Nothing Falls Off The Katamari

There is a game called Katamari Damacy in which you roll a small sticky ball around the world collecting things.

At first, the ball is tiny. You pick up drawing pins, sweets, buttons, perhaps a small frog if fortune is smiling on you. Then the ball gets bigger and you can pick up books, chairs, bicycles, people, cars, buildings.

The logic is wonderfully simple: growth happens through contact, and nothing falls off.

It is difficult to spend much time inside a mature organisation without occasionally thinking about Katamari. A process picks up an exception. A project picks up a spreadsheet. A report picks up an audience. A pilot picks up users. An application picks up an integration. A meeting picks up another purpose. A workaround picks up a procedure.

None of these things necessarily began badly. That is important.

The usual story about organisational complexity imagines poor design, bureaucracy, indecision, or some malign administrative instinct slowly making everything worse. Often the truth is less satisfying. Most additions made perfect sense at the time.

The problem is that organisations are very good at adhesion. They are much less naturally equipped for shedding.

The Meeting That Changed Species

Consider a recurring meeting. It begins because a programme needs coordination. There are decisions to make, dependencies to manage, people who need to know what other people are doing.

Fair enough.

Then the meeting begins to collect things. Someone gives an update there because everyone is already in the room. A decision gets made because the right people happen to be present. Finance starts giving a monthly update. Leadership occasionally joins. Another team discovers it is useful for something adjacent.

The meeting becomes part coordination forum, part announcement channel, part decision-making body, part social infrastructure.

Then the original programme ends.

Can we stop the meeting? Not really, because it is no longer the original meeting.

This is one of the stranger properties of organisational things. They survive not only because their original purpose continues, but because new purposes accumulate around them. The meeting was created for one reason. Its existence created opportunities for other uses. Those uses became dependencies, and those dependencies became part of the argument for keeping the meeting.

The Katamari does not merely collect objects. The objects begin collecting functions.

The same thing happens elsewhere. A report depends on a spreadsheet, the spreadsheet depends on an application, the application feeds a governance process, the governance process depends on a recurring meeting, and the meeting requires the report. At that point the problem is no longer size. It is entanglement.

A tiny spreadsheet can be harder to remove than a large application if enough things have quietly learned to depend on it. A three-step process can be almost impossible to change if one of those steps has become the informal place where five other things happen.

Mature organisational complexity is not a pile of old things. It is a network of old things that have learned to depend on one another.

This is where the Katamari metaphor begins to break down. In the game, things stick because the rules say they stick. In organisations, they stick because the organisation adapts around their existence.

That is much worse.

Organisational Scar Tissue

A lot of complexity arrives as an exception. Something goes wrong - an audit raises a finding, a supplier behaves unexpectedly, a project encounters a situation the normal process cannot accommodate, or something embarrassing happens in front of somebody senior. A control is introduced. An approval appears. A field becomes mandatory. Another report is requested. A special route through the process is created.

Organisational scar tissue, earned one exception at a time.

The temptation is to mock this, usually unfairly. Every scar may have had a wound. The approval may once have prevented something genuinely dangerous. The exception may have been the only sensible response to a real edge case. The report may once have answered an urgent question.

The problem is temporal. The wound heals, the conditions change, and the scar remains.

Organisations are often much better at preserving remedies than preserving the circumstances that made those remedies sensible.

The Decay Of Rationale

Organisational structures often have longer half-lives than the reasoning that created them.

Sometimes the explanation is lost through ordinary human fallibility. People forget, compress and reconstruct. Even the person who made the original decision may remember the conclusion more accurately than the circumstances that produced it.

Sometimes the people themselves disappear. Projects close, consultants leave, suppliers change, teams are reorganised. Sometimes the rationale technically still exists, buried in old minutes, archived email, forgotten tickets, or a folder called FINAL containing seventeen files. It has not vanished. It has simply become more expensive to recover than continuing whatever it was meant to explain.

And sometimes the surviving explanation is not entirely neutral. A team may own the process. A manager may derive authority from the committee. A role may exist partly because the surrounding system is difficult to navigate. A vendor may benefit from an application remaining indispensable.

