Chapter 7: The cascade
Things fall apart; the centre cannot hold. — W. B. Yeats, 'The Second Coming'
So far
Chapter 5 ended with a warning. When fabric thins, orientation becomes an individual burden rather than an organisational responsibility. People navigate alone; in a complex system, that is exhausting. Chapter 6 showed what the three fabrics look like when they are working, and the conditions for each. This chapter describes what happens when they stop working, and in what order.
That order matters. Fabric does not fail all at once. It fails in a sequence, each stage making the next more likely, and the sequence has a direction: from the outer world inward, and from the present into the past. By the time the last stage is reached, the organisation has lost not just its connective tissue but a significant part of what it knows.
And on to chapter 7, what are the five stages of the fabric degrading
If you are trying to understand what happened, or what is happening, it helps to have a shape for it. The cascade tends to run in five stages.
- Overwhelm makes people look inward.
- The connecting things stop being tended.
- Belief in those things erodes.
- Practice stops.
- Diffusion of ideas and answers dries up.
Each stage follows from the one before. Each is also a stage of forgetting, and by the time the fifth stage is reached, the organisation has lost not just its connective tissue but a significant part of what it knows. It is working harder than it was, and it is doing so without noticing what is gone.
Stage one: overwhelm
It begins with something that looks like being busy. The messages arrive from different directions, each reasonable on its own: a priority from one part of the organisation, a question from another, a decision that needs to be made before the information to make it has arrived. The problem is not the volume, it is the incoherence. The signals do not add up to a picture, but they point in different directions, different timescales, or address different versions of what the work is for.
The natural response is to turn inward. When you cannot orient to the signals arriving, you fall back on what you can see clearly, for example, the immediate work, the team around you, and the deadlines in front of you. Cross-team conversations start to feel optional; so does the community of practice, the check-in, the sideways relationship that keeps two parts of the organisation in touch. These things are not optional, they are the fabric, but they are the first to go when you are trying to find ground.
At this stage, the fabric isn't failing yet, but it is degrading. The connections are still there; they are just being used less. The people who maintain them are still present; they are just more stretched. At this stage, it is possible to reverse the direction. What is required is an accurate diagnosis of what is happening and a decision to treat it seriously. That diagnosis is rarely made because, from a distance, the signal that the fabric is degrading looks like people being appropriately focused, or just struggling with volume, and it will get better. If it is not a volume problem but an orientation problem, it won't.
Stage two: the connecting things stop being tended
More signals are the community of practice that met monthly cancels twice and stops scheduling, the cross-functional forum that used to surface problems early becomes a status update and then stops happening, and the check-in between two teams that kept each other oriented goes from weekly to monthly to whenever someone remembers to send the invite.
None of this is decided, and maybe the pattern goes unnoticed. There is rarely a meeting at which someone says: we are stopping the community of practice. Instead, a series of reasonable decisions is made, each in isolation by people who are stretched. The meeting is cancelled because this week is particularly bad. It is not rescheduled because next week looks the same. After a while, the absence becomes the norm.
This is not the absence of the connecting things, but the gradual displacement from the rhythm of the week. The rhythm is what the fabric depends on. Chapter 6 described how leadership fabric works through cadence; the regularity of the conversations that orient people, not their content alone. The same is true of community fabric: it is the meeting that happens even when nothing urgent is on the agenda that sustains the relationship, trust, and shared understanding that make the community useful when something urgent does arrive.
Remove the rhythm and you have not removed the community, but you have removed the condition that keeps the community useful.
Stage three: belief erodes
The third stage is easier to miss than the second, and possibly more damaging. This is the stage at which people stop believing the things that connect them matter. They have lived without them for a while, and leadership seems to be comfortable with how things are.
This is not cynicism, though it can look like it, and there are signals to back it up. The community of practice has not met in four months is a signal about what the organisation values. The forum that used to surface problems early has become a status theatre is a signal about whether honest signal is welcome. The check-in between teams was dropped; that signals whether the relationship it sustained was real or only a nice-to-have.
People read these signals accurately, and they adjust their behaviour accordingly. They stop bringing the problems they would have brought to the community, stop maintaining relationships they can see the organisation does not reward, and stop investing in the trust the rhythm had been building, because the rhythm has gone and the trust was built on it.
What is being lost here is not just the behaviour but the expectation behind it. Stage two is about the absence of the connecting things. Stage three is about people updating their model of the organisation: concluding that the connecting things were not, in the end, what this organisation was going to be built on. Once that conclusion is drawn, it is difficult to reverse. The evidence that it was wrong would come from the connecting things themselves: the community of practice that surfaces problems early, the cross-team conversation that shows the organisation is still one thing. But when those are removed, they also stop serving as messengers of healthy fabric; the diagnostic indicator is no longer there.
Stage four: practice stops
In stage three, people stop believing, but in stage four, they stop doing.
