Housekeeping Inspection: Why 100% Is the Wrong Target

A hundred percent inspection rate is not a quality program. It is a statement about how much walking a supervisor did — and on a property where the villas are minutes apart, it is usually a statement about how much walking she is recorded as having done.

That is not an accusation. It is what an impossible standard does to the people asked to meet it. The standard is not the thing to enforce. What matters is what a housekeeping inspection program looks like once you accept that a supervisor’s day contains a finite number of walks — and that on a villa property, the walk is the expensive part of everything.

An inspection rate is an input

Rooms cleaned, rooms inspected, checklists completed: all counts of activity. None of them is the outcome. The outcome is villas released in a state the arriving guest does not have to phone about.

Inspecting more is one way to move that outcome, and it works — until it starts competing with itself. A supervisor who must inspect everything inspects quickly. Sixty fast looks find less than fifteen careful ones, and the sixty cost the whole day. Past a certain point, raising the inspection rate lowers inspection quality, and the reported number goes up while the thing you wanted goes down.

There is also a second cost that is easy to miss: a supervisor walking a mandatory route is a supervisor who is not available for the villa that went wrong. Inspection capacity spent uniformly is inspection capacity that cannot respond.

The sample nobody designed

Here is what happens when the standard says everything and the day allows a fraction.

The selection still occurs. Geography and fatigue make it, and nobody’s judgment is consulted. The villas near the office get inspected. The villas at the end of the path, the ones behind the generator shed, the ones where the WiFi is worst and the supervisor’s line of sight has never reached — those go to ready on trust.

That is the worst available sample, because distance and knowledge run in opposite directions. The villas you already know least about are the ones the accidental sample checks least. And because the standard says everything is inspected, nobody records which villas were skipped, so the gap cannot even be looked at afterwards. It is the same mechanism that makes verification stop being free the moment rooms stop sharing a corridor, which is one of the assumptions corridor-built systems get wrong on an island.

An inspection program starts the moment that selection becomes deliberate. Three things can trigger a look, and only one of them is a schedule.

Consequence: some villas have no slack

The first trigger is what happens if the villa is wrong.

A villa cleaned on a departure with nothing arriving until Thursday has slack. If something was missed, there are two days to find it. A villa cleaned this morning for a guest landing at two o’clock has no slack at all: it will be wrong in front of the person it is wrong for. Same cleaning job, entirely different consequence, and the only variable that matters is the arrival.

Properties know this and handle it informally — the supervisor keeps the day’s tight turns in her head and tries to see those. Making it structural is a small change with a large effect, because “in her head” fails on exactly the day the list is longest.

On our own turnover boards a checkout that matches a same-day arrival into the same villa is marked as a hard turn when the job is created. That pulls its deadline forward to the arrival, and it means the job always requires an inspection when the cleaner finishes. The villa moves to awaiting-inspection, and the inspection goes out as its own job to the pool of housekeepers who hold inspect rights.

Pooling the inspection is not a detail. Part of why a hundred percent standard collapses is that it is one person’s walking distance; an inspection any qualified person nearby can claim happens in minutes instead of whenever the route gets there.

Sample: a rule, not a mood

The second trigger covers everything else — the villas with slack, which are most of them.

Those need a sample, and a sample is only a sample if it is chosen by a rule. Chosen by a person, it reverts to proximity within a week. Chosen by a rule that can be re-run, it stops being a sample the first time someone re-rolls it.

Ours is one turnover in four.

The coverage that produces is easy to reason about, which is the other reason a fixed rule beats judgment. Take a day with twenty-two departures, eight of them turning the same day. All eight hard turns are inspected. Of the fourteen with slack, the rule picks three or four. Eleven or twelve inspections, chosen on purpose — against a hundred percent standard that in practice delivers a similar number chosen by whichever villas were nearest.

Same walking. Completely different information.

Evidence: the trigger that costs no walking

The third trigger is the one properties most often leave on the table, because it does not look like inspection at all.

When a housekeeper marks a checklist item as failed — a cracked tile, a shower that will not drain, a count that comes back below its range — that failure can raise a maintenance job by itself: high priority, unassigned so the nearest engineer can claim it, carrying the photo the housekeeper took, sitting on the same board as everything else. Nobody walks anywhere to generate it, and it arrives on the day the problem exists.

Whether a failed item raises a job is a per-item setting. Somebody has to sit down with the template and decide which failures deserve one. That decision is the evidence trigger. Skip it and the third leg of the program does not exist.

What a failed inspection does to the villa

A program also needs a defined path for a fail, or the fail becomes a corridor conversation.

For the program, the villa has only one acceptable destination: back out of ready. On our boards a failed inspection reopens the turnover and returns the villa to in-progress, so it cannot be allocated while the problem is still inside it. A system that records the fail and leaves the villa parked at ready has given you an audit trail and an unhappy arrival. Where the job goes after that — who picks it up, what the second failure triggers — is a routing question, and a good one, but it belongs to whoever designs your task flow rather than to your inspection program.

The reopen goes to the person who cleaned the villa because that person is standing closest to the problem.

The number a program should be able to state

Not the inspection rate. This one:

Of the villas you released yesterday, how many were released without anyone looking — and was that a decision?

A property with a hundred percent standard cannot answer it, because the standard says the number is zero and everyone knows it is not. A property running a selective program answers it immediately, in three parts: the ones that always get looked at, the ones the rule picked, and the ones nobody looked at on purpose.

An inspection program is a selection function, and it fits in a sentence. Write yours down — what always, what sampled, what on evidence — and the villas nobody looked at stop being an accident. If you want to watch the selection run against a turnover board, a walkthrough takes about as long as inspecting two villas properly.