A supervisor covering eighty rooms can walk into a handful. Random sampling finds the rooms you happened to visit — not the bathroom that was skipped on the sixth floor.
Room inspection is the quality system of the housekeeping department, and in most properties it works like this: a supervisor with eighty rooms on their floor plan gets through twelve to twenty of them, prioritises VIP arrivals and new starters, and covers the rest with experience and hope.
Nobody thinks this is ideal. It persists because the alternative — inspecting everything — does not exist at any staffing level a hotel can pay for. So the question is not how to inspect more rooms. It is how to choose which rooms to inspect.
Suppose one room in twenty has a real defect — a phase skipped, a bathroom rushed, a bin missed. Inspect fifteen rooms out of eighty at random and you will, on average, walk into one of them. You will also walk into fourteen rooms that were fine, and you will miss three defects entirely.
And random is generous, because real spot checks are not random:
The result is a quality system whose coverage is decided by geography and habit, and whose failures are discovered by guests.
"A spot check tells you about the room you walked into. It tells you nothing about the seventy-nine you didn't."
AiCarpus · Hotel IQTargeted inspection needs something to target. Hotel IQ produces it as a by-product of how it works: a Bluetooth beacon identifies each room, motion sensors recognise the cleaning activity, and the system groups what it detects into the six phases of a turnover — pre-spray, sheets, bathroom, vacuum, dusting, windows.
Which means that at any point in the shift, there is an answer to the question inspection exists to ask: did each room get each phase?
Illustrative example — the shape of the output, not data from a specific property.
The matrix is not a verdict on cleanliness. It is a map of where a human judgement is worth spending. Two flagged rooms out of eighty is a supervisor walk of ten minutes instead of a morning.
A cleaning phase is missing. The strongest signal there is — the bathroom was never worked, or the vacuum pass never happened. This is the room a guest would have found.
The turnover is far below the measured median for that room segment. Not a nine-minute stayover, but a nine-minute departure in a room type that reliably takes half an hour.
Present, but little cleaning detected. Time in the room without the movement pattern that goes with work — often an interruption, sometimes a room that was opened and abandoned.
The last rooms of a long shift. Where fatigue drift shows up in the effort metrics, defects follow. Inspecting the final two rooms of a heavy board beats inspecting the first two.
A new starter still learning the routine. Worth targeting — but with the data showing which phase they are consistently rushing, so the check turns into training rather than a verdict.
One warning, because it is the mistake that quietly breaks risk-based systems everywhere: if you only ever inspect flagged rooms, you stop being able to see what the flags miss.
Keep a small random control sample — a handful of unflagged rooms every day. Its job is not to find defects; it is to measure whether the flags are working. When a control-sample room fails, you have learned something more valuable than a dirty bathroom: you have learned that a category of defect is invisible to your signals, and you can go and fix the signal.
Motion detection sees whether a phase happened. It does not see a hair in the sink, a fingerprint on the mirror, or the smell of the room. Those remain human judgements, and always will. The point of targeting is to buy the time to make them properly.
"The aim isn't fewer inspections. It's inspections that find something — and a supervisor who spends the rest of the hour teaching rather than hunting."
AiCarpus · Hotel IQThe role shifts in a way most supervisors welcome once they see it. Detection — walking floors to find out what happened — is the least skilled part of their job and the part that eats the day. Handing that to the system leaves the parts that need a person: coaching a new starter through a bathroom routine, checking standards rather than compliance, dealing with the rooms and guests that actually need judgement.
There is a fairness effect too. When inspection is targeted by evidence rather than by habit, the experienced attendant stops being ignored and the new starter stops feeling singled out. Everybody gets checked when their rooms give a reason.
Targeting has to stay pointed at rooms. Hotel IQ produces no individual performance ranking, no leaderboard and no comparison between named people — it flags rooms, and a room belongs to a shift, not to a verdict about a person. No cameras, no microphones, no personal location outside cleaning zones, and a GDPR mode on every report.
Used the other way — as a hunt for the slowest attendant — this stops working within a week, because people optimise for the signal instead of the room.
Hotel IQ Watch is the edition this article is really about: phase-level verification per room is what makes a coverage matrix possible.
Hotel IQ Mobile still gives you a usable first signal with no hardware at all — which rooms were serviced, how long each took, and where movement does not match cleaning. That alone is enough to stop inspecting at random.
See the phase detection: Explore Hotel IQ →
See a full shift report: Open the sample report →
Hotel IQ confirms each cleaning phase per room and flags the ones that need a human eye — so inspection stops being a sample and starts being a decision. Book a demo and see it run in your property.