A skylight cleaning robot covers 300-600 m2 per hour of glass against roughly 60-120 m2 per hour for a ladder crew, and removes the fall exposure that causes most atrium cleaning injuries. The trade-off is a USD 22,000-34,000 upfront cost that a manual crew never carries.
Ladders still make sense on some roofs. Pretending otherwise loses you credibility with any facilities manager who has run a building for twenty years. So here is the honest split.
How much faster is a robot on real glass?
Speed depends far more on roof layout than on the machine. A clean rectangular bay with few frame breaks runs fast. A patchwork of small panels with mullions every 1.5 m slows any robot to a crawl.
| Metric | Ladder crew | Robot (mid range) | Note |
|---|---|---|---|
| Area per hour | 60-120 m2 | 300-600 m2 | Robots lose time at edges |
| Crew size | 2-3 people | 1 operator | Operator stays below |
| Cycle for 1,500 m2 atrium | 10-22 h | 3-5 h | Plus setup |
| Fall exposure | Constant | None above deck | Biggest difference |
What is the safety case that actually matters?
Every serious buyer has one incident in the memory of the team – a slip, a dropped pole, a near miss. Ladder and rope work on skylights is unforgiving because glass does not forgive a mistake the way a concrete floor does.
A robot keeps the human below the glass line. The operator handles the tether, the water, and the remote. That single change is often worth the price on its own, before you count a single square metre of cleaned glass.
Where do ladder crews still win?
Small jobs, odd corners, and buildings that will never buy hardware. A 400 m2 roof with five different levels and a dozen obstructions is quicker to send two people up for half a day. Robots need a run-in and a run-out; below about 300 m2 a cycle, the setup eats the gain.
Fragile or historic glass also favours a careful human, at least until a supplier proves a low-pressure setting on that exact glazing. Never assume the demo panel matches your roof.
What does a typical week look like for each option?
The ladder or rope crew arrives, sets up, cleans for a few hours, packs down, and the atrium is out of service for most of a day. The robot shows up with one operator, runs a quieter cycle, and can work in sections around trading hours. For a mall, that difference in disruption is often worth more than the direct cost saving.
Disruption has a price even when nobody invoices for it. Closed atriums, blocked escalators, and unhappy tenants all carry a cost that rarely appears in the cleaning budget but is felt by the operations team.
When does the robot lose on pure speed?
On roofs with heavy obstructions and short runs. If the machine spends more time repositioning than cleaning, a nimble human with a pole can win. This is common on historic buildings with small, irregular glass panels.
The honest test is a side-by-side on one representative bay. Time both, count the setup, and compare the finished quality under the same light. That single comparison settles more arguments than a year of datasheets.
How do insurance and liability shift?
A rope or ladder incident on a glass roof can halt a building for weeks and draw a serious claim. Insurers know this, and some now look more kindly on a site that keeps people below the glass line. The premium saving is not huge, but the risk reduction is real, and it shows up in how quickly a claim-free year adds up.
The robot itself becomes an asset to insure, so factor that in too. It is a smaller, more predictable risk than a suspended human, and that usually means a friendlier conversation at renewal.
What does the labour math look like in 2026?
A two-person crew costs USD 700-1,400 per day fully loaded in most Western markets. At four cycles a year on a 1,500 m2 atrium, that is roughly USD 12,000-24,000 annually. A robot with an internal operator and 2,500 of consumables undercuts that from year two onward.
Ask Skylight Cleaning Robot to model your atrium against a two-person crew – the crossover month is the only number that matters.
Key Takeaways
- Robots clean 300-600 m2/h; ladder crews manage 60-120 m2/h.
- The real win is removing fall exposure above the glass line.
- Below about 300 m2 per cycle, setup can erase the speed gain.
- Two-person crews run 700-1,400 per day loaded in 2026.
- Fragile or historic glass still favours a human until proven otherwise.
What about the noise and access angle?
A robot avoids the scissor lift and the road closure that a lift-based crew needs. On a city centre atrium, the cost of permits and traffic management can dwarf the cleaning itself. A machine running above the deck needs no street closure, which is a line item buyers only notice after the first invoice.
Noise is lower too, though not silent. Most machines sit in a range that neighbours accept during working hours and that malls tolerate at night. The ladder crew still needs a quiet-hours conversation with the landlord; the robot usually does not.
How do you measure whether it actually worked?
Photograph the same three bays before and after, from the same position, at the same time of day. Reflections make glass photographed at different angles look different even when the cleaning is identical. A fixed camera spot removes that argument entirely.