Ask three facility managers what "clean" means and you will get three different answers — one pointing at a shine, one at a checklist, one at the day the last inspector complained. On a shop floor, acceptance criteria have to survive an argument: a supplier saying the machine did its job, an operator saying the shift was too short, and a maintenance lead saying the housing was never cleared. Good criteria settle that argument with a written standard, a simple test, and a record. Here is how to build them, step by step.
1. Describe the surface, not the task
"Clean the machining centre" is a task; it is not a criterion. Break the machine into surfaces a person can look at and touch: the painted exterior panels, the chip conveyor covers, the way covers and bellows, the coolant tank lid, the floor beneath and around the feet, the walkway, the electrical enclosure and its cable entries. Write each one on the list with its material and its soil type — way oil and coolant residue here, dust and light mist there, caked chip sludge at the conveyor end. Acceptance is then per surface, which is the only way two people can disagree about a job and settle it by looking at the list.
Mark a second column for risk: what may be wetted, what only takes a low-moisture method, and what must stay dry. Enclosures, sensor cabling, bearing ends and open motor housings belong in the last group, and that decision has to be written before an operator improvises it at the machine.
2. Turn "clean" into a test that takes under a minute
Pick a check an operator can run without a laboratory. On a shop floor the practical ones are: a white cloth dragged across the surface, a visual inspection of joint lines and the floor-to-wall fillet, a feel test for tackiness on painted panels, and a look under the machine feet where oil creeps and hides. Write the pass condition in plain words — "no transfer on a white cloth across a 300 mm pass" beats "visually clean" because it can be argued about and settled. Do the check on the worst surface, not the easiest one.
3. Set the criteria with a controlled trial, not a brochure
Standards written from a datasheet fail because the datasheet describes a demonstration, not your soil. Take the easiest surface and the worst one you can find, photograph both, clean them with the method you intend to use, photograph them again, and time the whole job from set-up to clean-down including refills and run-off handling. If the two results diverge — they usually do — you need two standards, not one averaged routine. Anyone comparing machine shop grease removal options should run this trial on their own floor with their own soil, because no supplier's figures reproduce your layout or your access constraints.
4. Put the method limits in the same document
A criterion implies a capability. If the standard requires heat and dwell to lift hardened oil, the document has to state what the machine can deliver: supply, output and tank capacity set how long a continuous pass is and how often it stops for a refill. As an example of the figures to ask for, the Fixed Steam Cleaning Machine PowerDe HW02 product page lists a 220 V supply, 6 kW heating power, 160 °C steam temperature, 0.8 MPa vapour pressure and an 80 L tank. Compare those against your site: the rated circuit available near the work, whether a tap-water connection is within hose reach, and whether the unit fits through the doors and over the floor transitions on the route.
Anything the page does not publish — runtime and charger output on battery units, hose and wand reach, weight, noise — stays an open question and belongs in writing with the supplier before the order. A criterion built on assumed numbers inherits every assumption.
5. Add the criteria nobody puts in a cleaning spec
Two belong in any shop-floor standard. First, run-off: where the collected liquid goes, and confirmation that an oil-bearing wash water route is permitted at your site — floor drains are not automatically a lawful answer, and obligations vary by location. Second, drying: the time until the floor is safe for forklift traffic, measured in your shop in cold weather, not quoted from a demonstration. A surface that passes the cloth test but stays slick has not passed anything.
6. Record it, then let the record change the standard
One line per surface per job: date, operator, method, pass or fail, and a photograph for anything on a critical line. Six to eight weeks of that log tells you which criteria are real and which were aspirational — usually the cadence is wrong before the test method is. It is also the document a supplier's engineer reads before quoting a change, and the evidence you hold up when a specification was not met.
7. Close the loop with the supplier in writing
Hand the criteria to suppliers before they quote. Ask which surfaces each model is cleared for, which components must stay dry, what the documented steam temperature, pressure and capacity are for the specific model, what installation and drainage the site must provide, and what service interval, spares and training come with delivery. Requests that go out as customized steam cleaning equipment enquiries should carry the same page, because the customisation is decided against your criteria, not against a brochure.
Surfaces on a list, a one-minute test with a written pass condition, a trial on your own soil, the machine's real limits, the run-off and drying rules, and a log that argues for changes — that set of criteria is what makes a cleaning contract enforceable and a cleaning job something nobody has to argue about at the end of a shift.
