Why Companies Fail Without Being Attacked: Management as the Defense of Boundaries

There is a pattern that recurs under all the frameworks of management, and it is worth taking seriously. Every boundary that has to hold must be actively held, or something on the other side takes it. A customer relationship, a product's quality, a position in a market: none of these stays yours because you once won it. It stays yours while you keep projecting force across its edge. The founder who steps back and watches the business drift does not usually lose to a clever rival with a better plan. He stops attending the boundary, and the boundary is taken.

That is the whole claim, stated flat. The rest of this piece is the evidence for it, and the honest limits of that evidence.

What actually happens when you stop paying attention?

The default story of business failure is a duel. A competitor builds something better, attacks head-on, and wins. That story is comforting because it makes failure feel like a fair fight you could have won with a sharper sword.

The more common pattern is quieter. A function inside the company stops getting attention. A supplier relationship, a hiring pipeline, a once-loyal customer segment. Nobody is watching it. There is no dramatic loss, just a slow erosion at the edges. By the time someone notices, the territory is already half gone, and it went not to a frontal assault but to whatever was sitting next to it under pressure.

Management research has a precise name for the resource that runs out here. It is attention. The starting point is older than the management literature: in 1971 Herbert Simon observed that in an information-rich world, a wealth of information creates a poverty of attention, because attention is the limiting factor in the consumption of information. You cannot attend to everything. What you stop attending does not pause and wait for you. It changes.

William Ocasio built that intuition into a theory of the firm. In his 1997 attention-based view, he defined the thing managers actually spend as "the noticing, encoding, interpreting, and focusing of time and effort by organizational decision-makers," and argued that firm behaviour is the result of how a company channels and distributes that scarce focus. The premise is blunt. Organizational action is bounded by where managers point their limited attention. So the question is not whether you are a good manager in general. It is which boundaries you are actively holding right now, and which ones you have quietly left undefended.

It is worth saying where this sits, because the neighbour is well-trodden. Clayton Christensen's disruption story explains who attacks an incumbent and why the attacker's cheaper, worse product eventually wins. The attention-based view explains something adjacent: why the incumbent, who can often see the attack coming, does not respond in time. This piece works the seam between them. The boundary decays whether or not a disruptor ever shows up. The disruption literature supplies the rival; the attention literature supplies the neglect. I am claiming only the join: that the decay is a property of attention and pressure, not of whether a worthy opponent exists.

Is there a real timescale, or is this just a metaphor?

Here I want to borrow a picture from neuroscience, and I want to be careful about how far I lean on it. I am holding it as an analogy, not as proof that a company is a brain.

The neuroscientist David Eagleman, with his co-author Don Vaughn, published a theory in 2021 in the journal Frontiers in Neuroscience. They start from a well-documented fact about the brain. Regions maintain their territory through continuous activity. In their words, if activity slows or stops, for example because of blindness, the territory tends to be taken over by its neighbors. The visual cortex is not guaranteed to stay visual. It holds its ground only while it stays active.

The striking part is the speed. Eagleman and Vaughn note that the surprise in recent years has been how fast the takeover happens, and that it is measurable within an hour. They draw on work by Lotfi Merabet and colleagues, who found that when blindfolded volunteers performed a fine touch-discrimination task, touch-related activity showed up in the primary visual cortex after only forty to sixty minutes of darkness. In a separate study, sighted adults blindfolded for five days began recruiting the visual cortex for touch, and read Braille better for it. Remove the blindfold, and within about a day the change reverses.

Eagleman and Vaughn build a bolder claim on top of this. They argue that this is why we dream. The visual cortex is the one sense disadvantaged by nightly darkness, so the brain generates internal activity during sleep to keep neighbours from encroaching on idle visual territory. That specific dream hypothesis is contested and far from settled, and so, strictly, is the "territory and neighbours" way of describing what happens. What is actually well replicated is narrower: a deprived sensory region starts responding to a neighbouring sense within hours to days, by unmasking connections that were already there, which is the Merabet result. The "invading neighbours" language is vocabulary I am borrowing from the contested account; the fast cross-modal recruitment is the settled fact I am leaning on. I keep them separate on purpose, and only the second one is load-bearing.

Now the analogy, stated as an analogy. An organization is not a cortex. But it shares the relevant geometry. It is a set of bounded functions sitting next to each other under pressure, where holding a boundary costs ongoing energy and the surrounding regions are not idle. The lesson worth carrying over is not a number. It is a shape. The decay rate of an unattended boundary is set by the pressure around it, not by your intentions for it.

Why didn't Kodak win the digital photography it invented?

The clearest public example is Eastman Kodak, and the usual reading of it is wrong in a way that matters.

The folk version says Kodak missed digital photography. It did not. A Kodak engineer named Steve Sasson built the first working digital camera in 1975, a device that recorded a black-and-white image at 0.01 megapixels onto cassette tape. Kodak saw the future early. It owned the lab where the future was invented.

