, every service, every policy, every channel with one question – if we were not in this already, would we go into it now?

What this post does not do: generative strategy

This text is a diagnosis. It tells you how to find the assumptions your model rests on and how to see the openings where they have failed. It does not tell you how to generate what goes in their place.

That gap is older than the current crisis. In his 2010 Long Range Planning article, David Teece argued that a technology has no value in itself – value appears only through the business model that commercialises it – while noting that the design of business models has no sound theory behind it and remains closer to art than to science. Fifteen years on, that gap is still open. What arrived in the meantime were tools and pattern catalogues: description and transfer, not generation.

Analytical strategy describes and optimises what is. Generative strategy produces what could be. The second is the subject of the next post.

Lessons learned

1. Every business model is a bet on assumptions. The dangerous ones are not the obviously wrong ones. They are the ones that were right for a long time.
2. A certainty is an assumption that succeeded its way into invisibility. Success, not stupidity, is what makes a premise tacit.
3. Retrieval, not invention. Every one of these sentences was explicit once. You are not inventing a worldview; you are putting an old statement back into words so it can be checked. I published the same argument in 2009 and again in 2026, which tells you something about how long a retrievable assumption can stay buried.
4. Climate change has stopped being a projection. It is a measurement, everywhere. The models are no longer telling us what may happen; the record is telling us what did. A scenario that has arrived is not a warning any more, it is a schedule – so ask which of today’s scenarios you are still filing rather than acting on.
5. When two independent realities cancel the same premise, the argument about either is over. Geopolitics and climate landed on the same assumptions from different directions, in the same year.
6. Fossil energy is a subscription; built capacity is a purchase. Any dependency that must be renewed every month is a lever somebody else holds. That is a business model question long before it is an environmental one.
7. The flood of broken assumptions is the opportunity, not the diagnosis. Look at what broke at once: where our energy comes from and who controls it, who leads in the technologies of the next decade, what a building has to do for the people inside it, which raw materials we can count on, what a degree still certifies, who decides what a customer sees, and whether we can still agree on a fact. Any one of those would be a decade of work. Together they are not a longer crisis – they are a larger drawing board, and almost nothing on it is fixed.
8. Efficiency tells you how well you are doing something, never whether it is still the right thing. That is the trap that catches the most competent organisations, not the weak ones.
9. Change arrives either way. What you decide by acting early is not whether it happens, but who carries the cost.

So: which assumption has to stay true for your model to keep working, and what will you do when it stops?

Understand. Imagine bigger. Act.

---

Sources and further reading

• Peter F. Drucker, The Theory of the Business, Harvard Business Review, September–October 1994. Every organisation rests on assumptions about environment, mission and core competencies; the theory is a hypothesis that must be tested constantly, and successful organisations are the ones most likely to stop testing.
• Congressional Research Service, The Strait of Hormuz: Security Developments and Impacts on Oil, Gas, and Other Commodities, R45281, updated 7 August 2026. The primary chronology of the 2026 conflict, the collapse of transits and the share of world energy trade that passes through the strait.
• Congressional Research Service, Russian Hybrid Warfare Activities in Europe: Considerations for Congress, R49134, 11 August 2026. Documents a sustained campaign of sabotage, assassinations, political interference, cyberattacks and airspace violations, increasing in number and severity since 2022.
• International Energy Agency, Global EV Outlook 2026, May 2026. China’s share of battery cell production and key battery materials, the share of global electric car sales held by Chinese manufacturers, and the export figures cited above.
• MeteoSwiss, Klimabulletin Sommer 2026, Zurich, 10 September 2026, with BAFU and WSL. The record summer, the driest April to August since 1901, record low water and water temperatures, and the comparison with the average summer of a 3 °C world.
• MeteoSwiss, ETH Zurich and C2SM, Klima CH2025: Swiss Climate Scenarios, published 4 November 2025 under the National Centre for Climate Services, with contributions from the University of Bern (Oeschger Centre) and the University of Lausanne. Successor to CH2018; its innovation is to describe the Swiss climate at global warming levels of 1.5, 2 and 3 °C above 1871–1900 rather than by emission pathway. The scientific report is available in English. This is the source of the “future summer” comparison used above.
• OECD, Global value and supply chains. Treats resilience not as a return to domestic production but as the interplay of diversification, agility and adaptability.
• World Intellectual Property Organization, Global Innovation Index 2025, 16 September 2025. Switzerland ranks first among 139 economies for the fifteenth consecutive year; China enters the top ten for the first time. Some 80 indicators across innovation inputs and outputs, which is broader than a patent count.
• Swiss Federal Statistical Office, Szenarien zur Bevölkerungsentwicklung der Schweiz und der Kantone 2025–2055, Neuchâtel, 15 April 2025. Ageing advances fastest between 2025 and 2035; from 2035 all population growth is migration-driven because deaths exceed births.
• ADAC, Aktueller Spritpreis, daily national averages for German fuel prices, with the observation that the 2026 records are not explained by the crude price alone.
• German Environment Agency (Umweltbundesamt), Gesellschaftliche Kosten von Umweltbelastungen, based on the Handbuch Umweltkosten – Methodenkonvention 4.0 and the GIVE model. Recommends 350 euros per tonne of CO2 for emissions in 2026 at a 1 per cent time preference rate, and 1,000 euros at 0 per cent – far above prevailing carbon prices.
• Sandrine Dixson-Declève, Owen Gaffney, Jayati Ghosh, Jørgen Randers, Johan Rockström and Per Espen Stoknes, Earth for All: A Survival Guide for Humanity, a report to the Club of Rome, New Society Publishers, 2022. Fifty years after The Limits to Growth; argues that polarisation and the erosion of a shared factual basis are themselves a barrier to acting on any of the other findings.
• Alexis de Tocqueville, Democracy in America, 1835 and 1840. The classic outside account of a society built by people who did not bring the old European order with them.