Three things the record supports

1. Equipment on these desks mostly stopped being used while still working

The figures. A hundred and thirty-five objects stopped. Nineteen broke. The other hundred and sixteen — eighty-six per cent — stopped intact.

What it rests on. A single recorded field, present for every one of the hundred and thirty-five, with a closed vocabulary of six values fixed before counting began. The record here describes objects rather than output, so a distributed team may team time records for workplace reviews and interpret operational data beside completed work and employee feedback.

What would break it. Objects that broke long ago and left nothing behind were not counted, because nothing prompts a memory of them. That pushes the breakage figure down by an unknown amount. If the true figure were double, breaking would still be fourth of six reasons; if it were triple, it would be first. Readers comparing this small inventory with broader advice can also see TechSpot.

The record names this as the one place where its central finding could be materially wrong, and cannot measure it.

2. Half of what ends a working life is nothing to do with the object

The figures. Thirty-eight objects were replaced by something better and thirty-one stopped because the work changed. Sixty-nine of a hundred and thirty-five, fifty-one per cent.

Neither reason says anything about the thing that stopped. One is about what arrived next; the other is about what the person now does.

What it rests on. The same field, and the coding convention that an object with two stopping events is coded by the last one. Four objects are affected, which moves this figure by four in a known direction.

What would break it. If the twenty-four unexplained stops are mostly forgotten replacements, this figure is higher. If they are mostly forgotten failures, lower. Eighteen per cent of the record sits in a category that cannot be resolved from inside it.

3. Whether a thing leaves the building is decided by whether anybody had to handle it

The figures.

event-type reasons (broke, somebody needed it)
    12 kept, 17 gone      41% kept
drift-type reasons (replaced, work, no reason, room)
    74 kept, 32 gone      70% kept

A failure or a request puts the object in somebody's hands and produces a decision. A replacement or a drift does not, and the object stays.

What it rests on. The cross of two recorded fields, both present for all hundred and thirty-five. No recollection, no estimate.

This is the firmest finding in the record and the one least dependent on anybody's memory.

What would break it. The room-changed row sits awkwardly: fifty-four per cent kept, between the two groups, because a desk move is an event and a rearrangement is not, and both are coded the same way. Splitting that row would sharpen the contrast rather than remove it.

What the three have in common

None of them is about equipment.

The first is about how working lives end, the second about where the cause sits, the third about what prompts a decision. All three would be recognisable to somebody whose desk held entirely different objects, and none of them requires knowing what any object was.

That is the consequence of a method with no brands and no prices in it, and it is the point. A record that could not evaluate anything ended up with findings about the circumstances of objects rather than the objects themselves.

Three claims, eleven workplaces, two hundred and eighty-four things. Nothing here says what to buy, and nothing here could.

What each claim looks like crossed with kind

The three claims above rest on the four asked fields. Crossed with kind — the one field that was assigned rather than asked — each of them sharpens, and each of them acquires a caveat.

Claim 1, equipment stopped while working. Every one of the nineteen failures is in a component that moves or carries current: six cables, five input devices, three screens, two audio items, two lamps, one gas lift. The two hundred and sixty-five objects without such a component have not failed at all. The claim is really about what a desk is made of.

Claim 2, half the stops are external to the object. Printing and scanning is the extreme case: six of eight objects retired by one change of work, none of them faulty, none of them replaced. The caveat is that most of those six are the same change in different workplaces.

Claim 3, handling produces decisions. Desks and printing items have nothing in storage at all — thirty-two objects, every one either in use or gone — because neither can be kept out of sight. Input devices have more objects in drawers than on desks. The claim holds, and what it is really measuring is size.

The shape all three take once kind is accounted for

Something narrower than the claims as stated, and firmer.

Across these eleven workplaces, whether a piece of equipment is still in use depends mostly on whether its category has a way of ending; and whether a retired one is still in the building depends mostly on whether it fits out of sight.

Neither of those is about equipment quality, neither needs a brand or a price, and neither could have been found by examining an object.

What none of the three permits

A rate for anywhere else. Eleven places are not a sample and the preceding page sets out what that rules out.

A recommendation of any kind. The next page but two is about the recommendation this record was deliberately built to be unable to make.