The knowledge-sharing that used to happen informally, e.g. at the edge of a meeting, over lunch, in the five-minute conversation after someone presented, stops being initiated. This is not because it was ever formally required, but because it depended on a social fabric that made it feel natural and worth the investment. When that fabric has degraded enough, the initiation cost rises. The person with something to share is less sure it will be welcome, the person who might benefit is less likely to reach across the gap to find it, and the gap widens as both sides fail to act.
The same thing happens with the communities and practices that produced learning. The retrospective that used to examine what the team could do differently becomes mechanical, the peer review that used to surface assumptions becomes a formality, and the coaching conversation that used to build capability becomes an appraisal.
The activities may persist, but the practice inside them does not. The organisation is still doing the retrospectives, peer reviews, and coaching conversations, but it is not doing the actual work those activities were designed to do. The metric stays green while the capability drains, and people's heroic efforts keep it green until the tank is empty.
Stage five: diffusion dries up
The fifth stage takes the longest to notice and the longest to recover from.
Diffusion is the movement of ideas and answers between people. It is how the person who worked through a problem in one part of the organisation leaves a trace that someone in another part can follow, without ever knowing there was a connection. It depends on community fabric: on the relationships and the trust and the shared forums that make knowledge mobile. When those things are gone, knowledge stops moving. It stays with the people who have it, in a form they can use, inaccessible to anyone who was not in the room when it was made.
This is not primarily an information management problem. Better documentation systems, improved knowledge bases, or more comprehensive after-action reviews do not solve it. The knowledge that matters most in a complex organisation is tacit: it is held in the judgment people have developed, the pattern recognition built through experience, the sense of what will work here and what will not, in this specific context, with these specific people. That knowledge cannot be captured in a document. It lives in the community that holds it, and it moves through the relationships the community sustains.
This makes stage five hard to diagnose accurately. By the time diffusion has dried up, the visible symptoms are people not sharing and not being curious. Those are real, but they aren't the cause. They happen when questions stop being taken seriously, when leadership signals that curiosity isn't rewarded, and when the forums that made sharing natural are removed. The response to a misdiagnosis is to tell people to share more and be more curious. It is an easy instruction to give, but it targets the symptom and adds another signal to the ones that have already accumulated: that the organisation sees the problem as a people problem, not a fabric problem, and not a leadership problem. This is not a fantasy hypothesis; I've been there, and it was not a people problem.
When those relationships are gone, and those communities stop functioning, the knowledge does not go into the archive. It goes with the people who held it, and it goes with the conversations that might have happened but didn't. The organisation is left with its documented knowledge and its institutional memory, neither of which is the same thing as knowing what to do.
Memory
An organisation has two kinds of memory, and the cascade treats them differently.
The first is memory of its identity, which is what the organisation thinks it is. Its values, story and sense of its own character. Identity memory tends to be robust as it is recorded, repeated, displayed, and reinforced by the language of leadership. When an organisation talks about itself, it usually draws on identity memory, and this sort of memory matters. A shared sense of purpose and character creates the conditions for belonging and commitment, but it is not what keeps the organisation viable.
The second is capability memory, which is what the organisation knows how to do. This is the accumulated learning about how to solve the problems this organisation faces, in this context, with these constraints, and these people. It is the knowledge of which approaches do not work, which relationships are load bearing, which informal norms keep the system stable, and which people to call when a particular kind of problem arrives. Capability memory is not stored in documents. It is stored in communities, in shared practice, accumulated experience, relationships that make it possible to retrieve and apply what was learned.
When communities stop functioning, capability memory decays. The document that records the outcome of the last project remains. The process map that describes how the work is supposed to flow remains. What does not remain is the judgment built from doing the work, the understanding of why the process was designed that way and what the exceptions are, the network of people who know what happened and what it taught them. The document is a postcard from somewhere you can no longer visit.
You cannot offset history
Conservation biology has a term called biodiversity offsetting. The idea is that the ecological damage done in one place can be compensated for by ecological improvement somewhere else. For example, if you destroy a wetland here, you create or restore a wetland there, and the balance is maintained. It is a tidy theory, but it fails in practice for a reason relevant to this discussion.
The wetland you destroy is not just a collection of species and habitats; it is a system that has been developing, adapting and accumulating for decades or centuries. The relationships between its organisms, the soil chemistry, the hydrology, and the microhabitats that exist only here are not portable. You can create a wetland elsewhere with some of the same species and some of the same structures, but you cannot create one with the same history. The ancient woodland that takes three hundred years to develop its full ecological complexity cannot be replaced by a three-year-old plantation.
Something similar is true of the capability memory that communities have.
When a community of practice breaks up, a team is dispersed, or the people holding a practice together are restructured away from each other, the capability memory they carried does not transfer to the new arrangement. You cannot take the knowledge out of one community and deposit it into another. The relationships that made the knowledge retrievable and applicable are specific to the people who built them, in the context in which they built them. The glue roles, which are the people who connected different parts of the community, translated between groups, and maintained the relationships that kept ideas and problems moving, are probably gone now. When they go, what they know and did goes with them.
