How often should you inspect a cleaning site?
Monthly is the floor. Most companies believe they are doing monthly and are actually doing it when something goes wrong, which is a different thing with a different result.
Ask an owner how often they inspect and you will usually hear “monthly.” Ask when they were last in a specific building and the answer is often six weeks ago, or after a complaint. Inspection frequency is one of those numbers everybody reports as a policy and almost nobody runs as a schedule.
Here is how I set it, and why.
The baseline, by service days
Take it off the service frequency in the contract, because that is what determines how fast a building can drift.
- Seven days a week, or any building with heavy public traffic: every two weeks. Things go wrong faster and get noticed sooner.
- Three to five days a week: monthly, and hold that date.
- Once or twice a week, small footprint: monthly still. It is tempting to stretch it to six weeks on a small account, and the small accounts are the ones you lose without warning.
- Brand new account, any size: weekly for the first four weeks, every two weeks to the 90 day mark, then the regular frequency.
That new-account column is the one I would not compromise on. The first month is when you find out what the scope missed, and 90 days is when accounts start going quiet. Almost nobody loses a contract in a fight. They lose it somewhere around month three, and they find out by letter.
Add one during occupancy hours
Most inspections happen right after the shift, and they should, because that is the work as delivered. But once a month, walk a building in the middle of the day.
A washroom at 6am after the crew has left and the same washroom at 1pm are two different rooms. The 1pm version is the one the tenants are forming an opinion about. That inspection usually scores worse and it is the more honest number, because it tells you whether the frequency in the contract is actually enough or whether you need a day porter in there.
Inspect the person as well as the place
Frequency by site is only half of it. Run it so every cleaner gets inspected at least monthly too. If a supervisor keeps inspecting the three buildings on their way home, you can have full coverage on paper and two cleaners nobody has seen in a quarter.
When the same area fails across several different cleaners, stop looking at the cleaners. Either the workload is wrong or nobody was ever shown how. I have seen the elevator tracks come up over and over at site after site, and the fix was never discipline. It was fifteen minutes with one cleaner at one set of tracks.
What actually limits the frequency
Not the driving. The paperwork afterwards.
For years my supervisors would inspect a building in forty minutes and then spend another forty at home that evening turning it into something a customer could read: typing up the findings, pulling photos off a phone, renaming files, matching them to the right line. Half the time the photos never got attached at all. The inspection was never the expensive part.
Cut that second forty minutes and the same supervisor covers meaningfully more sites without working longer, which is a lot cheaper than hiring another one. That is the honest business case for doing this on a phone rather than on paper. Not the scores. The evenings.
Put it on a calendar per site
Frequency as a policy decays. Frequency as a dated calendar per site, where a missed visit is visible, holds. The difference between those two is the difference between believing you inspect monthly and being able to prove it.
And you will want to prove it. A customer thinks about you maybe four times a year: when they have a complaint, when they pay the bill, when renewal comes around, and when another company knocks on their door. At renewal you either sit down and say you think you have done a good job, or you sit down and say you were in the building 26 times, here are the dates, here are the scores, here is how long fixes took. One of those two conversations supports a price increase.
For the schedule that gets a new account through its first three months, read the 30-60-90 day plan for a new cleaning account. The walkthrough with the client is the quarterly version of this with the customer in the room, and KPIs for commercial cleaning companies covers what to do with the numbers the inspections produce.
Common questions
How often should a commercial cleaning site be inspected?
Monthly is the floor for any site under contract. A building cleaned seven days a week, or one with a customer contact who notices things, is worth every two weeks. A brand new account needs weekly for the first month. Whatever you pick, put it on a calendar per site rather than leaving it to whoever has a light afternoon.
How often should you inspect a brand new cleaning account?
Weekly for the first four weeks, then every two weeks to the 90 day mark, then settle into the regular frequency. The first month is when the scope and the reality stop matching, and the 90 day point is where accounts quietly start to fade if nobody has been watching.
Should cleaning inspections be announced or unannounced?
Unannounced for the regular ones, announced for the walkthrough with the customer. An announced inspection tells you what the crew can do on their best day, which is worth knowing once. The unannounced ones tell you what the building actually looks like, which is what the customer experiences.
How many sites can one supervisor inspect?
It depends far less on the number of sites than on how much of the supervisor's day goes into paperwork afterwards. When the report is generated from the inspection instead of being retyped at night, the same person covers noticeably more ground. That is the lever worth pulling before hiring another supervisor.
What time of day should you inspect?
Mostly right after the shift, so you are scoring the work as delivered. But run one inspection a month during occupancy hours, in the middle of the day. Washrooms and entrances look completely different at 1pm, and 1pm is the version the customer and their tenants are judging.