The Firebugs in the Attic: Why Sustaining Innovation Never Delivers Business Model Innovation

This post argues that sustaining innovation — accelerators, innovation labs, venture clienting, e-fuels, hydrogen engines — is not dangerous because it fails, but because it works. Working results are the reason nobody has to ask whether the business model still holds. The argument runs through four cases: the record labels and DRM, thirty-five years of German automotive from the 1991 BMW E1 to today, the research literature on venture clienting, and Christensen’s own decision to follow The Innovator’s Dilemma with a book that promised a solution. It ends with five lessons and no method, because the thing being described is a dilemma, and a dilemma cannot be solved — only decided.

Why sustaining innovation is so hard to refuse: a lesson from Max Frisch

In The Firebugs — also translated as The Fire Raisers or The Arsonists — a 1958 parable play by the Swiss writer Max Frisch, a man named Biedermann reads in his newspaper about arsonists who talk their way into people’s houses, store petrol in the attic, and burn the place down. That same evening a stranger appears at his door. Biedermann lets him in. Then a second one. They move into the attic. They roll in barrels. Biedermann sees the barrels, asks what is inside, is told it is petrol, and decides not to make a scene. In the final act he hands them the matches himself, because refusing would have been impolite.

Biedermann is not stupid. Every single one of his decisions is socially reasonable. Throwing a guest out is unpleasant. Accusing someone of arson without proof is rude. Each step is defensible. Only the sum is fatal.

I have been writing about business model innovation for seventeen years on this blog, and I have come to think this is the most accurate picture we have of what actually happens inside large companies. Not blindness. Not laziness. A long series of individually sensible decisions, each one avoiding a conflict, adding up to a catastrophe.

The guest we keep letting in is sustaining innovation.

Why business model innovation hurts

A partner at a large consultancy once put it to me like this: business model innovation always hurts, because there are no easy answers — only the realisation that the past was magnificent and is not a plan for the future.

That is the shortest version of everything below.

The reason it hurts so much is that a business model is a company’s identity, not just its profit mechanics. Questioning it does not mean saying “your processes could be better.” It means saying “the thing you are proud of no longer carries you.” You can buy an instrument without changing. You cannot buy a business model.

So the evasion into instruments is not stupidity. It is identity protection. And that makes it far harder to argue with than incompetence would be.

The rational no: Christensen’s innovator’s dilemma in one scene

This is Christensen’s resource dependence, dramatised. In The Innovator’s Dilemma, the customer who rejects the disruptive solution is not dumb or slow. He is rational. Measured against his criteria, the new thing genuinely is worse: lower margin, smaller volume, immature performance. And the harder version of the argument says that even if management wanted to allocate resources differently, its own best customers and its own return expectations would prevent it.

That is worth sitting with before we get indignant about anyone. The people who say no are usually right about their own numbers.

“You don’t understand our market”: the killer argument against business model innovation

Notice what the town does not do. It does not test the water.

This is the move I have seen most often and recognised most slowly. The diagnosis is not refuted. The diagnostician is reclassified. “You don’t understand our industry.” “You don’t understand our market.” “That’s academic.” It is the most economical way to dispose of an inconvenient analysis, because it costs nothing — no counter-evidence, no data, no argument. It only requires assigning the speaker to a category from which nothing binding can be said.

I have been on the receiving end of that sentence more than once, and I want to be precise about why it works. It is not censorship. Nobody is silenced. The analysis stays in the room, it is simply no longer anybody’s business. And that is enough, because a diagnosis that nobody has to act on is indistinguishable from one that was never made.

E-fuels, hydrogen and HVO100: sustaining innovation you can buy

Here is what sells instead.

Tell a manufacturer of combustion engines that e-fuels are coming — synthetic petrol made from renewable electricity, poured into the tank of a car built yesterday — and you are not selling a technology. You are selling a sentence: you may remain who you are. The piston stays. The plant stays. The supplier network stays. The competence stays. The existing fleet keeps running. Only the molecule changes.

