Six years against four

By kind

Desks have no figure in the right-hand column because only two of the twenty-four ever stopped, and two objects do not make a median. Replacement decisions are easier to evaluate when the baseline includes real project hours, which is why a team can use this page as a practical reference before changing its equipment policy.

Every row that has both figures shows the same direction: the things still in use are older than the working lives of the things that went. The gap is between one and two years in every case. For independent context beyond these eleven workplaces, review material from Laptop Mag.

What makes the gap

Three things, and they stack.

Survivorship. An object that has lasted six years has demonstrated something about itself and its circumstances. The stopped group contains every object that failed early; the in-use group contains none of them, by definition. This is most of the gap and the next page is about it.

Censoring. Every object in the in-use column is still accumulating age. Its six years is a minimum, not a final figure. The stopped objects are finished and their four years is complete.

So the comparison is not like for like at all: one column holds partial lives and the other holds whole ones. The honest version of the headline is that unfinished lives already exceed finished ones by two years, which is a stronger statement than it first appears.

Composition. The in-use group is weighted towards the kinds that last. Desks, storage and seating are fifty-one of the hundred and forty-nine in use, and three of the four longest-lived kinds.

Those three kinds, stripped out

Strip desks, storage and seating out of both columns:

in use, less those three        median 5 years   (n=98)
stopped, less those three       median 4 years   (n=113)

The gap narrows from two years to one, and both figures drop.

So about half the headline difference is composition — the in-use group simply contains more of the durable kinds — and about half is survivorship and censoring acting on comparable objects.

That is the most careful thing this section can say about the six and the four, and it is less dramatic than the headline. A reader who takes away things in use are much older has taken away a result that is half an artefact of what kinds of object happen to be in which column.

The cables

Six years for cables in use, two for cables that stopped, and the widest gap of any row.

Cables are the record's most numerous category, its largest share of breakages, and its largest share of the cupboard. A cable that is going to fail fails early, and one that survives its first couple of years appears to go on indefinitely: the oldest cable in use is eleven years old.

That is the shape of an early-failure pattern, and the record can gesture at it without measuring it, because measuring it needs stopping dates for the failed cables and only four of the six broken ones have an estimate.

What the four-year figure actually rests on

Recalled durations for a hundred and nine of the hundred and thirty-five stopped objects. Twenty-six have none.

Nineteen of those twenty-six come from the group that stopped being reached for, where nobody can place the moment. If those nineteen ran longer than four years, the true median is higher.

The record reports four and flags it wherever it appears. It is the only number on this site built from recollection rather than from a recorded field, and it is the one most likely to be wrong.

The two rows with no right-hand figure

Desks have no median for stopped objects because only two of the twenty-four ever stopped, and two objects do not make a median.

Printing and scanning has the opposite problem: six of its eight stopped and only two remain, so the left-hand figure of seven years rests on two objects and is printed with that said.

Those are the two kinds at the extremes of the record — the one nothing ends and the one a single cause ended almost entirely — and neither can contribute a comparison. Everything the table is used for comes from the eight rows in between.

What the gap looks like per kind

kind                      in use   stopped   gap
cables and adapters          6        2        4
lamps and lighting           6        5        1
storage and organisers       8        6        2
seating                      7        5        2
screens and stands           5        4        1
input devices                5        4        1
other                        5        4        1
audio                        4        3        1

Seven of the eight rows show a gap of one or two years. One shows four.

Cables are the outlier and the mechanism is visible elsewhere in the record: a cable that fails, fails early — six breakages, and two of the five shortest working lives in the inventory are cables at six and eight months — while a cable that survives goes on indefinitely, the oldest in use being eleven years old.

That is the shape of an early-failure pattern, and the record can point at it without measuring it, because measuring it would need stopping dates for the thirty-four cables that stopped and only some of them have even a recalled year.

What does not follow

That older equipment is better equipment, or that anything in the in-use column was a better purchase than anything in the stopped column.

Nothing here evaluates an object. A desk outlasts a cable because a desk is a surface and a cable is a flexing wire, and that comparison tells a reader nothing they did not know.

What the figures add is the size of the effect in eleven real workplaces, and the fact that half of it dissolves once you ask which kinds of object are being compared.