Framework
Do robot vacuums actually clean the whole floor?
Usually not fully, and the apps obscure it. Coverage splits into percent mapped, percent driven, and percent actually completed, which often disagree. A robot can map everything and drive most of it while leaving edges, corners, and embedded soil untouched, yet still report a flattering figure.
Open the app after a run and a robot will tell you it covered your home. The word does a lot of quiet work. Coverage sounds like a single, verifiable fact, the way a thermometer reports a temperature. It is not. At least three different measurements hide inside it, and a robot can score well on one while leaving the floor that matters untouched.
Pulling the three apart is a small framework that survives across categories, from vacuums to mowers to floor-scrubbing arms. Once you see the seams, the coverage map stops being a victory lap and starts being a diagnostic.
Term
CoverageAn umbrella word that conflates three separate claims: the area a robot built a map of, the area it physically passed over, and the area it actually completed to standard. Apps usually report the most flattering of the three.The first promisePercent mapped: what the robot knows exists
Mapping coverage is how much of the space the robot has built an internal model of. It is the precondition for everything else and the easiest to get right, because building a map does not require doing any work in the mapped area. A robot can map a room perfectly and never clean it well. High mapping coverage tells you the robot understands the floor plan. It says nothing about results.
Figure
Three claims that sound like one
- Floor areawhat the room measures, and the number a spec sheet quotes
- Reachableminus what the chassis cannot get under, and the strip it leaves at every wall
- Sweptminus what the brush cannot reach into. This is the figure that matches the floor afterwards
The second promisePercent driven: where the wheels went
Driven coverage is the path the robot actually traced, the line the app draws over your map. This is the number most coverage displays are really showing, and it is genuinely useful: it reveals the rooms skipped behind a closed door, the zone abandoned when the battery ran low, the corner the robot decided was unreachable. But driving over a patch is not the same as cleaning it. A vacuum passing once over embedded grit, or a mop gliding over a dried spill, has driven the area without finishing it.
The three measurements
Same word, three different claims
| Promise | What it measures | What it hides |
|---|---|---|
| Percent mapped | Area the robot modelled | Whether any work happened there |
| Percent driven | Area the wheels crossed | Whether the pass was effective |
| Percent completed | Area finished to standard | Hardest to measure; rarely reported |
The third promisePercent completed: what actually got done
Completion coverage is the one that matters and the one almost no robot reports, because it is the hardest to measure. It asks whether the area was finished to standard: the grit lifted, the spill removed, the grass cut to height, the edge actually reached. A robot has no clean way to verify its own results. It knows where it drove. It does not reliably know whether driving there worked. So the app reports driven coverage and lets you infer completion, and the inference is where expectations break.
Hidden inside one number
3claims
Mapped, driven, and completed are different measurements. The app usually shows the most generous one and labels it coverage.
The gap between driven and completed is the structural limit of current consumer robots, and it should be named. These machines are good at being thorough about going everywhere and weak at confirming that going everywhere accomplished anything. Edges, corners, transitions, and embedded soil are where driven and completed diverge most. A robot that reports 98 percent coverage has almost always measured the path, not the outcome.
For assessment, this is why we resist a single coverage figure. The useful questions are which of the three a number represents, and how wide the gap is between where a robot went and what it finished. A high driven number with a known completion weakness is a more complete picture than one confident percentage that quietly means only the easy thing.
A robot can map everything, drive over most of it, and finish less than you assume. Coverage is the word that hides the difference.