That is why e-fuels have the political and emotional pull they do. They are the one option that requires nobody to give anything up — not the manufacturer, not the supplier, not the driver, not the region that builds engines. Everything else on the table asks somebody for a sacrifice.

Hydrogen burned directly in the cylinder makes the same promise one step down, and HVO100 makes it one step further: a drop-in fuel for engines already on the road. I have written about why a clean fuel is not a business model, so I will not repeat the argument here — only the part that matters for this post. None of these are fantasies. They work. They can also come down a learning curve. But they sit far up that curve with very few cumulative doublings behind them, while batteries have had a decade of volume doing the work. Volume is the variable, and volume is exactly what these options do not have.

None of that is the reason they sell so well. They sell because they promise change while leaving the business model untouched.

Venture clienting and ambidexterity: sustaining innovation as a process

Now move one level up, from the product to the organisation, and you find the identical construction.

Venture clienting. Ambidextrous organisation. Innovation labs. Corporate accelerators. The promise is the same: the evaluation criteria stay, the budget logic stays, the career incentives stay. Only the origin of the idea changes.

Take venture clienting, the current favourite. Gregor Gimmy coined the term at BMW in 2014 and built the first venture client unit, the BMW Startup Garage, in 2015. The idea is genuinely elegant: don’t take equity in a startup, become its customer while it is still a venture. Buy the prototype, run it in a real project, turn the first order into follow-up orders. “We are not an accelerator, we are a venture client” is how Gimmy draws the line, and he is right to draw it — this is not a demo-day operation.

I want to be careful here, because the model is better than most of what surrounds it. But read the selection criterion its own inventor states. You should only consider venture clienting when your internal solution is significantly worse or does not exist at all — not even within your network of established partners. If a startup is only marginally better, stay away.

That is a sensible rule. It is also a precise description of what the model is built to do: fill gaps in the value creation you already have. It looks for technologies your current suppliers cannot deliver, for the architecture you are currently running. The reference point is always the existing business.

Look at the case everyone uses to explain it. BMW wanted cars to drive themselves out of the factory, solved it with sensors along the track rather than in the vehicle, and did it with two startups the Startup Garage found and integrated into the purchasing system. Excellent project. Real value. And a process improvement inside the plant.

Now hold that against Christensen’s actual argument, because this is the part that gets lost. In his framework the autonomous unit does not work because it is organised autonomously. It works because it has different customers — a new market that values the immature offering precisely for the attributes the mainstream market rejects. Autonomy without its own demand is just a sheltered room.

A venture client unit, by construction, has the same demand as its parent. Its internal buyers are the existing business units, applying the existing criteria, serving the existing customers. That is not a flaw in the execution. It is what the model is for.

And you do not have to take my word for it, because the research says the same thing in its own vocabulary. A comparative study of corporate venturing modes, built on 99 interviews with corporate venturing managers, finds that corporate venture capital tends to operate as an exploratory mode oriented toward future growth, while venture clienting serves as an exploitative mode focused on the near term. Exploit and explore are March’s terms, and exploit is the academic name for what I have been calling sustaining.

Note what that finding is not. It is not a verdict on whether the model works. It works. The pilots run, the sensor gets into the plant, the cycle time drops, and the numbers show up in somebody’s target agreement this year. Venture clienting is sustaining innovation at its best, and that is exactly what makes it dangerous.

A bad instrument gets found out. A good one gets loved. Every working result is an argument that the problem is being handled — and the better the results, the more convincing the argument. Firebugs who turn out to be useless get thrown out. The useful ones stay in the attic.

This is the point where most critiques of these instruments go wrong, mine included on earlier attempts. The question is never whether they deliver. It is what they deliver against. Sustaining innovation is effective in the short term, reliably and repeatedly, and it does nothing whatsoever to secure long-term survival — because it is aimed at a different axis. The plant runs better while the question of whether you will still need that plant goes unasked.

