Score Cleaning Quality in Commercial Buildings: A 100-Point Scorecard

The account you lose is almost never the one that failed an inspection. It is the one you scored 97 percent on for eleven straight months, right until the facility manager walked into a first-floor restroom at 9 a.m., found an empty soap dispenser, and decided the whole quality program was theater.

That is what a weak scoring system actually costs. Not a deduction. A rebid on a three-year contract, which is expensive whether you win it or not.

Score cleaning quality by inspecting a fixed set of weighted items in a random sample of rooms, marking each pass or fail against a written standard, then dividing points earned by points possible. A 100-point weighted score is defensible when restrooms and touchpoints carry more weight than baseboards.

Below is the sequence in the order you actually have to do it. Each stage has an input, an output, and a gate you need to clear before the next stage is worth your time.

Stage 1: What standard are you scoring against?

You cannot score a condition you have never defined. Most inspection programs fail here and never recover, because the inspector is grading against personal taste and the client is grading against memory.

Input: your signed scope of work, the building floor plan, and a list of space types. Output: a one-page written statement of acceptable condition for each space type.

The most widely used vocabulary for this in the United States is the APPA five levels of clean. It describes outcomes, not tasks, which is exactly what you need.

Write the target level next to each space type: restrooms at Level 1, open office at Level 2, warehouse mezzanine at Level 3. Then get the client to read that page. Ten minutes of disagreement now beats a year of disagreement later.

Gate: do not build a checklist until every space type in the building has a named target level in writing.

Stage 2: Which items belong on a commercial cleaning inspection checklist?

An inspection item is not a task. It is an observable condition that two different people would grade the same way.

Input: the scope of work and your Stage 1 target levels. Output: an item bank of roughly 60 to 120 inspection points, grouped by category, each written as a condition.

Three rules for writing items that hold up:

• Observable in ten seconds: if the inspector has to move furniture or run a finger along a hidden ledge, it is an audit item, not a routine inspection item.
• Binary-decidable: "Restroom is clean" is useless. "No visible soil at the base of the toilet or urinal" and "No streaks or smudges on partitions below 60 inches" are gradeable.
• Inside the contract: never score your crew on work you did not sell. Scoring high dusting on a monthly account weekly is how you train supervisors to ignore the score.

Group the bank into categories now, because Stage 3 weights categories, not individual items. A workable set for commercial office: restrooms, floors, touchpoints, trash and recycling, break rooms, dusting and vertical surfaces, entryways and glass.

Gate: hand ten items to a supervisor and to yourself, grade the same room separately, and compare. Any item where you disagree is written badly. Rewrite it before you move on.

Stage 3: How to weight inspection items so the score means something

An unweighted checklist says a dusty baseboard and a restroom with no toilet paper are the same event. Your client does not believe that, and neither do you.

Input: your item bank plus your complaint log from the last 6 to 12 months. Output: category weights that sum to 100, plus a separate critical defect list.

Pull every complaint email and service request from the last year and tally them by category. Weight follows two things: how often complaints come from that category, and how much health or reputation risk sits there. Here is an illustrative weighting for a Class B multi-tenant office building.

Weights are yours to set. What is not optional is that they come from your own complaint data rather than from a template you copied.

Separately, build a critical defect list: conditions that are unacceptable at any score. No soap or paper in an occupied restroom, blood or bodily fluid left unaddressed, a propped exterior door, chemicals left unsecured, a wet floor with no sign. Anything involving blood or other potentially infectious material also puts you inside OSHA's Bloodborne Pathogens standard, so it is a training and compliance event, not a point deduction.

Gate: weights total exactly 100 and the critical defect list is posted where crews can see it.

Stage 4: How many areas should you inspect per building?

Inspecting everything is not a plan, it is a fantasy that ends with supervisors backdating checklists in the truck. Inspect a rotating random sample instead.

Input: a numbered list of every inspectable area in the building and your planned inspection frequency. Output: a written rotation plan with a fixed number of areas per visit.

Use this: areas per inspection = total inspectable areas ÷ inspections per quarter. That guarantees every area gets seen at least once a quarter without inflating the workload.

Worked example, illustrative. Northgate Financial Center is 62,000 sq ft over three floors, cleaned five nights a week. You count 46 inspectable areas: 8 restrooms, 3 lobbies, 3 elevator banks, 18 open office zones, 6 conference rooms, 3 break rooms, 3 stairwells, 2 loading areas. You inspect twice a month, so 6 inspections per quarter. 46 ÷ 6 is about 8 areas per inspection.

Two adjustments every operator should make:

• Restrooms are never in the rotation. Inspect at least two restroom sets every single visit. They generate the complaints, so they get the coverage.
• Randomize the rest. Number the areas and pull them with a spreadsheet random function or a shuffled deck. If supervisors can predict the route, they clean the route.

Gate: the rotation plan exists on paper and covers 100 percent of areas within a quarter.

Stage 5: Pass/fail or a 1 to 5 rating?

Input: your item bank. Output: a scoring rubric with photo anchors for the items people argue about.

Use pass/fail. Five-point scales feel more precise and are not: inspectors cluster on 4, avoid 1 and 2 because it feels like an accusation, and the resulting scores compress into a narrow band that never triggers action. Pass/fail forces a decision against the Stage 1 standard.

