Your company is dying. Stop trying to save it.

A note on scope: this is about large, established enterprises: capital-intensive, with real product portfolios, real governance, real dependencies. Not startups. Not pivot stories. Not another "innovate or die" sermon. If you've ever sat in a strategy review where someone said "growth" more often than "customer," this article is calibrated to your audience.

Every quarter, your leadership team looks at a board deck. Revenue is roughly on plan. The strategy slide is intact. Somebody says "growth" 19 times. The product roadmap is full. The transformation programme is broadly green.

And somewhere in the same company, something has already started to die.

It won't be on the roadmap. It won't be in the risk register. And it usually won't show up in the board pack until the dying business stops being profitable, which, in large enterprises, tends to happen years after the actual death began. By the time you admit it, the cheap window to act is gone. What remains is the expensive window.

So the most useful question your leadership team can ask this year is not "how do we grow?" It's:

Which part of us is currently dying and why hasn't anyone written it down?

What's really happening

The data is not subtle. Innosight's corporate longevity forecast shows the average tenure of a company on the S&P 500 has dropped from 33 years (1964) to 24 years (2016), and is forecast to shrink to 12 years by 2027. Their warning is basically the corporate version of "winter is coming," except nobody brings a coat to the meeting.

Foster and Kaplan called the underlying phenomenon "creative destruction" two decades ago and argued that companies "built to last" are also built to underperform the market over time. This is not new news. It is, however, news your strategy department tends to rediscover every few years as if it just happened.

The conventional reading is "innovate or die." where as the more useful one is harder:

The numbers aren't telling you to avoid death. They're telling you that survival is controlled amputation of products, business models, governance, complexity, and assumptions, preferably before the market turns it into people.

The companies that endure aren't immortal. They're simply the ones that are willing to kill parts of themselves in a controlled way and keep operating between the deaths.

Take LEGO. In 2004, LEGO was near the edge. It was losing nearly US$1 million a day and was carrying around US$800 million in debt. The cause wasn't a lack of innovation. It was too much innovation in too many direction and complexity that spiraled beyond control. The reset under Jørgen Vig Knudstorp was brutally simple: cut complexity, divest what wasn't core, stop adding optionality like it's free.

Now here's the part people love to say next: "Yeah, but we're not LEGO."

Good. You're worse.

LEGO could cut complexity without negotiating with 14 functions, three regulators, five steering committees, and last year's operating model. LEGO didn't have a global compliance regime, a multi-year union agreement, and an ERP landscape that treats basic reporting as a heroic act.

If they almost killed themselves with complexity, imagine what "innovation" looks like in an enterprise that can't even agree on a single definition of "customer."

Companies don't die in one dimension. They die in several, often at the same time:

• Death of the product (Nokia's phone; BlackBerry's keyboard; dedicated GPS units)

• Death of the business model (newspaper classifieds; rental video)

• Death of distribution (pre-Amazon retail; pre-streaming media)

• Death of relevance (Yahoo)

• Death of capability (the acquired startup six months after acquisition; the regulated company trying to ship like a product company)

Most enterprises survive one of these while being quietly killed by another. The product still sells; the business model has stopped scaling. Distribution still functions; the category is fading. Culture scores are "strong"; the capabilities required for the next decade aren't present in the building.

Which is why "Are we innovating?" is a useless executive question. The honest version is: Which death is currently underway and which one are we pretending not to see?

Why this keeps happening

1) The dying business is usually still profitable. When market share leaks slowly but margins still look great, every dashboard reads "fine." Revenue is a trailing indicator. Momentum is generous. By the time finance agrees something is wrong, the trend has usually been running for a long time.

2) Path dependency and sunk cost aren't personal, they're institutional. Factories, supplier contracts, headcount, and the political weight of the executive who runs the dying division are very real. The cost of reallocating resources is immediate and visible. The cost of not reallocating them is abstract and later. Boards reliably choose the immediate option because boards are built to do exactly that.

3) Risk governance is optimized for the current business. Stage gates, approvals, compliance reviews, control frameworks; they were built to protect the existing operating model. New work has to clear governance designed to prevent old work from breaking. Rational at small scale. Lethal when an entire category is dying.