None of this requires conspiracy. People naturally become defenders of the structures through which they exercise competence, authority, status or security.

So rationale does not simply disappear. It gets replaced.

The programme meeting becomes the coordination meeting, then the alignment forum, then the leadership touchpoint, then eventually an important part of the governance architecture. The structure persists while its explanation evolves.

Organisations do not merely inherit structures. They inherit stories about why those structures must remain.

Rediscovering the original story does not necessarily solve anything either. The important question is no longer why something was created. It is what now depends on the fact that it exists.

Reasons decay faster than consequences. Purpose decays faster than dependency.

That is where legacy begins.

Locally Legible, Globally Illegible

This does not mean nobody understands anything. Quite the opposite. The application has an owner. The report has an owner. The process has an owner. The committee has terms of reference. The policy is documented. Everything looks reassuringly governed.

And still nobody understands the whole.

The important structure often exists between the components. Someone downloads a spreadsheet before the official workflow starts. Another team uses a report for something its creator knows nothing about. A person manually reconciles two systems every Thursday. A supposedly obsolete meeting survives because it happens to contain everyone required to make a particular decision.

The pieces are legible. The relationships are not.

An organisation can be locally legible and globally illegible.

That matters because removal is a question about relationships. You are not really asking, "Do we still need this report?" You are asking, "What has learned to depend on this report?"

That second question is much harder.

Ritual

Once causal understanding decays far enough, behaviour begins to look ritualistic.

"We have to produce this before month-end."

Why?

"Finance needs it."

Who in Finance?

"Not sure."

What do they do with it?

"They always have."

The action remains while the explanation weakens. Eventually we arrive at something close to superstition: we no longer know exactly why we do it, but we have learned not to disturb it.

That sounds absurd until you remember that there genuinely might be a warehouse somewhere depending on column G.

Which is why organisational superstition can be perfectly rational behaviour. The structure is too entangled to understand completely, the consequences of removal are uncertain, and continuation has at least demonstrated that the organisation can survive another day.

Familiarity begins to act as a proxy for correctness.

I once reimplemented an agricultural forecasting system and was instructed to reproduce known bugs from the old one so that reports from the two systems would match.

The error was understood. Correcting it was not the safe option.

If the new system reproduced the old numbers, everyone knew where they stood. If the numbers changed, even because they were now more correct, the discrepancy had to be explained. Historical comparisons shifted. Confidence in previous reports became awkward. Somebody had to own the change.

The mistake had become part of what correct looked like.

Continuity was easier to defend than correction.

That is the less romantic form of organisational conservatism. Not a belief that the old way is better, but an unwillingness to become the author of a new failure.

Building On Ruins

This is why so much organisational change that presents itself as greenfield is actually brownfield.

You do not begin with "What should this be?" You begin with "What already exists, what has grown around it, and what will break if we disturb it?"

A new system inherits reporting conventions from the old one. A new process inherits its exceptions. A new team inherits responsibilities nobody consciously assigned to it. We are often building on the ruins of previous organisational civilisations.

The builders have gone. The original purpose has disappeared. The structure remains, and somebody builds something new around it.

Eventually archaeology and architecture begin to look the same.

Time launders contingency into necessity.

This is legacy in its literal sense - not merely the old thing that remains, but the consequences the new thing is required to inherit.

Legacy is the gap between what we would choose today and what we can safely stop doing.

The Burden Of Subtraction

This is where accumulation becomes self-preserving.

Adding something and removing something are subjected to different standards. A new report needs a plausible audience. A new meeting needs a reason. A new control can be justified by a failure. Subtraction asks a harder question: are we confident enough about the consequences to remove the thing?

In a globally illegible organisation, that can approach an impossible evidential standard. We are asking someone to reason confidently about a counterfactual inside a system nobody fully understands.

The inability to demonstrate that something remains useful can matter less than the inability to demonstrate that removing it is safe.

But even that is slightly too noble. The organisation is rarely asking for proof of an absence of harm. It is asking something more personal.

Who owns the consequences if we are wrong?

A report can continue for ten years without anyone consciously deciding to renew it. A meeting recurs because it is already in everybody's calendar. A system stays live. An approval remains in the workflow. Gradually, continuation stops presenting itself as a decision.