If you need a middle ground, use 2 points for pass, 1 for partial, 0 for fail, and define partial precisely: the task was done but the condition is short of the target level. Nothing vaguer than that.

Then handle N/A properly, because it is where scores get quietly inflated. If a conference room was locked and occupied, mark those items N/A and remove them from the denominator. Never score an uninspected item as a pass. Never score it as a fail either.

Photo anchors settle arguments. Take one reference photo of an acceptable restroom floor and one of an unacceptable one, and attach them to the item. Do this for the ten items your team disputes most.

Gate: a new supervisor can read the rubric and grade a room without calling you.

Stage 6: How to calculate the 100-point score

Input: a completed walkthrough with pass, fail, or N/A on every item. Output: one weighted score out of 100, plus a category breakdown.

Two lines of arithmetic:

1. Category score = (items passed ÷ items scored, excluding N/A) × category weight
2. Building score = the sum of all category scores

Here is why the weighting matters, using two illustrative inspections of the same building on different nights. Both nights have the identical raw pass rate: 73 of 84 items passed, or 86.9 percent.

Same raw pass rate, three points apart. Night B is the night you get the phone call, because seven restroom failures land on tenants and seven dusty ledges do not. A raw checklist percentage would have told you the two nights were identical.

Round to one decimal and stop there. Reporting 88.83 implies a precision your sample size does not support.

Stage 7: What score should trigger a rework?

A score with no consequence attached is a number in a folder. Publish the action ladder before the first inspection, and apply it the same way on your best account and your worst.

Input: the weighted score and the critical defect flag. Output: a named corrective action with an owner and a due date.

Add one trend rule that catches slow decay: three consecutive inspections below 90 at the same site triggers a staffing and training review, regardless of whether any single score looked survivable. Sites rarely collapse. They erode.

If the contract carries a performance SLA, your internal thresholds should be stricter than the contractual one. You want to catch the problem a level before the client is entitled to a credit.

Gate: every score you have recorded this month has a closed or open action attached to it. No orphan scores.

Stage 8: How do you keep two inspectors from scoring the same building differently?

Scores only have value if they are comparable across people, sites, and months. That takes calibration and evidence.

Input: two or more inspectors and one building. Output: a documented agreement rate and a revised item list.

Run a calibration walk quarterly. Two inspectors grade the same 10 areas independently, no talking, then compare item by item. If they disagree on more than about one item in ten, the problem is almost always the wording of the items, not the judgment of the inspectors. Rewrite the disputed items and re-run.

Three more controls worth the effort:

• Rotate inspectors: nobody inspects the same site more than two or three months in a row. Familiarity inflates scores.
• Photo every failure: a timestamped photo turns a disputed opinion into a fact. It also protects your supervisor when the client claims something was never addressed.
• Photograph a sample of passes too: a folder of dated photos showing an acceptable restroom on 40 different nights is the single best thing you can bring to a rebid meeting.

Then send the client something short: the weighted score, the trend against the prior three inspections, photos of every failure with the corrective action and due date, and closure photos from the last report. Two pages. Facility managers do not read more than that, and the ones who do are looking for the closure evidence.

Frequently asked questions

What is a good cleaning inspection score?

There is no industry-published passing grade, so the number only means what your thresholds say it means. A common operational pattern is to treat 90 and above as acceptable and anything under 85 as requiring a documented correction. What matters far more than the level is the trend and whether your scores move when performance moves.

How often should you inspect a commercial building?

Tie frequency to risk, not to square footage. Many operators inspect new accounts weekly for the first 60 to 90 days, then settle into twice monthly for stable office accounts and weekly for medical, food, or high-complaint sites. Any site with an open corrective action gets inspected again before the action can be closed.

Should the client join the inspection?

Once a quarter, yes. Joint walks calibrate expectations faster than any report and surface scope gaps you are being blamed for but never sold. Keep the other inspections internal, because clients who attend every walk begin adding items informally, and your score quietly starts measuring unpaid work.

Can you tie a subcontractor's pay to inspection scores?

You can, if the scoring method, the sample size, and the thresholds are all written into the subcontract before work starts. Withholding based on a checklist the subcontractor has never seen invites a dispute you will lose. Also be careful that performance-based deductions do not blur the independent contractor relationship, which is a legal question for your attorney.

How long should you keep inspection reports and photos?

Keep them for at least the length of the contract term plus the notice period, and longer if the account has ever had a claim or a dispute. Inspection history is your primary evidence at renewal, during a rebid, and if a slip-and-fall complaint alleges the floor was never maintained.

Where CleanTrack360 fits

Everything above works on a clipboard. It works better when the checklist, the weights, and the math live in one place: CleanTrack360 runs quality inspections with custom checklists, photo evidence attached to failed items, and automatic scoring, so the supervisor grades and the score calculates itself. Inspection reports appear in the browser-based client dashboard alongside schedules and service requests, which means the facility manager sees the closure photo without you emailing a PDF. Reports export to CSV when you want to chart the trend yourself.

Crews and supervisors work from the phone browser today, so there is nothing to install before your next walkthrough. Plans start at $99 per month for Starter with up to 5 team members, $199 for Pro up to 20, and $249 for Business up to 50, priced per plan rather than per user, with a 14-day free trial and no credit card required.