The most common skylight cleaning robot buying mistakes in 2026 are oversizing suction, ignoring roof access, skipping spare parts and never naming an operator. None of them are about price.
After enough installs you see the same failure patterns. Buyers compare spec sheets hard and then forget the boring operational questions. What follows is the list worth pinning to the wall before you sign anything.
Mistake 1: buying for the steepest pitch only
People choose a heavy 30 kg machine because one small section of the roof hits 35 degrees, then run it over the other 90 percent of flat glass. The result is slower cleaning, more cup wear and a bigger lift. Match the machine to the majority of the roof, and treat the steep patch as a special case – often by slowing a lighter unit like the Lingfeng S1 rather than picking a heavy one for everything.
Mistake 2: no access plan
A robot without a safe route to the roof is a paperweight. Check how the machine and its water get to roof level. Stairs, a lift, a hoist? An 22 kg machine plus water is more than one person wants to carry up four flights weekly. Budget a walkway or a hoist if access is awkward, and put it in the project cost from the start.
Mistake 3: forgetting spares and consumables
Cups, brushes and filters wear out. A site that orders them only when something fails loses days each quarter. Keep a small spares box on site from day one. The cost is trivial next to a missed cleaning window in a building that closes access for events.
- Cups: replace every 6-12 months
- Brushes: 3-6 months
- Filters: 2-4 months
- Battery: 500-1,200 cycles
Mistake 4: no named operator
Train two people, name one as owner. Machines with no owner drift, get misused, and end up with flat batteries and clogged filters. Handover should include a signed checklist and a stored route map. See the commissioning checklist for what to insist on.
Mistake 5: ignoring power logistics
Battery swaps or a 16 A supply at roof level – decide before purchase. A tether unit without a suitable socket turns into an extension-cable nightmare. A battery unit without a charging spot at roof level means hauling packs. Both are cheap to fix at design time and expensive later.
Mistake 6: chasing the lowest price
Cheap units often save on cups and sensors, which is exactly where reliability lives. A 20 percent lower price that costs three extra support calls a year is not cheaper. Ask for the suction spec, cup material and warranty terms in writing, then compare like for like.
Mistake 7: no measurement after install
If you never measure m2 cleaned per hour or daylight before and after, you cannot tell whether the purchase worked. Set two or three numbers on day one and review at 90 days. Sites that measure tend to fix route plans and hit payback; sites that do not blame the machine. Read the 2026 payback numbers to set targets.
How do you avoid regret after the purchase?
Slow down before signing. Visit a similar roof if you can, or watch video of the machine on a roof like yours – sloped, framed, dirty. A demo on a clean flat floor tells you nothing about how it handles a raised seal in the rain. Buyers who insist on a real-roof demonstration report far fewer surprises.
Then commit to the boring parts: spares, training, measurement. The machine is rarely the problem. The failures come from a missing cup, an untrained operator, or a route nobody updated. Get the operating discipline right and almost any competent machine will deliver. Get it wrong and even the best one sits idle.
Key Takeaways
- Match the machine to the majority of the roof, not the steepest patch.
- Plan roof access and water before buying.
- Keep cups, brushes, filters and a battery on site from day one.
- Name one operator and train two.
- Measure m2/h and daylight at 90 days to prove the purchase.
Avoid all seven by planning early. Talk to our team before you specify a machine.
How do you compare two machines fairly?
Put them side by side on five numbers: vacuum in Pa, weight in kg, area per hour, runtime per charge, and cup service life. Ignore marketing words. If a supplier cannot give all five, that itself is an answer. Then ask for one reference site of a similar building and call it – not for praise, but for the parts they have replaced and how often.
Watch the units. Area per hour depends on speed and pass width, so a machine that is fast but narrow can lose to a slower wide one. Ask for m2/h at a realistic speed, not a maximum sprint that wears the machine and leaves streaks.
What belongs in the contract?
Four clauses. Handover of one to two days with a signed checklist. A starter spares kit. A stated cup life in months. And a training clause naming two operators. These cost little to ask for and prevent the most common first-year problems. A supplier who resists all four is telling you something about after-sales.
Finally, agree how you will judge success. Set m2 per hour and a daylight or haze target at day 90. A purchase with a target gets reviewed and tuned; a purchase without one drifts until someone blames the machine. Measurement is not bureaucracy, it is how the machine earns its keep.