Why the research looks the way it does

The serious academic work on venture clienting deserves credit rather than suspicion. Tobias Gutmann at EBS Business School has published in Research Policy, the Journal of Business Venturing and the Strategic Entrepreneurship Journal, and co-authored what is now the standard book on the subject. He is candid about the state of the field: venture clienting, he notes, remains young and conceptually fragmented in academic terms, with few dedicated studies, so the theoretical anchors have to be borrowed from adjacent research streams — corporate venturing, open innovation, and supplier development from supply chain management.

Hold on to that third anchor. Supplier development is the discipline of making your existing supply base better. If that is one of the frames the phenomenon is understood through, then the field has already told us what kind of thing it is looking at. Not a mechanism for entering new markets. A mechanism for improving the value chain you are running.

Why does the literature look like this? I have written before about the paradox of management science: the discipline is best at explaining what no longer matters, because it rewards research that explains what has already happened. That is not laziness, it is the shape of the field. A process can be described while it runs. Whether a process carries a company through a structural break is knowable only afterwards — and by then it is not a research question but a post-mortem.

So the incentives line up neatly on every side, and nobody has to act in bad faith for the result to be what it is. Researchers describe what exists, because that is what can be researched. Consultants sell what can be implemented. And established companies love an instrument that lets them be innovative without giving anything up. Everyone is doing their job well. That is precisely the Biedermann problem: the fire does not require a villain.

The critique lands even harder on ambidexterity. It requires corporate leadership to allocate resources against its own return logic and to sustain that conflict indefinitely. That is not a structure. That is a feat of character. And a strategy that presupposes individual willpower as its mechanism is not a strategy.

The music industry ran this experiment for us

Nobody can claim the record labels slept through the digital transition. They fought Napster in court. They built their own platforms with real money and real people. In April 2001, RealNetworks, AOL Time Warner, Bertelsmann and EMI announced MusicNet, with RealNetworks holding forty per cent and each of the three label parents twenty. Universal and Sony answered with PressPlay. Enormous effort, front-page announcements, executives explaining that this was the online distribution solution they had been waiting for — that what they had needed all along was someone who would make music available “in an appropriate way” over the internet.

Appropriate is the word to hold on to. It meant: in a way that does not endanger the album.

MusicNet launched at $9.95 a month with roughly 75,000 tracks, and the music disappeared when your subscription lapsed. PressPlay let you keep tracks while you paid, but forbade CD burning and playback anywhere except the PC. Its basic tier allowed 500 low-quality streams, thirty downloads and ten burns per month. Two competing catalogues, neither complete, because each consortium licensed only its own rights.

DRM was not a regrettable compromise in these products. It was the showpiece — the engineering achievement you could present internally, and simultaneous proof that the existing business was safe. The feature that stopped customers from listening was, inside the building, the mark of quality.

And the question nobody worked on was the only one that mattered: how do I find the music that fits the mood I am in right now? That is the job. Not owning titles. Not even buying titles. Getting to the right thing at the right moment. The labels never understood that as their task, because they defined themselves by the catalogue rather than by access to it.

The rest is known. Apple did not have better technology. Apple asked a different question and sold the single track. The album had never been what the customer wanted — it was what could be shipped in a box. Eleven songs you didn’t want for the one you did. Once the physical carrier fell away, the justification for the bundle fell away with it.

Then streaming arrived and sold neither the album nor the track, but access and discovery.

Three steps: bundle, track, access. The industry that lived through all three initiated none of them. It reacted every time, with considerable effort every time, and every time it built the thing so that the existing business stayed protected.

Nobody in those buildings was asleep. They were being hospitable to the most useful guest they had.

German automotive: thirty-five years of knowing

The same clock is running now, and we can read it.

The Tesla Model S went on sale in June 2012. That is the moment the question stopped being whether electric drive works and became whether the incumbents can build this. Not a niche forecast — a series production car in the premium segment, which is precisely where German manufacturers earn their money.

