Most buildings maintain their atrium glazing reactively, and almost nobody decided to. It happens by default, because the roof is out of sight, nothing is obviously wrong with it, and there’s always something on the register with a date against it that the glass hasn’t got.
So the glass waits until it leaks. Then it gets attention, at the worst possible moment, on someone else’s timetable.
What reactive actually costs
The repair itself is usually the smaller half.
An emergency call-out is priced for emergency. Out-of-hours labour, equipment mobilised at short notice rather than booked, and a contractor who has to drop something else to come. The same work done in a planned slot in March, at a time the building chose, is a different number.
Then there’s everything the leak did on the way to being noticed. Water in an atrium runs horizontally inside framework before it drops, so by the time it shows on a ceiling it has been in the structure for a while. Wet insulation stops working. Steel starts corroding. Plasterboard, paint and floor finishes need making good. None of that is glazing cost and all of it is on the same invoice trail.
The disruption is the part that genuinely hurts in a trading building. A leak over a retail floor or a hotel lobby means closed areas, moved furniture, barriers in a space designed to look open, and in some buildings lost revenue that dwarfs the repair.
And reactive work is nearly always a patch. Someone is there to stop water coming in today, not to understand why it did, so the fix addresses the symptom and the underlying defect carries on. Buildings get into a cycle of three or four repairs on the same roof before anyone asks a wider question.
What planned maintenance is actually buying
Not the absence of defects. Defects happen regardless. What it buys is finding them while they’re still cheap and dealing with them at a time of your choosing.
Perimeter seals degrade predictably. Gaskets harden and shrink back. Outlets silt up. Fixings loosen. Sealed units lose their edge seals, and the ones on a given roof fail in a cluster because they were all made at the same time. Every one of those is visible at close access long before it produces water indoors.
Catching them early changes the instrument you need. A perimeter joint cut out and remade before it splits is a maintenance item. The same joint left to fail becomes a leak investigation, a repair, corroded steelwork to deal with and interior making good.
There’s a budgeting argument too, and it’s the one that usually carries a board. Planned work can be forecast, phased and spread. Reactive work arrives as an unbudgeted request with a note saying it has to happen this week.
What a programme looks like in practice
Twice-yearly cleaning suits most urban commercial atria, with an inspection attached to each visit. Quarterly for sites near heavy traffic, a flight path or industrial plant. Annual is sometimes defensible on a quiet rural site.
The habit that makes it worth the money is combining the clean with the survey. Someone is already at height next to glass nobody else sees properly from one year to the next, and a quarter of an hour more gets you seal condition, gasket condition, fixing condition, and whether the outlets are clear.
Recording it is what turns a visit into a programme. One inspection tells you the state of the roof. Five tell you the rate it’s changing at, and the rate is what lets you forecast. A misted unit count that goes from two to five to eleven over three years is a replacement programme announcing itself with plenty of notice.
Smoke vents and automatic opening vents belong in the same schedule, since they’re mechanical, they’re part of the fire strategy, and nobody tests them by looking up.
The Centre for Window and Cladding Technology publishes guidance on inspecting and maintaining glazed envelopes, and the Glass and Glazing Federation is a sensible first check on who’s accredited to carry it out.
The objection, and the answer to it
The push-back is always the same: planned maintenance is money spent on a roof that isn’t leaking.
The answer is that the alternative isn’t zero. It’s the same money spent later, at emergency rates, with the interior damage and the trading disruption added, and without the choice of when. Asset management works the same way everywhere, and the new-against-used calculation on plant purchases runs on the identical logic: whole-life cost, not purchase price.
For a building that has never had a programme, the starting point is one close-access condition survey so the first set of numbers exists. Contractors who hold the access and the glazing together, https://skyaccessglazing.co.uk/commercial-glazing-services/maintenance among them, can do that survey and the first clean as one visit.
Book that survey against a date in the autumn diary rather than adding it to a list of things to consider. A dated item gets done and an undated one waits for a leak.