What Kodak had was a boundary it stopped defending. At its 1976 peak it held about 90 percent of the United States film market; into the mid-1990s it still carried a market value near 28 billion dollars and employed more than 140,000 people. That dominance was the thing to protect, and protecting it meant attending honestly to the digital edge that threatened it. Instead, digital stayed a side project, subordinate to film, often priced and placed so it would not disturb the profitable core.

Scott Anthony, the innovation researcher and Innosight managing partner, made this point directly in a 2016 piece for Harvard Business Review titled, plainly, "Kodak's Downfall Wasn't About Technology." The failure was not blindness to the future. It was the failure to commit attention and action to a boundary the company could already see. Kodak filed for bankruptcy in 2012. Sony, Canon, and Nikon did not beat Kodak by inventing something Kodak lacked. They moved into territory Kodak stopped actively holding.

This is the claim in one case. The boundary was the film franchise. The neighbouring pressure was digital imaging, which Kodak itself had built. The founder-scale move, walking away from the edge, was not literal departure. It was the slow withdrawal of serious attention from the place where the boundary was being tested.

I should be honest about what one case can and cannot do. Kodak is also told as a story of hubris, and as a story of business-model inertia, and as a textbook case of Christensen-style disruption; the same facts sit comfortably under each. So Kodak illustrates the attention reading, it does not prove it over those rivals. The thing that would actually separate them is the prediction further down, decay tracking ambient pressure rather than any direct attack, and that test has not been run on a clean set of cases.

A historical parallel, and its limit

The Maginot Line is a useful parallel for one half of the claim, and I want to be exact about which half. It was the fortified border France built against Germany after the First World War: real, expensive, and genuinely strong along the stretch it covered. The French attended that boundary with enormous care, and that care is the point.

The 1940 invasion did not come through it. German forces moved through the Ardennes forest to the north, a stretch treated as too difficult to need heavy defending. The strong wall held and was simply bypassed. So the part of the analogy that survives is the misallocation of attention: the defenders bound their effort to the fortified perimeter while the decisive perimeter went under-attended. The part that does not survive is the rate claim. The Ardennes did not quietly erode under idle adjacent pressure because it had gone unattended; it was overrun by a directed assault aimed exactly at the weak point. The loss arrived as a frontal attack at the under-defended spot. So this case illustrates "attention pointed at the wrong boundary," but it does not support the sharper half, that decay is set by ambient pressure rather than by attack, and I will not stretch it to.

How would you know if this claim is true for you?

A claim worth anything has to be falsifiable, so here is the operational version. Pick any boundary your organization depends on. A key customer relationship. A core part of the codebase. A defensible niche in your market. Stop attending it, on purpose or by neglect, and watch what happens.

The claim predicts something specific. The boundary will decay at a rate set by the surrounding pressure, not by any frontal attack. A customer with three hungry alternatives nearby erodes fast. A customer in a sleepy market with no real substitute erodes slowly. Same neglect, different decay rate, because the rate is a property of the neighbourhood, not of your attention alone. To make that a real test and not just a saying, pin both sides to numbers chosen in advance: measure decay as one observable, the time for a customer cohort to half-leave, say, or for a codebase's unaddressed defects to double, and measure pressure as the count and funding of substitutes already pressing on that boundary. The sharp version is an ordering prediction: across cases, the boundaries that decay fastest should be the ones facing the most substitution pressure, and the decay should not line up with who launched a direct attack. If instead you find that unattended boundaries hold steady regardless of what surrounds them, or that decay tracks open attacks rather than ambient pressure, the claim is wrong and you should ignore me.

What this rules out is the comforting excuse. If you lose a position, the claim says do not first look for the brilliant competitor who beat you. Look for the boundary you stopped projecting force across, and ask what was sitting next to it. Most of the time the answer is not a duel you lost. It is a fence you stopped mending.

So what is management, actually?

Strip away the vocabulary and a manager's job is narrower and harder than the org chart suggests. It is to decide, against a fixed and scarce budget of attention, which boundaries get actively held this quarter and which are allowed to soften. That is Ocasio's scarce resource spent on the thing it is uniquely for.

The founder who walks away rarely loses because someone outfought him. He loses because a boundary stops being projected, and the territory next to it was never idle. Attention is the force that holds the edge. Stop spending it on a given boundary, and the only open question is how fast the neighbours arrive. That speed is set by them, not by you.

I will not pretend the biological story proves the management one. The brain is not a company, and I have kept the cortex strictly as an analogy. I will be as careful with the management side: that a held boundary gets eaten when you stop attending it is well supported, but the larger claim, that all of management reduces to this single act, is an organizing hypothesis I find useful, not a proven law. I offer it as a lens, not a verdict. What I will stand behind is the pattern both share. A boundary you stop attending gets eaten. That is the whole claim, and a business audience can check it without a microscope: in a market. Most of the work is deciding which boundaries are worth the force.