What an ageing maintenance backlog is telling you
Open jobs sorted by how long ago they were raised is one of the most honest numbers in a venue. Here is how to read the buckets and what each shape of backlog usually means.
Take every maintenance job that is still open and sort it by how long ago it was raised. Three buckets is enough: under a week, one to four weeks, and over a month. The shape that comes out is one of the most honest numbers in an operation, because unlike most reports nobody has to fill it in.
The shapes, and what they usually mean
A backlog concentrated in the first bucket is healthy. Work arrives, gets dealt with, and leaves. There will always be something in there; a venue with nothing raised in the last week is more likely a venue where nobody is reporting than a venue where nothing is wrong.
A backlog with a fat middle bucket is usually a triage problem. Things are being reported and then sitting, which almost always means the reporting is happening in a place that nobody owns. Jobs raised into a WhatsApp group have this shape, because a message is read and then scrolls away.
A backlog with a fat third bucket is different and worse. Anything over a month has usually been decided against without anybody saying so. It is either work nobody intends to do, or work that is genuinely blocked on a quote, a part or a landlord, and those two look identical from the outside.
The tell that something else is wrong
A very large backlog with no history of anything being missed or cancelled is worth a second look. If nothing ever leaves the list except by being completed, you are probably looking at jobs that will never be completed and cannot be closed.
The commonest version of this in a multi-site business is a site that closed. Jobs keep generating against a venue that no longer trades, nobody can complete them, and the backlog grows forever. The number stops meaning anything and people stop looking at it, which is the real cost.
Clearing one
Go through the third bucket first and be honest. Anything you are not going to do gets closed with a reason, not left to age. A backlog you have decided to ignore is worse than a shorter one you believe.
Split the genuinely blocked ones out. Waiting on a part, a quote or a landlord is a real state and it should be visible as one, so those jobs stop reading as neglect.
Give the middle bucket an owner each. Almost everything stuck there is stuck because it belongs to nobody in particular.
Then look at what is left and ask what would have stopped it being raised at all. That list is your planned maintenance schedule.
The last step is the one that matters
Clearing a backlog is a one-off. The same backlog reappears within a quarter unless something changed about how work arrives and who owns it.
Two changes do most of the work. Reporting has to happen somewhere with a queue rather than somewhere with a feed, so that a report is a thing with a state rather than a message that scrolls past. And every job needs one owner, a person or a team, from the moment it is triaged, because a job assigned to everybody is assigned to nobody.
Paddl gives you both: a request goes into a board with a triage column, gets assigned to a person, a team or a contractor, and carries a deadline set from its priority. The ageing report is on the same data, so the backlog you read is the backlog you have rather than a spreadsheet somebody updated last month.