They had known the technology far longer than that. VW ran electric Golf prototypes on public roads from the mid-seventies, BMW showed the fully functional E1 at the 1991 Frankfurt motor show, and between 1992 and 1996 prototypes from VW, BMW, Mercedes and Opel went into a government field trial on the island of Rügen, where residents used an electric car as their only vehicle. The verdict was that the technology was not ready. Measured against lead-acid batteries in 1994, that verdict was correct.

And this is exactly why staying with sustaining innovation is so defensible. A disruptive technology is genuinely bad at the start — that is Christensen’s whole point. It improves along its own trajectory while the incumbent technology improves along its, and for years the gap looks permanent. Then the lines cross, and the customers who were most demanding, the ones who had the best reasons to reject it, switch last and fastest. The Rügen trial measured the right thing at the wrong altitude. It asked whether the battery worked in 1994. The question that mattered was which curve it was sitting on.

But the car was the smaller half of it, and this is the part that was misread at the time and is still being misread. Tesla did not sell a vehicle. It sold the ability to actually use one. The Supercharger network went up alongside the product, because the bottleneck in electric mobility was never the drivetrain — it was the answer to where do I charge on the way to the mountains. Whoever solves that owns the value proposition. Whoever builds only the car has made a component for somebody else’s system.

And charging was only the most visible piece. Tesla built its own cell factories, because a car company that buys its batteries has handed the largest cost block and the steepest learning curve to a supplier. It sold directly, in shopping centres rather than dealerships, because a franchised dealer earns his living on servicing combustion engines and has no reason to talk a customer into a car that barely needs servicing. And those showrooms did something the industry had not needed for a century: they let people experience the thing before buying it. A car has always been an experience good, but the experience used to be familiar — everybody knew what driving felt like. An electric car in 2013 was unfamiliar and slightly frightening: the range, the charging, the silence, the question of what happens when it goes wrong. The showroom existed to take the fear out, not to close the sale.

Value proposition, value creation architecture, revenue model, and the people who carry it — all four moved at once. That is what a business model innovation looks like, and it is why the incumbents could not answer it by building a better electric car. There was nothing single to copy.

A manufacturer who sells cars and leaves charging to utilities, cells to suppliers and customers to franchised dealers has defined himself as a maker of vehicles. That was a perfectly good definition, and it was the right one for a hundred years, because refuelling, batteries and retail were solved problems that other people had solved well. The moment those problems came unsolved, the division of labour stopped being neutral and became a structural disadvantage. European manufacturers are still living with it today. The convenience of electric mobility is more than the car, and the customer buys the convenience, not the drivetrain.

What did the software resources go into over the following three years? Into defeat devices. That is the hardest version of this argument, and I want to state it carefully: it was not inaction. It was enormous effort, highly qualified engineers, sustained over years, spent defending the existing product. The combustion engine was to pass limits it could not pass. That is sustaining innovation in its purest and most expensive form — you don’t optimise the product, you optimise the measurement. It is also the point in the play where Biedermann stops asking what is in the barrels.

In 2017 I ran a workshop on the competences needed for the connected car. Five years after the Model S, two years after the scandal broke. The young managers in that room were not in denial. They were frustrated, because they could see precisely how little capability existed for what was obviously coming. It was named. It was documented. It was in the room. I wrote a frustrated post that same year about a company that had set up an innovation unit and asked whether it could possibly work that easily.

What came of it? By 2026 I am writing about SANY building an electric truck every six minutes and what that means for European OEMs.

Thirty-five years from the E1, fourteen from the Model S. And note what fails here — it is not diagnosis. The diagnosis existed in 2017, made by the people it concerned. Naming the problem in a workshop is not the same as naming it in the resource allocation. What was missing was never the analysis. It was the decision to give something up.

Innovation theatre: why sustaining innovation sells so well

Steve Blank gave this its name in 2019: innovation theatre. Hackathons, design thinking classes, innovation workshops — activities that shape culture, he grants, but rarely deliver anything shippable. Visible, photographable, and leaving the allocation of money and people exactly where it was.

