Short answer: on a clean, flat, same-size panel run, a skylight cleaning machine wins on time from roughly the fourth panel onward. On site, the crossover is decided by setup, not cleaning. Our 2026 time study measured a trampoline-setup suction robot at about 3 to 5 minutes per panel once running, against 7 to 11 minutes per panel for a two-person manual crew with pole and squeegee. But the robot spent 25 to 50 minutes on setup, staging and suction checks, and the manual crew spent about 10. So a 5-panel roof is usually faster by hand. A 60-panel roof is not close.
This page reports what we measured and what we did not, because the honest version of this comparison is more useful than the marketing version. Every minute below comes from a stopwatch on a real roof, and the assumptions are stated so you can reject them if your roof is different.
How many minutes per panel does each method actually take?
The manual crew ran 7 to 11 minutes per panel depending on how much of the load was hardened dust versus film. The spread is wide because manual work slows down on the panels where the pole cannot reach the top edge without repositioning, and because two people on poles interfere with each other on narrow runs.
The machine ran 3 to 5 minutes per panel on the same roof, with the low end on panels that needed one wet pass and the high end on panels with dried bird residue, which needed a pre-soak and a second pass. That number excludes setup and excludes any panel the machine could not reach, which on this roof was 6 of 74 panels along the ridge and the end walls.
The unreachable panels matter in the arithmetic. If you add them back at manual rates, the blended machine figure rose to about 4.7 minutes per panel on this site. That is still a large gap, but it is the number you should use for planning, not the 3-minute best case.
What does setup overhead really add?
Setup is where most comparisons quietly cheat. The manual crew arrived, carried poles and buckets to the roof access point, filled water, and started. Ten minutes, and most of it was walking.
The machine required: staging at the access hatch, assembling the safety line anchors, checking suction margin on a test panel, filling the tank, running the power lead to the nearest safe outlet, and a short calibration pass. Median 34 minutes across the visits, with a long tail when there was no convenient outlet and the lead run went across a fire route.
Teardown added another 8 to 15 minutes for the machine against about 5 for the crew. So the true fixed cost of a machine visit is roughly 45 minutes, and the crew’s is roughly 15. At 2 minutes saved per panel net of the blended rate, you need about 15 panels to break even on a single visit. Below that, manual wins on time.
Which roof conditions flip the answer?
Panel count is the obvious one. Anything under about 15 panels in a single continuous run rarely justifies the machine on labour time alone, unless the access method is the real cost. On a roof where a person would need rope access, the machine’s advantage is not measured in minutes saved per panel but in the hours of permit and rigging work avoided entirely.
Panel pitch is the second. On steep runs, manual crews slow down dramatically because they are working against gravity and repositioning constantly. Machine times hold closer to constant. On nearly flat podium skylights the gap narrows, because manual work is easy and there is no fall exposure to design around.
Dust load is the third, and it is the one buyers underestimate. A machine is efficient on light film and much less so on hardened residue, where the manual approach of a long soak followed by targeted scrubbing is hard to beat. If your roof collects industrial fallout, budget for pre-soaking either way, and expect the machine’s advantage to shrink to a minute or two per panel.
The site’s access profile decides more than the cleaning. Our notes on high-rise facade and skylight access break down how rigging and permit time compares with machine staging on tall buildings, which is usually the dominant number on anything above four storeys.
How should the comparison be measured on your own roof?
Run a one-day trial on a single section. Time three things separately: fixed setup, per-panel cleaning, and non-cleaning movement between panels. Most comparisons fail because these are lumped together into one number that nobody can audit.
Time a section with the manual crew on a normal day, not a rainy one, and accept the real panel count rather than the ideal one. Then time the machine on the same section within the same week, so the dust history matches. If the machine wins on the blended number and your section is at least 20 panels, the arithmetic will hold across the roof.
Write down the panels the machine missed and time them at manual rates. Include the teardown. Include the walk to the outlet. The single most common error in these comparisons is quoting the per-panel cleaning rate as if it were the job rate, then being surprised in the annual review.
If you are weighing this against a fully manual rope team, the safety-side comparison is worth reading next to the time comparison: our analysis of crossing obstacles and edges on facades covers where the machine must stop and a person must take over, which is where the time saving quietly leaks away.
Key Takeaways
- Measured 2026 figures: machine 3-5 minutes per panel running, crew 7-11 minutes per panel. Blended for unreachable panels, the machine was about 4.7.
- Fixed machine cost is roughly 45 minutes per visit including teardown; the crew’s is roughly 15. That gap, not the per-panel rate, sets the crossover.
- Below about 15 panels in one run, manual usually wins on time. Above 20, the machine usually wins clearly.
- Heavy hardened dust narrows the gap. Plan a pre-soak whichever method you use.
- Time setup, per-panel work and inter-panel movement separately on your own roof before committing to an annual plan.