4) Culture defends yesterday's winning formula. The behaviours and hiring patterns that produced 20 years of growth become the same things that prevent the next 20.

Which brings us to a painful benchmark: BCG's research on large-scale tech program implementations (basically the backbone of many transformations) found only 30% fully meet expectations for timeline, budget, and scope. [5] That success rate hasn't exactly become inspirational - despite every consulting firm on earth (including ours) trying to improve it.

One reason is simple: organisations keep treating transformation as something with a start and end date. But any company that wants to be around in 15 years doesn't "do transformations.". It operates in transformation mode.

Think about this for a moment

You have a product line contributing ~40% of EBIT. It's the dashboard hero. Every monthly review features its margin. It has also been losing category share for years.

Nobody has labelled it. Because the dashboard reports revenue and margin, not direction. The decline won't become undeniable until volume finally cracks — and by then you'll be reinventing under crisis, with fewer options and more fear.

This isn't unusual. It's the normal path of a dying-but-profitable business unit in a large enterprise. The dashboards are the last to know. The question is whether your organisation is structured to see that trend early - while you can still afford to act.

Most aren't.

The capability that actually matters: abstraction

Most large enterprises are now running "AI literacy" programmes. They teach people how to use ChatGPT, Copilot, Claude. They buy licences. They publish policies. They celebrate adoption numbers.

In three years, they'll do it again for the next tools. And the next. And the next. That's teaching the tool, not the skill. The skill is abstraction.

Abstraction is the ability to separate what you sell from what you're actually good at.

That's what Fujifilm did when the film business collapsed. They didn't "pivot from film to cosmetics" in the cute LinkedIn sense. They asked what they genuinely knew how to do: precision chemistry, coatings, materials, managing degradation under light and oxidation. Then they applied those capabilities to new domains: healthcare, imaging, displays, pharmaceuticals, and more. Result: Fujifilm is now a diversified company with annual revenue measured in the trillions of yen. Kodak, starting with broadly similar category assets, defended the legacy and filed for bankruptcy in 2012.

This is the muscle you want, because it compounds across multiple "deaths." Tool literacy doesn't. People who can abstract teach themselves the next tool. They don't need a programme. They don't wait for permission. They recognize that the tool is surface-level and their capability is portable.

And yes, there's a framework for this that's been around for years: Teece's dynamic capabilities - sensing, seizing, transforming - is essentially a structured description of abstraction in action. It is also routinely ignored in favour of transformation programmes with end dates and PowerPoint sunsets.

What to do about it

Not a pivot manifesto. A practical operating approach for organisations that must keep functioning while they evolve.

1) Name the death. Every quarter, ask: Which of our products, business models, distribution channels, markets, or capabilities is in decline that the dashboard doesn't show yet? Make someone accountable for the answer. Make it routine. Make it boring.

2) Teach abstraction, not tools. The next AI literacy programme will be obsolete faster than the governance approval for it. Teach people to separate what they do from what they know how to do and apply that capability to a different domain as an explicit exercise. Repeat until it becomes reflex.

3) Adapt while you can still afford it. The cheap window to evolve is while the existing business is still profitable. The expensive window is when it isn't. Most enterprises wait for the expensive window and then act surprised by the bill.

4) Govern new work differently than old work. Putting new businesses through gates designed to protect old businesses is how enterprises kill their own optionality. This doesn't mean "no governance." It means proportional governance. Heavy gates for mission-critical work. Lightweight guardrails for experimental work.

5) Plan for the next death. Whatever you build this cycle starts dying the moment it works. The only question is whether you'll see it earlier next time, with cheaper interventions and less drama. Treat the current adaptation as the first of several, not the last.

Closing

Your company is dying. The question was never whether. It was which death, in what dimension, on what timeline.

The companies that endure aren't the ones that figured out how not to die. They're the ones that learned to die in pieces, on purpose - on a schedule of their choosing rather than the market's.

Most enterprises won't do this. They'll keep running tool-literacy programmes, avoid uncomfortable conversations, and treat the next transformation as an exception.

Rest in peace.