His diagnostic test is one question, and it is brutal in its simplicity: ask what happens to the output of the successful teams. How does it reach the sales channel? If the answer is that they are working on it, you have your answer.

The chronology is long: corporate venture capital in the late nineties, accelerators, incubators, intrapreneurship, innovation labs and digital units around 2015, design thinking, lean startup in the corporation, company builders, now venture clienting and AI labs. The striking feature is the half-life of roughly five to seven years — long enough that the launch, the operation and the quiet dismantling never appear in the same annual report. The successor fashion starts before anyone has to measure the effect of its predecessor.

One distinction matters here, and it cuts against the easy reading of Blank. Some of these formats were theatre in his strict sense: activity with no output reaching the business. Others, venture clienting among them, are not theatre at all. They work. And the ones that work are the harder case, because Blank’s test cannot catch them. Ask what happened to the output and you get a real answer, with numbers. The question that would catch them is different: not did anything reach the business, but did anything change about which business we are in.

Biedermann’s guests were not idle either. They carried things upstairs, they were pleasant at dinner, they made themselves useful. That was the whole problem.

Rumelt’s test separates all of this in one sentence. Every one of these formats is additive. They cost budget. They demand no sacrifice. Who does venture clienting hurt? Nobody. That is why everyone agrees to it, and that is why nothing happens.

I should be honest that I have been watching this loop rather than breaking it. In 2009 I wrote that I was getting bored of the mantra that design thinking would solve large corporations’ problems — impressive client list, unimpressive output. In 2017, the innovation unit. In 2026, venture clienting. That is not three observations about three fashions. It is one observation, confirmed three times against whatever instrument happened to be current. I will come back at the end to what that makes me.

Three questions for anyone selling self-disruption

So let me put the burden of proof where it belongs. Nobody needs to prove that these instruments produce results — they do, and I have just said so. The claim that needs proving is the larger one: that they answer Christensen. Anyone making it should answer three questions:

Name me the self-disruption — an established large corporation that attacked its own profitable core business.

Name me the structure that carried it — the ambidextrous organisation or venture client process that was demonstrably the mechanism.

Name me the new customers — the less demanding buyer segment that was opened up in the process.

The list stays short. The counter-list is long, and it is devastating in a specific way. Kodak’s engineer Steve Sasson built the first portable digital camera in the company’s own lab in 1975, presented it to senior management, and was told it was cute but that he should not tell anyone — Kodak patented it and left it there. GM built a working electric car in the EV1, leased just over a thousand of them, ended production in 1999 and later took the cars back and crushed them. Nokia had an internet-ready touchscreen phone prototype in 2004, three years before the iPhone, and shelved it along with a proposal for an online app store. Xerox PARC is the canonical case: maximum autonomy, maximum inventiveness, almost no adoption by the parent.

In fairness, none of these are as simple as the folklore suggests, and the Kodak story in particular gets told too neatly — Sasson’s prototype weighed nearly four kilos, produced 0.01-megapixel images and needed twenty-three seconds to write one to tape. Management was not wrong that it was unsellable in 1975. That is exactly Christensen’s point, and exactly why the counter-list matters: at the moment of decision, the rational reading of the numbers pointed the other way every single time.

Every one of these organisations had the invention in the building. Several had exactly the separated, autonomous structure the ambidexterity literature recommends — Xerox PARC is the purest example ever built. They had it, and they failed anyway. Which is fairly strong evidence that the structure is not the mechanism.

Look at the cases that did work and the pattern is different. Netflix had no explore portfolio running alongside the DVD business; Hastings converted the core business head-on, against customer protest and a collapsing share price. Microsoft had no autonomous cloud unit that won an internal argument; Nadella changed the target metric for the entire company and rewired compensation to match. Apple did not spin out the iPhone.

The mechanism was the same everywhere: a leadership decision involving sacrifice, in the core, against resistance. Structure is a symptom of the will, not its cause.

