Four things it refuses

1. That desk equipment is reliable

The reading. Nineteen breakages in two hundred and eighty-four objects. One in fifteen. Equipment on these desks almost never fails. The useful question is whether a change removes friction without adding unnecessary oversight, so managers may operational data for equipment decisions before setting the pilot measures.

Why it fails. Three reasons, and they stack. Another useful source on technology, maintenance or purchasing is Trusted Reviews.

Breakage competes with other ways of leaving service, and loses. Thirty-eight objects were replaced while still working and would have been available to fail if they had not been retired first. An object removed from service cannot subsequently break, so the share of endings that are failures is lower than the share of objects that would eventually fail.

A third of the record is furniture, which has almost no failure mode. Ninety-five items, three failures, and the three are two lamps and a chair's gas lift — the only moving parts furniture has.

And objects that failed and were discarded long ago are the ones most likely to be missing from the inventory altogether, because nothing remains to prompt a memory of them.

What survives. That in these eleven places, when a working life ended, the thing that ended it was usually not a failure. That is a statement about endings, not about reliability.

2. That people keep too much

The reading. Twenty-seven per cent of everything ever on these desks is kept and unused. Seventy-eight intact objects, eight broken ones, and six returns to service in the whole record.

Why it fails. It treats the figure as the output of a choice, and the record shows no choice being made.

Across eleven workplaces there is no instance of anybody deliberately reviewing their equipment and deciding to keep something. What the accounts describe is an object leaving a desk to make space and then nothing happening. The destination acquired the name cupboard afterwards.

The one test the record can run supports that. The share of objects ending up kept sits between twenty-five and thirty-four per cent across all seven origins — a thing somebody chose and paid for and a thing that was simply in the room when they arrived end up kept at nearly the same rate. A preference would show variation there. There is none worth the name.

What survives. That when nothing obliges anybody to decide, things stay. Eighty-six times out of a hundred and thirty-five.

3. That the figures transfer

The reading. Fourteen per cent of stops are breakages, a quarter of cupboards are cables, the median age in use is six years — useful benchmarks.

Why it fails. Eleven workplaces, selected by whoever agreed to be counted. No population, no weighting, no second observation.

Worse, the eleven are not independent of each other: most of the thirty-one work-changed stops are the same shift away from paper, counted once or twice in each place. The figure of thirty-one is nearer to three or four events than to thirty-one.

What survives. That these patterns occur, and how the parts of this record relate to each other. Nothing about frequency anywhere else.

4. That any of this is advice

The reading. If things rarely break and often sit unused, then buy fewer and better, or audit your cupboard, or do not replace what works.

Why it fails. Every one of those requires a price, and no price was recorded anywhere in this inventory. Without one there is no way to weigh twenty-one held cables against four avoided purchases, or a replacement against a thing that still worked.

Every one of them also requires a judgement on objects, and no object in this record was evaluated. Not examined, not assessed, not compared. Nothing here knows whether anything was any good.

Every one of those readings needs either a price or a judgement on an object, and this record contains neither.

What survives. A count of what became of two hundred and eighty-four things. The page after next sets out why the record was built so that the obvious recommendation would be impossible to make from it, and the one after that says what a second inventory would settle.

5. That any kind of object is better or worse than another

The reading. Desks last nine years and audio equipment four. Storage lasts eight and screens five. There is an order here.

Why it fails. The order is a list of how many ways a thing can stop being used, not a list of how well things are made.

A desk can stop for one reason and did so twice in twenty-four objects. A screen can stop for five and did so twenty-one times in forty-one. Printing equipment sits fourth on the list on the strength of two surviving objects while six of its eight were being retired wholesale.

And cables sit fifth — above screens and input devices — while being the most frequently retired, most frequently broken and most frequently forgotten kind in the whole record.

What survives. That in these eleven places, things with no moving part and no better version available stayed longer. Which is a statement about categories and was not in doubt.

What the five refusals have in common

Four of the five fail on something missing: no stopping dates, no prices, no second observation, no object ever examined.

The fifth fails on something present. The kind field is in the record and is the one field that was not asked of anybody — assigned afterwards from a list of ten settled at the pilot — and every per-kind figure on this site is a count of that classification rather than of an answer.

Where the record is firmest is where it used only the four asked fields. Where it is most interesting is usually where it did not.

The shape of all four refusals

Three of the four fail on the same two absences: no stopping dates, and no prices.

The fourth fails on something chosen rather than missing. The record could have held brands and prices; it was built not to, and the cost of that decision shows up here, as four readings it cannot deliver.