Direct answer. Size a window cleaning robot fleet by facade, not by floor area. A robot works one pane, or one continuous glass run, at a time, so the planning unit is the number of independent glass surfaces one operator can reach, start and move on from inside a shift. One unit covers a villa, a shopfront or a single sunroom. A tower with independent curtain-wall sections, or a daily route across several buildings, is normally planned as two or more units with one supervisor — so a stopped machine does not stop the day. Buy one unit, time a real shift on your own glass, then scale from what you measured.
Why square metres cannot answer this question
Two buildings with identical glass area can need completely different numbers of machines, because the robot does not sweep a surface. It adheres to one pane, runs a planned route across that pane, and then has to be physically moved to the next one.
Every move costs repositioning time, and every repositioning needs an anchor point for the rope, somewhere safe to set the machine down, and a decision from the operator about whether the next pane is safe to start. A tower where each floor is one long ribbon of glass is a small number of long runs. An older building with divided panes and proud mullions is a large number of short, awkward ones. Same square metres, very different shift.
Five steps that give you a defensible number
1. Count independent glass planes, not floors. Group them by what you can reach and start from one position: one anchor, one power point, one setup.
2. Sort the groups by accessibility. Continuous curtain-wall runs are one class. Standard framed panes are another. Awkward geometry — A-frame, triangular, narrow gaps — is a third. Internal glass such as sunrooms, atrium panels and shopfronts is a fourth.
3. Time one shift on one plane from each class. The published figure for a large panel is around 20 minutes. Use it as the expectation and then measure on your own glass. What you measure is what you plan with.
4. Add repositioning, power and water. This is the step most plans skip, and it is where an efficient-looking schedule falls over. Find out where power and water come from on each plane, and how long the move, anchor and restart actually takes.
5. Decide your redundancy before you decide your total. For a route that has to be finished, size so that one stopped machine does not stop the schedule. Two units and one supervisor, rather than one unit and a hope.
Matching a fleet shape to the job
| Situation | Fleet shape | Why |
|---|---|---|
| Villa, shopfront, single sunroom | 1 unit | One operator, few runs, no schedule pressure |
| Hotel or office building, standard framed glass | 1 unit first, then measure | Reposition-heavy: the constraint is setups, not cleaning time |
| Tower with independent curtain-wall sections | 2+ units, 1 supervisor | Sections run in parallel, and one stalled unit must not stop the route |
| Daily route across several buildings | 2+ units, 1 supervisor | Travel and setup dominate; a spare machine protects the route |
| Awkward geometry only — A-frame, triangular, narrow gaps | 1 unit with the right head | This is a geometry problem, not a capacity problem |
The last row is the one buyers get wrong. Not every fleet problem is solved by adding capacity. If your obstacles are panes a flat machine physically cannot reach, the answer is a different body — the G8 slim kit pairs a 295 × 145 × 72 mm body with a triangular squeegee head and a telescopic pole — not a second identical unit. Adding units to a geometry problem multiplies the failure rather than the throughput.
The two numbers that set your price
Pane count and cycle time price a job. Floor area does not. That is why the handover document is a commercial instrument, not just an operational one. A usable handover records:
- the panes or zones covered;
- roughly how long each takes;
- where power and water come from;
- which areas are excluded, and why;
- what happens in bad weather;
- who gets called when a unit stops mid-panel.
Photograph the glass before the first cycle and keep the images. “Clean” is an argument you will have later, and those photographs plus the pane count are the evidence. They are also the fastest way to quote the next building, because the two numbers you need are already measured rather than estimated.
Sizing a mixed fleet
Portfolios rarely need one model. A villa-heavy channel, a shopfront route and a curtain-wall project in the same market usually want different bodies — a square ultra-thin unit for tight framing, a 3,000 Pa entry model for low-rise domestic work, a 6,000 Pa commercial unit for tower sections.
That is easier to buy than it sounds. Mixed-model orders are supported for qualified wholesale and OEM buyers, so the portfolio can be built without meeting a separate minimum on every sku. The order structure — whether the minimum applies per model or per order, and whether bands apply across the whole mix — is worth settling before you decide the fleet, because it changes what a sensible first order looks like. The wholesale and OEM supply page sets out how those bands work.
When we would tell you not to buy the second unit yet
- You have not timed a real shift. Sizing a fleet from a data sheet is how capital ends up frozen in the wrong number of machines. Buy one, run it for a week, measure, then scale.
- The glass is cleaned once or twice a year. A robot earns its place on frequency. If a facade is cleaned annually, a rope-access team is the honest answer and the robot is the wrong purchase.
- The site requirements are still open. Access to power, a controlled drop zone, water for the spray tank and a written safety line are prerequisites. Adding machines to a site that has not settled those questions multiplies risk, not output.
FAQ
How many commercial window cleaning robots does a tower need?
Plan by independent curtain-wall sections rather than by floors. One unit handles a villa, a shopfront or a single section under supervision. Two or more units with one supervisor is the normal shape once sections can run in parallel or the route carries a completion commitment.
What is the real bottleneck — battery, water or cleaning time?
On most commercial routes it is repositioning, power and water between runs, not the cleaning cycle. Measure setup and teardown per plane and the schedule usually explains itself.
Should the spare machine be the same model as the working machine?
Usually yes, if the constraint is schedule. Pads, nozzles, seals and battery packs are then shared across the fleet, which halves the consumable inventory — a point we cover in what a distributor should stock. If the constraint is geometry instead, the spare should be the model that covers the panes the first one cannot reach.
Can one operator run more than one unit at a time?
On a single continuous glass run, no — the machine needs moving and starting. On a tower with independent sections, yes, provided someone is still controlling the zone below and the safety line has an owner. Unattended operation means the operator can do other work in the building; it does not mean the building can be empty.
Tell us your building types and the shape of your route, and we will confirm a fleet size and a mixed-model order structure. Contact us →