Chapter 8: How you know the fabric is degrading
You see, but you do not observe. Arthur Conan Doyle, A Scandal in Bohemia
So far
Chapter 5 said what the fabric is, and Chapter 6 described the three forms it takes. Chapter 7 followed it as it fails: five stages, each making the next more likely, with the organisation forgetting as it goes and some losses beyond recovery.
All of that is description. This chapter is about your own organisation, and how to tell which of those states it is in. Chapter 5 sets out the difficulty: working fabric is invisible, and the dashboards stay green while it degrades. The usual instruments will not tell you, so you must find out another way.
And on to Chapter 8, and ways of looking...
There is no score for fabric, and this chapter offers none. A number would be convenient, but people would treat it like any other target and manage it for appearance.
What follows are several ways of looking. Each is a question, or a small set of them, that can be asked of a team, a function or a whole organisation. They rely on watching what people do and asking what has stopped. Most of them need no authority or budget, which matters for the person who has neither. They do need someone who can cross boundaries, because a view from inside one team shows only that team. That is the work of a glue role, and where those roles have been lost, the wider view is harder to get.
None of them gives a verdict on its own. Taken together they show where the fabric is working and how far along the cascade the weaker parts of the organisation have gone. That last point decides whether tending what is left will be enough, or whether the fabric must be rebuilt before anything else changes, as chapter 9 will argue.
What working looks like, and what absence looks like
Chapter 5 offered images of both. Here, they become things you can check.
Start with orientation, since it is the clearest sign that fabric is doing its work. Ask a handful of people in different places what matters most this month and why. Where the fabric is working, their answers will differ in detail and agree in substance, and they will be able to say how they know. Where it has degraded, you will get the official priorities recited back, or several confident answers that cannot all be true. Where it has gone further, you will get chaos, with no shared answer at all and each team working to its own priorities.
Then watch how bad news travels. In a healthy part of the organisation, it travels quickly and arrives as it was sent; the person who raised it can tell you what happened as a result. Where the fabric is weak, bad news is softened at each step, or held until it can be presented with a solution, and the people closest to the problem stop expecting a reply.
Look at how help crosses a boundary. Two teams can share a floor and still deal with each other only through a queue and a service agreement. The test is whether someone can ask someone on the next team for help and get it without raising a ticket or asking permission. If help only travels through the process, sharing a floor doesn't mean the two teams have joined.
Also look at meetings with nothing urgent on the agenda. Chapter 7 described the rhythm the fabric depends on: the community that meets when there is no crisis, the check-in that keeps two teams in touch. Where these still happen, and people turn up because the hour is useful to them, the fabric is being tended. Where they have turned into status reports, or lapsed, it is not.
Two cautions go with these signs.
The first is that silence is easy to misjudge. A room that has learned that questions are unwelcome goes quiet, and from the front of the room that quiet looks like agreement. When people stop telling leaders things, the silence is interpreted as stability. An absence of challenge is worth treating as a finding to be explained; it is weak evidence that all is well.
The second is that a close group can still be unhealthy. It can be tightly bound and quick to help its own members while being hard for a newcomer to enter and organised around defending how things used to be. Fabric is a means of keeping the organisation viable in pursuit of its purpose. A network that closes ranks against that purpose is a problem, however strong it is. So, the question to ask of any close group is whether newcomers and news can pass in and out.
The blank space on the map
Alongside the nomads of the Mongolian steppe, chapter 6 described the map most organisations draw of themselves, with a confident middle and a great deal of blank space at the edges. That map gives a practical test.
Every team draws its own map of the organisation, with itself in the middle. The most important map is the one drawn at the centre by head office and senior leaders, because it is the one used to decide what changes. It is made of the organisation chart, the operating model, the reporting packs and what senior people carry in their heads. Take that map and look for the places where it says little. These are the teams known only by headcount and output figures, and the functions described in a line because nobody at the centre could describe them in a paragraph.
Those are the blank spaces. The mapmaker hasn't been there, but in each one someone sits at the centre of their own world, with knowledge the organisation depends on.
A blank space tells you two things. It tells you where community fabric has degraded between the centre and the edge, because a connection that was in use would have put something on the map. It also shows where the organisation is most likely to do damage without meaning to, since what a leader cannot describe, they cannot protect when the next change comes.