It becomes environment.

Removal is different. Someone has to cancel the meeting, stop producing the report, remove the control, decommission the application or delete the data.

Subtraction has an author. Continuation often does not.

That changes the economics of organisational behaviour. The cost of continuation is distributed while the responsibility for subtraction is concentrated.

Keeping a report might waste hundreds of hours each year, but those hours arrive in tiny increments across dozens of people. Nobody sees an incident called REPORT CONTINUED.

Cancel it and discover three months later that somebody depended on it, and the causal story becomes beautifully simple: we stopped the report, then this happened.

Nobody was ever fired for buying IBM. Very few people are punished for not removing a report either.

Inaction is seldom judged as harshly as action because inaction is harder to locate. Remove something and a decision appears. With the decision comes a decision-maker.

The Presumption Of Continuance

This produces a peculiar reversal.

At creation, the question is: why should this exist?

Years later, it becomes: why is it safe for this not to exist?

The burden of proof has moved. Meanwhile the structure has become more entangled and the reasoning behind it has decayed.

Continuation therefore acquires a structural advantage. It no longer has to demonstrate value. It merely has to remain less visibly dangerous than removal.

Persistence creates dependency. Dependency reduces reversibility. Rationale decays. Causal understanding weakens. Removal becomes harder to defend. Continuation creates further dependency.

The Katamari begins protecting its own shape.

Not because everything still matters, or even because everyone believes it does. It becomes self-preserving because it is increasingly difficult to know what does not.

Retirement Is Not Shedding

Organisations do retire things. Applications are switched off, committees dissolve, reports are cancelled and processes are replaced. But retirement is not necessarily shedding.

A new application reproduces the old reports. A new process inherits the old exceptions. A new system carries forward the old data structures. The committee disappears, but its functions are distributed into three new meetings.

The object has gone. The accumulated shape remains.

The Katamari can lose an object while preserving its form.

This is why replacement programmes can produce remarkably little simplification. The old system disappears, but its history survives.

The Anti-Katamari Department

There are obvious responses: sunset clauses, expiry dates, annual reviews, lifecycle governance, application rationalisation, perhaps even committees responsible for closing committees.

All of these can help. They also contain a small problem.

Organisations are exceptionally good at responding to complexity by creating new organisational machinery.

"We have too many controls."

"Let's establish a controls rationalisation board."

Excellent.

The Katamari has picked up the anti-Katamari department.

The answer cannot simply be another process whose purpose is to make us have fewer processes. At some point, continuation itself has to become visible again as a choice.

A useful question is not only "Can we prove this is safe to delete?" but "If this did not exist today, would we create it?"

Then comes the harder brownfield question: if we would not create it, what has become dependent on it since we did?

The distance between those answers tells us quite a lot.

Shedding Is Part Of Care

I have argued elsewhere for the importance of maintenance, and I still believe that. Creation gets more attention than upkeep, while much of what keeps organisations functioning is quiet, unglamorous care.

But care and preservation are not the same thing.

An organisation that cannot accumulate cannot learn. An organisation that cannot retain anything cannot build memory. But an organisation that can only add and preserve eventually loses its ability to adapt.

Maintenance without shedding becomes preservation.

Healthy systems do not remain healthy by keeping everything they ever acquired. They maintain themselves partly by letting things go.

The difficult part is knowing what.

Mature organisational complexity is not simply clutter. Some of it is capability, some memory, some scar tissue, some ritual, some archaeology, and some of it is unexpectedly load-bearing.

That is why simplification is much harder than tidying.

Nothing Falls Off

At the beginning of Katamari Damacy, nothing falling off is funny. It is the rule that makes the game work. The ball grows, the world disappears into it, and the scale becomes ridiculous.

In organisations, the problem is not simply that the ball gets large. It is that the things stuck to it begin to depend on one another while the reasons they were attached slowly disappear.

Eventually the organisation can no longer confidently distinguish between what is valuable and what is merely dangerous to disturb.

The mature organisation does not preserve everything because it knows everything matters.

It preserves everything because it no longer knows what doesn't.