For windows too high to reach, four methods are honest in 2026: water-fed pole, rope access, cleaning robot and building maintenance unit. Cost per pane runs from about $1.80 to $9, and the right one depends on height, glass geometry and cycle frequency.
The question sounds simple and the answer is a decision tree. Most of the cost of cleaning high windows is not cleaning. It is reaching them safely, legally and repeatedly.
Here is how each option behaves in 2026, with the numbers that actually differ, and the two cases where the sensible answer is to keep paying a crew.
How do you clean windows that are too high to reach?
Start with height. Up to two storeys a water-fed pole is the cheapest correct answer and needs no certification in most jurisdictions. From three to eight storeys, rope access crews or a pole from a platform become usual.
Above eight storeys you are choosing between rope access, a BMU and a facade robot. Above 20 storeys robot or BMU takes over for regular cycles, because rope crew time per square metre climbs steeply with height, not glass area.
Skylights break the pattern. You are not reaching sideways, you are reaching up and over, and poles stop being viable above about 6 m because control at the top of a wet pole is guesswork.
| Method | Practical height | Cost per pane | Best for |
|---|---|---|---|
| Water-fed pole | Up to 12 m | $1.80–3.50 | Low-rise, frequent cycles |
| Rope access | Any | $2.50–9 | Complex facades, irregular cycles |
| Cleaning robot | Any with an anchor | $0.45–0.90/m² | Regular cycles, glass roofs |
| BMU | 15+ storeys | Amortised capital | Fixed facade, new build |
What do the four methods cost in 2026?
Water-fed pole, $1.80–$3.50 per pane on a 12 pane facade with ground access. Rope access, $4–$9 per pane at low frequency, falling to $2.50–$4 at monthly cycles. Robot, $0.45–$0.90 per m² amortised over five years. A BMU costs $80,000–$250,000 plus annual certification and only makes sense above 15 storeys on a fixed facade.
Frequency moves these numbers more than height does. Monthly cleaning justifies a robot in 24–30 months. Twice-yearly cleaning usually does not, because the amortised cost lands above rope access.
What are the safety rules that catch people out?
Rope access requires certified technicians in most markets, plus anchors with documented ratings and a written rescue plan. Small facility teams discover this after commissioning a quote, which delays the job by weeks.
Robots remove the person from height, which is the biggest safety gain available, but they add anchor and tether duty. A roof anchor rated for fall arrest is not automatically rated for a robot pulling 30 kg downslope, and that distinction is missed constantly.
For skylights, glass loading is the quiet risk. A technician standing on a skylight is a collapse hazard because most glass roofs are not rated for foot traffic. Robots avoid that entirely, which is underrated in buying decisions.
When is a robot the wrong answer?
Heavy geometry. Fins, sunshades, deep reveals and irregular curtain wall patterns cut robot throughput until the labour saving disappears.
Low frequency. A once-a-year clean has a higher amortised cost with a robot than with rope access, and a machine idle for 11 months sits out of warranty before it sits out of usefulness.
Very small glass area. Below 150 m², the capital outlay rarely pays back inside a reasonable horizon unless access equipment hire is expensive.
How often should high windows be cleaned, realistically?
Buildings that host clients clean 4 to 12 times a year. Industrial and residential towers run 2 to 4. Below annual, glass biofilms start bonding and the first clean after a long gap costs 30–50% more in labour because the soil has hardened.
Skylights are worse because they sit flat and collect everything. A north-pitch skylight in a wet climate grows visible algae inside two seasons, and algae needs a different chemical mix from dust.
What does the contract actually need to specify?
Scope by surface, not by hours. State the square metres, the number of cycles per year, the chemical pH range, and who supplies water. Contracts written in hours encourage speed, and speed on glass shows.
Add a before-and-after photo requirement for the first three cycles. It resolves most disputes early and it gives the supplier a reason to bring a better brush.
Does a water-fed pole work above 8 metres?
Partly. Reach is available to 20 m with carbon poles, but water weight, wind and control degrade badly past 12 m. Above that you spend more time positioning than cleaning, and results at the pole tip are visibly worse than at eye level.
What are the risks to plan for?
Two failures repeat on real sites, and both are cheap to prevent once someone names them in the contract.
- A fall-arrest anchor is not a robot anchor, and vendors rarely say so.
- Skylight panels can fail under a person even when they look solid from below.
Key Takeaways
- Pick by height first, geometry second, frequency third.
- Above 12 m a water-fed pole stops being a control problem and becomes a quality problem.
- Glass roofs are usually not rated for foot traffic, so robots remove a real hazard.
- Below 150 m² of glass, robots rarely pay back.
Before you commit to a model, read our Lingkong K3 review and all skylight robot reviews. Both are written for buyers running real rooftops rather than spec-sheet readers.