The second half of the test is to go there. Ask the people in the blank space who they depend on and what they know that nobody has asked about. Then ask the question from their side: what does the centre look like from here, and how much of it is blank? The test works in both directions. If the people at the edge can no longer say what the centre is trying to do, they have lost orientation, and that belongs on the map too.
The cascade as early warning
The cascade in chapter 7 was told as a history. It can also serve as a set of early warnings, because each stage leaves evidence, and the earlier stages can still be turned around.
The difficulty is that each stage looks acceptable from above. A team that has turned inward looks focused, for example, and a lapsed meeting looks like time saved. So, the useful question is the one Chapter 7 ended on: what has stopped happening that used to happen, and how long ago did it stop?
At stage one, look at what people decline. Invitations that cross a boundary are the first to be turned down, with workload given as the reason. Look too at who is tiring, because someone is doing extra work. When shared meaning is not supplied from the top, individuals and informal networks rebuild it from fragments, and they can only keep that up for a while. The people who do most of this work often notice incoherence earliest and feel it most, and they are frequently the glue roles. They overload and pull back before anyone else does, and it is usually taken as a personal matter, with someone said to have become fragile or difficult. Treat it as an early warning: when the people absorbing the system's incoherence stop absorbing it, the fabric is under strain, and the rest of the organisation is about to feel what they were shielding it from.
At stage two, the diary is the evidence. List the recurring conversations that crossed team boundaries a year ago, and mark which still happen, which have become status updates, and which lapsed without anyone deciding to end them.
At stage three, listen to how people talk about the connecting things. Phrases such as "nice to have" and "when things calm down" show that the organisation model in people's heads has been updated. So does the absence of complaint, since only people who still believe the community matters will object when it is cancelled.
Stage four is the hardest to see, because the activity continues after the practice inside it has gone. One way to test it is to take something the organisation says it protects, such as a design review or a retrospective, and list the habits it depends on. A habit here is something people do, unprompted, that makes the thing work. In a design review, the author sends the design round early enough for people to think about it, and someone senior asks the awkward question. Describe each habit by what it achieves. Then check when each one last happened, and whether anyone who joined in the past year has learned it. A review can stay in the diary long after the habits that made it useful have stopped being passed on.
At stage five, look for the same problem being solved separately in several places, and for questions that have nowhere to go. Ask people who they would ring about a problem outside their own area; if they cannot think of anyone, diffusion has stopped for them.
Three conditions deserve attention at any stage, because they stop signals from landing even when someone is willing to send them. One is a single dominant account of how things are going, repeated at every level like a mantra until it becomes fact. Another is a cost for disagreeing with it, whether formal or social. The third is language that hides decisions, so things are "agreed" or "deprioritised" and it is hard to find who made those decisions. Where all three are present, a true signal can be raised and will change nothing, and people soon stop raising them.
How far the cascade has run in each place matters for what comes next. In the early stages, the connections are still there and tending them is enough. In the late stages, the question changes, and chapter 9 takes it up.
Three kinds of behaviour
Most of what you can observe in an organisation is behaviour, and the same behaviour can mean different things. Sorting what you see into three kinds helps before deciding what it tells you.
Purposeful behaviour comes from what a person believes about themselves and their situation. Functional behaviour is a sensible response to present conditions: the signals people receive about what is rewarded and what is risky. Persistent behaviour is residue from an earlier manager or an older process; it was purposeful or functional once, and it has carried on because nobody has given permission to stop.
Take someone who copies their manager into every email. They may believe that being seen as busy keeps them safe, which would make it purposeful. It would be functional if the last person who decided without cover was publicly blamed for it. It would be persistent if a manager who left three years ago insisted on it and the habit has outlived the reason.
The action is identical in each case; the reason for it is different, and so is what it tells you. The first says something about the person and the story they are working from, which later chapters return to. The second is information about the fabric as it is now. The person is responding sensibly to the conditions around them, so if the behaviour is to change, the conditions must change first. The third is information about the organisation's history, and the right response is often to remove it by stating plainly that it is no longer required.
Sorting behaviour this way guards against the most common mistake: treating everything as the first kind. When all behaviour is attributed to attitude or character, the response targets individuals, and the conditions that produced the behaviour remain unchanged. Chapter 2 argued that people respond to signals more than to the stated design. Functional behaviour shows this, making it the most useful of the three for judging the state of the fabric.
Persistent behaviour has its own use. A large amount of residue suggests an organisation that has changed its structures many times without clearing up afterwards. People are carrying the requirements of several past models at once, and that load contributes to the overwhelm at the start of the cascade.