I want to be clear about the status of that claim, because it is easy to mistake for a verdict. It is not one. I would like to be wrong about this, and I am asking to be. Those three questions are an invitation, not a trap. If you know a large incumbent that attacked its own profitable core, carried by an ambidextrous structure or a venture client process, and opened a genuinely new customer segment doing it — write to me. I will publish the case and revise the argument.

I would also be inconsistent to claim otherwise. In complex situations nobody has finished answers, and that includes me. What I can say is what I have been able to find and what I have not. So far the counter-list is long and the list is short. That is an observation about the evidence available, not a law of nature.

Christensen against Christensen

Here is the part I find hardest to argue with.

The Innovator’s Dilemma, Clayton Christensen’s 1997 book, is an impossibility statement. It says the incumbent cannot, for reasons that are rational and structural. A book like that is difficult to sell, because it leaves the reader holding nothing.

So The Innovator’s Solution followed in 2003.

I am not making an accusation. Christensen was a serious scholar and the second book contains real work. I am pointing at something else: the same market force described in the first book acted on the author of the first book.

And look closely at what the shift actually offers. The first book says: here is a dilemma. A dilemma is not a problem. A problem has a solution you can find; a dilemma has two options that both cost you something, and it cannot be solved — it can only be decided. Under uncertainty, with sacrifice, and with nobody to confirm afterwards that you chose right.

The second framing says: yes, there is a dilemma, and here is the solution. That sells incomparably better, and not because buyers are foolish. It sells better because a procedure is the only thing anyone can put in a box. You cannot deliver a decision. You can only make one.

That is the whole chain in one sentence. E-fuels, hydrogen in the cylinder, HVO100, the accelerator, the venture client unit — every one of them offers a procedure at the precise spot where a decision belongs. Which is exactly why they are welcome.

If that is true — and I think the publishing record makes it hard to deny — then the expectation that a board can escape the same force by purchasing a process is not optimistic. It is naive.

But notice the conditional, because everything turns on it. The dilemma does not always win. It wins whenever it is not spoken aloud. And every offer described in this post shares one property: it permits action without diagnosis. E-fuels, hydrogen in the cylinder, HVO100, the accelerator, the venture client unit — each one lets you do something visible, expensive and genuinely useful while leaving the question untouched.

That last part is what makes them effective as anaesthetic. A useless instrument would provoke the question sooner. A working one postpones it, and keeps postponing it for as long as the results hold.

The labels never spoke their dilemma. They turned it into a feature and called it DRM.

The fair version: when sustaining instruments are the right choice

These are good ideas, as long as they stay inside the logic of the existing business model. That is not an excuse; it is a condition you can take seriously. An accelerator that qualifies suppliers is legitimate. A venture client unit that gets a better sensor into the plant three years earlier than procurement would have is legitimate and genuinely valuable. E-fuels in aviation, where no battery will do the job, are legitimate. So is HVO100 in the machines no electric drivetrain will fit.

As sustaining instruments.

The deception is rarely the inventor’s doing — the venture client model states its own scope honestly, right down to the rule about when not to use it. The deception happens in the market afterwards, in what the label comes to promise. These formats get sold as the answer to Christensen when by construction they serve horizon one and two. The honest version reads: “venture clienting accelerates our sustaining innovation and closes gaps our suppliers cannot close.” That is a real benefit, worth real money. Nobody buys it, because it addresses no fear.

The real cost of delay: why sustaining innovation makes it worse

The damage is not that sustaining innovation solves nothing. The damage is that it consumes the time in which a solution was still possible.

Every year of innovation theatre is a year in which the gap widens and the options narrow. What you face at the end is not the same problem you had at the beginning — it is a larger one, and the means available to address it are smaller. The German automotive industry did not lose fourteen years standing still. It spent them, and the cost of the answer rose while its own capacity to pay for it fell. What you could have built in 2012 you now have to buy, from suppliers who are no longer suppliers.

