The eight that are broken and kept anyway

What they are

The chair is the largest and the most clearly finished. A chair that will not stay at a height is not a chair, and it has been not a chair for some time. Readers comparing this small inventory with broader advice can also see iFixit.

Why they stayed

The accounts divide into three, and the first covers five of the eight.

An intention to repair, not acted on: 5. The object was set aside to look at later. Later has not arrived. In two cases the owner said plainly that it was not going to.

Uncertainty about whether it is actually broken: 2. A cable that works intermittently and a headset that works on one side. Neither is usable and neither is definitively dead, and that ambiguity is enough to prevent a decision.

No account at all: 1. The screen. It is in a cupboard, it does not switch on, and nobody can say why it is still there.

The repair intention

Five of eight, and the most interesting thing in the group, because repair appears nowhere else in this record.

Two hundred and eighty-four objects. Nineteen failures. Not one successful repair recorded anywhere, and five objects held in a state of pending repair that nobody expects to perform.

Two accounts gave a reason. One was that the object was not worth the trouble, which is an evaluation the record does not price. The other was that repairing it would require knowing what was wrong with it, and finding that out was itself the task nobody had done.

The record contains five intentions to repair and no repairs.

Whether that reflects these eleven workplaces, or the kinds of object in them, or something wider about equipment of this sort, the record cannot say. It can say that in this inventory, a broken thing is either thrown away fairly soon or kept indefinitely, and nothing in between ever happens.

Why eight rather than nineteen

Breakage is the only stopping reason in the record where more objects left than stayed: eleven gone against eight kept, forty-two per cent retention against sixty-four per cent overall.

That is the mechanism the whole cupboard section turns on. A failure is an event. Somebody picks the object up, finds it does not work, and is standing there holding a decision.

Most of the time they make it. Eleven times out of nineteen, the decision was to get rid of the thing.

Compare the twenty-four objects that stopped being reached for, where nobody was ever holding anything, and nineteen of twenty-four stayed.

The ambiguous two

A cable that works sometimes and a headset with one working side.

These are the hardest objects in the record to code, and they are coded as broken on the owners' own description. Neither is in use. Neither is fully dead.

They are also the two most likely to stay indefinitely, because the only way to resolve them is to test them, and testing them requires somebody to care more than anybody does.

A record with a field for condition rather than a binary for working would have caught these two properly and would have required every one of the two hundred and eighty-four entries to carry a judgement, which is exactly the kind of judgement this method was built to avoid making.

What this group costs

Very little in space: three cables, two keyboards, a headset, one screen, one chair.

The chair is most of it. A broken chair occupies the footprint of a working chair and offers nothing, and it has been doing that in one of these eleven workplaces for an unknown period, because the record has no dates.

Where the eight sit in the cupboard

The kept group is eighty-six objects: seventy-eight intact and these eight. By kind, the eight fall where the cupboard is already heaviest.

kind                broken kept   intact kept
cables and adapters          3            21
input devices                2            14
screens and stands           1            11
audio                        1             8
seating                      1             4

Five kinds, and no others. Nothing broken and kept is a desk, a lamp, a storage item or a printing item — the four kinds that between them hold eighty-three objects in this record and produced three failures, two of which left the building.

So a broken object is kept in exactly the categories where working objects are kept, and in roughly the same proportions. Whatever makes somebody store a working cable makes them store a dead one.

The proportion that stays broken

Nineteen failures, eight kept, eleven gone. Forty-two per cent.

Compare the intact stops: a hundred and sixteen objects stopped while working and seventy-eight of them are in storage, which is sixty-seven per cent.

A failure therefore roughly halves the chance that an object stays in the building, and that is the largest single effect any stopping reason has on any fate in the record.

It is also the only one with an obvious mechanism. A broken thing has to be picked up to be found broken, and being picked up is the whole of what produces a decision.

Why they are counted separately

Seventy-eight intact things in a cupboard can be read as a store. Eight broken ones cannot be read as anything except a backlog.

Folding them together would have produced a single figure of eighty-six that the prudence reading could claim in full, and that would have been false for eight of them. Nobody is keeping a chair that will not hold a height against future need.

So the record publishes two numbers: seventy-eight that are available in principle, and eight that are not available in any sense and are present anyway.

The second number is small. It is also the one place in the record where keeping something cannot be defended at all, and leaving it inside the larger figure would have made the larger figure look better than it is.