This is not a failure of documentation. The community might have excellent documentation and much of it is valuable, but the document is like a postcard from somewhere you can no longer visit. It tells you someone was there and what they found, but it doesn't give you the judgment to know what to do with it in your situation, right now. The judgment lived in the community, but it is gone.
The metaphor has a practical implication. It means that some of what the cascade destroys cannot be rebuilt from scratch. It can be built again, over time, by people working together on real problems in conditions of trust, but it cannot be rebuilt by decision, by announcement, or by the kind of reorganisation that creates a new structure and assumes that the capability will follow. The structure can create the conditions, but it cannot create the history.
The fresh start myth
Organisations do something, usually when the cascade is advanced, that makes things worse while feeling like progress. It goes by various names: transformation, reset, new chapter. In practice, it involves a new leader or leadership team deciding to discard what came before and start again.
The instinct is understandable, as the organisation is underperforming and the previous approach didn't work. A new direction is needed, and new directions require a clean slate. The language of fresh starts is everywhere in organisational life: "we are building from the ground up; we are not constrained by what was done before; this time will be different".
The problem is that "what came before" is not uniform. It contains formal models and processes that didn't work and can be discarded. It also contains the culture and habits that produced the dysfunction and need to change. But it also contains the capability memory the organisation has accumulated. This is the hard-won understanding of what works and what does not in this specific context, the relationships that make it possible to get things done, and the judgment built through years of encountering real problems. That cannot be discarded without cost, because it is not easily replaceable.
New leaders who make this move do not, in general, intend to destroy capability memory. They intend to change direction, but changing direction by discarding what came before has a consistent pattern of outcomes. The new direction borrows its ideas from wherever the new team came from, which is a different organisation facing different problems, it repeats cycles of change that the organisation has already been through because the people who remember those cycles are no longer there, and it spends its first year rebuilding the capability that was destroyed, usually at some cost, before it can begin to build anything new.
The question a leader should ask here is specific. The question is not should we change? It is: what are we preserving because it is who we are, and what are we preserving because it is what we know how to do? Identity memory and capability memory are not the same thing, and they do not have the same relationship to change. The first can be argued with, adapted, replaced. The second needs to be identified and carried forward, especially as the direction changes.
What cannot be recovered
One version of the cascade argument treats it as a cycle: fabric degrades, it is rebuilt, and the organisation learns and recovers. That version is partly true, and fabric can be rebuilt, communities can be re-formed, and trust can be restored. The conditions for capability memory can be created again, and new capability memory can be built.
However, the cycle version leaves something out.
The metaphor of debt is sometimes applied to this kind of degradation: technical debt, change debt, the accumulated cost of deferred maintenance, and it is a useful metaphor as far as it goes. The idea that an organisation has borrowed against its future capacity, and that the borrowing will eventually need to be repaid, captures something true about how the cascade works.
However, the debt metaphor fails in one important respect: debt implies that what was lost can be repaid in full. You can pay back what you borrowed, plus interest, and you are where you would have been. The cascade is not like that, because some of what it destroys is not recoverable; it depended on specific people, relationships, and accumulated experience that cannot be recreated from scratch.
The community that broke up five years ago, dispersed across a restructuring, cannot be reassembled by putting the same people back in the same rooms. The relationships are lost, and the shared understanding built through working together on real problems over real time is gone. What can be built in its place is a new community, with its own capability memory, over its own time. It is worth building; it may, eventually, be more capable than what it replaced, but it is not the same thing, and the five years of capability memory that would have been built in the interval are simply absent.
This is what makes the cascade more serious than a debt. Some of the loss is irrecoverable, but assuming you can always rebuild what you destroyed, if you decide to, is a form of the fresh start myth applied in reverse. The harder problem is that what was lost was rarely visible while it was there. The people who held the capability memory weren't labelled, and the communities that sustained it weren't on any org chart. When they go, there is only a gradual sense that things are harder than they used to be, and it is difficult to tell. why. The question is not whether rebuilding is possible, but it is whether the organisation has the honesty to understand what it has lost, and the patience to build what comes next.
A moment to pause
If you are living inside this system: You have almost certainly been inside a cascade. The exhaustion you felt was real. The sense that the things which used to make the work feel worth doing had disappeared was real. The growing suspicion that nobody was quite sure where the knowledge went is real. This chapter gives that experience a shape. It doesn't make it easier, but identifying it accurately is the first step toward doing something about it.
If you are leading: The cascade is easy to miss because each stage looks, from a leadership vantage point, like something else. Stage one looks like focus. Stage two looks like efficiency. Stage three looks like realism. Stage four looks like compliance. Stage five looks like normal. By the time it is visible as a cascade, the damage is advanced. The diagnostic question is not how the organisation looks now; it is what has stopped happening that used to happen, and how long ago it stopped.
If you are working to change this: The cascade tells you something important about sequencing. If the fabric is in late-stage decline, adding new structure may not work. The structure needs somewhere to take root. The minimum viable fabric, the conditions that make any model viable, must come first. The next chapter says what that looks like.