The keystone question
It is tempting, having accepted that connections matter, to try to map them all. A network survey (organisational network analysis, the method most associated with Rob Cross) can produce such a map, but it dates quickly, and it treats every connection as equal when a few matter far more than the rest.
The ecologist Robert Paine cleared the starfish from a stretch of rocky shore and kept them off. The mussels they had been eating spread and crowded out most of the other species. Paine called the starfish a keystone species: one whose removal causes the whole system to reorganise into something poorer. His finding suggests a narrower and more useful question than "what is connected to what". Ask instead: what, if removed, would leave this organisation still running, and worse?
The wording matters. Losing a keystone role rarely produces a failure anyone can point to. The work continues, and the reports are filed, while the system settles into a slower arrangement with each team less aware of the others; since nothing has visibly broken, nobody investigates.
There are a few practical ways into the question. Ask people who they go to when two teams need the same thing at once, or when a piece of work is stuck between departments; the same few people tend to come up, and they are seldom the ones the chart would suggest. For any single connection, ask whether cutting it would change an outcome that someone would lose sleep over. If it would, it is important, whatever its formal status.
A related question applies to knowledge. For each thing the organisation depends on, such as a core system or a pricing model, ask who still understands what sits beneath it. Some capabilities are retired on purpose and openly, as the sundial gave way to the watch. Others are lost by drift, with each new group of staff having slightly less reason to learn what the last one knew, until nobody can question the thing they all rely on. If one person is left who can go below the surface, find out whether that person is being protected or is next in line for a saving. Redundancy exercises rarely look at this level of detail. Someone decides they no longer need a role, and nobody asks what else depends on it.
Ask the same question before any improvement. A local arrangement that looks untidy from the centre may fit conditions the centre has not seen. Asking what this part knows that improving it would destroy is cheap. Standardising first and finding out later risks the kind of loss chapter 7 described, which cannot be made good.
The leader's questions
Anyone can ask the questions so far. Two more belong to people with authority, because they concern what leaders do with what they see.
The first is a test to apply before changing anything. Chapter 7 separated two kinds of memory, identity and capability, and described the fresh start myth: the new leader who clears away what came before and takes the organisation's working knowledge with it. The test is the question that would have prevented it. Of each thing we are keeping, or proposing to remove: is it here because it is who we are, or because it is what we know how to do?
The answer is not always obvious, because people come to treat a way of working as part of who they are, and the two become hard to tell apart. A team may defend a process as part of its identity when it is only the current method. The opposite mistake is a leader sweeping away a practice as an old habit when it contains knowledge that exists nowhere else. A supplementary question helps: would removing this break trust, or only familiarity? Familiarity returns with time, but trust and working knowledge do not.
Applied early, the test turns a new leader's first months into a study of what the organisation already knows. That is slower than announcing a direction, and the pressure to produce something visible is strong, but it is the cheapest point at which to find out.
The second is a question to ask in person, of the people doing the work. Leaders are buffered from the constraints under which work gets done, and the usual remedies, such as site visits and town halls, collect opinion and little else. The question beneath opinion is: what made this hard, and what did the system force you to choose between?
It is best asked about one specific piece of work, followed from start to finish. The answers show where the formal design and the real conditions part company: the two targets that cannot both be met, the approval that everyone routes around. They also show the state of the fabric in passing. Someone who answers freely, with examples, tells you that honest signal can still travel upward. A polite assurance that everything is fine suggests that it cannot, which returns us to the caution about silence.
The answers to these questions tell you where the fabric is working and how far it has degraded elsewhere; what you can do with that knowledge depends on whether you have the authority to change the conditions. The next chapter is for the reader who can now see all this and cannot redesign any of it; chapter 9 then turns to those who can.
A moment to pause
If you are living inside this system: You have probably been noticing these things for some time. The meeting that lapsed and the colleague who stopped speaking up are both evidence. Write down what has stopped and roughly when, because a dated list is harder to dismiss than a feeling.
If you are leading: A quiet organisation may only have stopped telling you things. Pick one of the questions in this chapter and ask it yourself, in person, about one piece of real work. Then notice how easy or hard it was to get a straight answer, because that tells you something too.
If you are working to change this: Look before you produce a design. The pull towards producing a target model early is strong, and a clean design made from the centre can remove what the blank spaces know without anyone deciding to. Your first job is to find out what is there, and it is worth saying so to whoever commissioned you.