And in the meantime the energy goes into windmill fights. The labels against Napster. The publishers against Google. The manufacturers against regulation. Always committed, always expensive, always aimed at something that is not the actual problem. Don Quixote does not mistake the windmill for a giant out of laziness. He does it out of conviction, with courage, at full tilt. That is the tragic version of the error, not the stupid one — and it is the version that actually occurs.

Frisch understood the mechanism precisely. The house does not burn because Biedermann is careless with fire. It burns because every individual act of hospitality was reasonable, and because by the time the barrels were visible in the attic, asking about them would have meant admitting he had been wrong to let them in.

Lessons for business model innovation

If you want to carry something away, carry these five.

1. Define yourself by the promise, not by the product

Netflix could give up the DVD because its reference point was the value proposition, not the thing in the envelope. Get that right and you can abandon almost anything without losing yourself. Get it wrong — define yourself through what you make — and every alternative arrives as an attack, which is exactly how the record labels experienced a customer who wanted one song, and exactly why charging networks, cell plants and showrooms all looked like somebody else’s business to a company that had decided it was in the business of making cars. The product is only one building block in fulfilling a value proposition, and what business you say you are in determines which futures remain available to you.

2. Structures do not disrupt anybody. Decisions do.

Every case that worked ran on a leadership decision with a cost attached, taken in the core, against resistance. Every case that failed had the structure and lacked the decision. If your transformation plan can be drawn as an org chart, you have not yet made the decision — you have made an arrangement for postponing it.

3. Judge an instrument by its label, not by its results

The results will be fine. That is not the test. The test is what the instrument is being sold as. As sustaining innovation, venture clienting and clean fuels are honest, useful and worth the money. As an answer to disruption, they are a sedative. Ask any provider one question: which part of our current business does this ask us to give up? If the answer is none, you have learned what you are buying.

4. The customer who says no is right about his own numbers

Do not lecture him. His criteria are real and his arithmetic is sound — that is Christensen’s finding, not a character flaw. The move is not to argue with the existing customer but to find or construct a buyer whose numbers point somewhere else. Demand is a position, not an attitude.

5. A dilemma is not a problem, and it stays a dilemma

None of the above amounts to a method, and I am not going to pretend otherwise. What changes is whether you treat the thing as something to be decided — with a cost you accept in advance — or keep buying procedures that let you postpone the decision while looking busy.

---

Frisch gave his play a subtitle: Ein Lehrstück ohne Lehre — a morality play without a moral.

That is the most accurate description I know of thirty years of innovation management literature. Everything has been described. Christensen laid out the mechanism in 1997. The counter-list has been public for decades. And still, every five to seven years, a new instrument arrives that promises transformation without sacrifice — and it is bought, and it works well enough that nobody has to ask the other question, and by the time anyone does, the answer has become more expensive.

There is one more part of the play worth knowing. Frisch put a chorus of firemen on stage — a deliberate parody of the Greek chorus, standing there in full uniform, commenting on the action in elevated verse, warning Biedermann, seeing the barrels in the attic and knowing exactly what they are. They never put out a single fire. They are not incompetent and they are not lying. Foreknowledge without resolve simply achieves nothing, and they are not the ones who decide what happens in that house.

I have been part of that chorus for seventeen years on this blog. So has the literature, so has the research, so has every consultant who has ever presented a disruption slide to a board. We describe the fire very well. We do not put it out, because the matches are not ours to hold.

The arsonists are not hiding. They never were. They are helpful, they are polite, they lower your costs, and they are sitting in the attic right now. They will still be there tomorrow, doing useful work, delivering results you can put in a report.

Biedermann’s tragedy is not that he failed to see them. It is that he handed them the matches to avoid an awkward conversation.

The only question that has ever mattered is whether you will have that conversation — and say out loud what your most useful guests actually are.

---

Dr. oec. Patrick Stähler · fluidminds · Zurich

Understand. Imagine bigger. Act.