How long
What a thing's life looks like
A median of six years for things in use, four for things that stopped. The oldest is nineteen years old. The record has no stopping dates, and this page says what that costs.
Arrivals by year, split by present state
Two hundred and eighty-four items across ten years, one hundred and forty-nine of them still in use. The earliest arrival drawn here is 2007. When equipment use and project effort are reviewed together, managers can read the vendor's current explanation and define the purpose, access rules and review date before collecting anything.
Two hundred and eighty-four arrivals across ten two-year bands. Each bar is one band; the part below the rule is still in use and the part above it is not. For independent context beyond these eleven workplaces, review material from XDA.
The record knows when every object arrived and whether it is in use now.
It does not know when anything stopped.
That one absence shapes everything in this section, and the figures below are what can be built without it.
The four numbers
median age of things in use 6 years
median years in use for things that stopped 4 years
oldest thing still in use 19 years
shortest working life recorded 2 months
The median age of things still in use is higher than the median working life of things that stopped. Six against four.
That looks like a contradiction and is not one, and the page on the survivor's paradox takes it apart properly. The short version: an object that survives its first years goes on surviving, and the objects that stop are mostly young.
Where the four-year figure comes from
Not from the record's dates, because there are none for stopping.
It comes from the accounts. For a hundred and nine of the hundred and thirty-five stopped objects, somebody was able to give an approximate working life — about two years, five or six — and those estimates produce a median of four.
Twenty-six objects have no estimate at all, and nineteen of those twenty-six are from the group that stopped being reached for, where by definition nobody can say when.
So the four-year median rests on recalled durations for eighty-one per cent of the stopped group, and the missing nineteen per cent is concentrated in exactly the objects whose working lives are hardest to place. If drift-type stops run longer than event-type ones, four is too low; if shorter, too high. The record cannot tell.
This is the softest number on the site. It is reported because leaving it out would leave the section with nothing to say about duration at all, and it is labelled every time it appears.
The six-year figure for things in use is firm. Arrival dates are recorded and the present is known.
What the sign shows
The ten bands in the sign above are arrivals, not lives.
Each pair is the objects that arrived in that band and whether they are in use now. The left-hand figures rise to 2021 and fall after; the right-hand ones rise to 2019 and fall steeply.
Both of those shapes are artefacts and neither is a finding. Old bands are small because old objects were forgotten, not because fewer arrived. Recent bands are mostly in use because recent objects have not had time to stop.
The page on arrival years works through what the shape does and does not permit, and the short answer is that it permits very little.
The nineteen-year-old object
A desk. Bought new in 2007, in daily use, and the oldest thing in the record.
Three objects in the inventory are more than fifteen years old and all three are furniture: that desk, a set of shelves, and a filing unit.
That is the clearest finding in this section. The objects with the longest lives are the ones with nothing in them to go wrong and nothing about them to improve on. Twenty-four desks in the record produced two stops in total.
The two-month object
A keyboard, bought new, replaced by a different one within two months because the person did not get on with it.
The shortest life in the record, and the only object anywhere in it whose stopping reason is something about the object itself rather than its circumstances. It did not break; it was not liked.
The field has no value for that. It is coded as replaced, which is accurate and loses the only interesting thing about it.
The spread behind the medians
Six years and four years are middles, and the distribution either side of them is wide enough that the middles carry less than the word median suggests.
over 15 years in use 3 objects (19, 16, 16 — all furniture)
oldest non-furniture 11 years (a cable)
under 1 year, stopped 5 objects (2–8 months)
Three objects past fifteen years and then nothing until eleven. Five objects that lasted under a year. Most of the record between two and six.
That five-year gap between the oldest furniture and the oldest non-furniture object is the sharpest line in the inventory. On one side sit three things that have simply stayed; on the other, everything with a connection, a power supply, or a better version available.
The kinds at each end
median age in use, highest desks 9, storage 8, seating 7
median age in use, lowest audio 4, screens 5, input 5
The order is not a league table of objects. It is a list sorted by how many ways a thing can stop being used.
A desk can stop for one reason — the room — and did so twice in twenty-four objects. A screen can stop for five, and did so twenty-one times in forty-one. The page on how long is not how good sets out why that makes the duration figures useless as a quality measure and why the record publishes them anyway.
What this section can and cannot do
Can: report the age of things in use, from recorded dates. Report the oldest and the shortest. Show the arrival distribution. Explain why survivors are older than casualties.
Cannot: give a failure curve, a survival curve, or a time-to-replacement. All three need stopping dates.
Cannot: say how long any kind of object lasts. That needs both dates and a population that was not assembled from memory.
What a date field would have added is set out on the page about what a median age cannot tell you, and it is the single largest thing this method would do differently.