Cleaning Inspection Scoring Systems: Pass/Fail, Ratings & Weighted Scores

None of these is universally right. The rest of this guide explains how each behaves, so you can pick deliberately. If your standards aren’t yet written clearly enough to judge, start with how to write cleaning inspection standards — no scoring model can rescue vague criteria. For how scoring fits alongside tasks, specifications and SLAs, see Commercial Cleaning Standards.

An inspection score summarises the condition observed against the standards actually inspected, at the time they were inspected.

That is all it says. A score of 94% does not necessarily mean:

• every item passed
• every corrective action has been completed
• no critical control failed
• every standard in the template applied to this site
• the site is “certified clean”, or that the client will be satisfied

A score of 94% can still conceal something the client cares deeply about. Keeping that in mind shapes almost every decision below.

Pass/Fail is crude. That is sometimes exactly why it works. Each standard asks one question — was it met? — which is quick to answer, easy to train and hard to argue about. It also scales: a 60-point checklist is still manageable when every decision is binary.

N/A isn’t a rating. It is an applicability state: the standard wasn’t scored because it doesn’t exist here.

The weakness is obvious. A single missed splash and a basin nobody touched both read as Fail. Weighting, critical controls and the written finding carry that difference instead. And Pass/Fail is only as good as the standard behind it — “basin is clean” gives two supervisors room to disagree; the version above doesn’t.

A three-point scale adds a middle state. An illustrative set of labels for a mirror:

That middle state is real information: a cleaner can fix streaking in thirty seconds, and a report that separates it from a filthy mirror is fairer to everyone. What the labels are worth numerically — whether “Needs attention” counts as half, nothing or something else — is a policy each organisation has to decide and write down. There is no standard conversion.

If inspectors cannot reliably agree on the difference between those states, the extra option is creating noise rather than insight.

Sometimes.

Five levels give more granularity, and they ask supervisors to make four boundary decisions instead of one. Without written anchors for each level, two competent people will score the same room differently:

Same room, same dust. The difference is the supervisor. Across a portfolio, that shows up as one site “improving” when its inspector changes — a hard grader/easy grader effect that has nothing to do with cleaning. Five-point scales aren’t more sophisticated by default; they are more demanding.

If supervisors cannot explain the observable difference between 3/5 and 4/5, the system is recording personal judgement with extra decimal places around it. Averaging those ratings across a month produces a figure like 3.72 that looks precise and isn’t.

clear standards + clear rating definitions + inspector calibration = precision

More scoring choices appear nowhere in that sum. A well-calibrated Pass/Fail inspection is usually more trustworthy than an uncalibrated 1–5 one.

A test worth applying before adding rating levels: what decision will we make differently because we know this item was a 4 rather than a 3? If there is no meaningful answer, the extra granularity probably isn’t worth the inspection time and subjectivity. (This is CleaningQA’s practical guidance, not an industry rule.)

Take ten equally weighted, applicable standards. Nine pass and one fails.

9 ÷ 10 × 100 = 90%

The percentage is only the summary. The failed standard — say, “bin exterior in the breakroom has visible residue” — is the part someone can act on. A report that shows 90% without the finding beneath it has thrown away the most useful line.

A weight decides how much a standard contributes to the final score. In CleaningQA’s model:

Three of four standards passed, yet the score is 66.7%, because the one failure carried a third of the weight.

Why weight inspection items at all?

Equal weighting gives every standard equal influence. A contract rarely sees quality that way: a smeared entrance door seen by every visitor, a washroom that generates complaints, a standard written into the contract, an area with a history of repeat failures. Weighting lets the score reflect those priorities.

Weights should represent genuine inspection priorities agreed in advance — never be adjusted after the fact to manufacture a desired score.

The same failure, two weightings

Nothing observed has changed. In the second version the basin makes up five-eighths of the applicable weight, so failing it removes five-eighths of the available score. That is weighting working as designed — which is exactly why weights need to be chosen carefully and explained to anyone reading the report.

A template has ten standards. Two don’t apply to this site. Of the remaining eight, seven pass and one fails.

Correct: 7 ÷ 8 = 87.5% · Wrong: 9 ÷ 10 = 90%

The wrong version treats the two N/A items as successes. N/A is not a generous Pass. It means “we didn’t score this because it doesn’t apply” — not “this standard was met”.

Unanswered is not N/A

Urinals in a washroom that has no urinals: N/A. Mirrors that nobody checked: unanswered. The first is a fact about the site; the second is a gap in the inspection.

Treating unanswered items as Pass inflates scores. Treating them automatically as Fail punishes an inspection for being incomplete rather than for anything the inspector saw. CleaningQA leaves unanswered items out of the calculation until they are answered, and when nothing applicable has been answered the result is Not scored yet — not 100%, and not 0%.

A partially completed inspection

Ten equally weighted standards: 4 Pass, 1 Fail, 2 N/A, 3 unanswered.

4 ÷ (4 + 1) × 100 = 80%

That 80% describes the five standards actually judged — not the whole template. Which is why completion context (“5 of 8 applicable standards answered”) belongs next to the number.

Those are two separate pieces of information, and a report needs both. A critical control is a standard the contract treats as non-negotiable — the one that should never be averaged away by twenty passes elsewhere.

CleaningQA keeps critical failures visible separately rather than quietly imposing a score penalty. That isn’t the only valid approach, but any alternative should be explicit. Rules such as “any critical fail subtracts 20%” are fine if they have been deliberately designed and agreed; they are a problem when they sit inside unexplained maths. If a critical failure should override an inspection outcome — for example, “any critical fail means target missed” — write that policy down.

There is no universal target that applies to every cleaning contract, and no reliable “industry standard” percentage. A sensible target depends on the contract expectation, the site type, the weighting model, the rating method, the client’s requirements, the historic baseline and how much deficiency is tolerable. The 90 used in CleaningQA’s examples is an illustration, not a benchmark.

A responsible way to set one:

1. 01Finalise the standards first.
2. 02Decide the rating and scoring model.
3. 03Run several representative inspections.
4. 04Look at what different scores actually look like on site.
5. 05Agree which result genuinely represents an acceptable inspection.
6. 06Document the target against the contract.
7. 07Review it only when the underlying contract or standards change materially.

Lowering a target because teams keep missing it turns the target into a description of current performance rather than a standard.

The score records the condition at inspection time. Corrective action records what happened next. When the fourth correction is eventually verified, the inspection stays at 91.3 — it does not become 100, because what the supervisor observed on the day has not changed.

Both timelines belong in a client-quality record. Rewriting the score would hide the fact that four issues were found; ignoring corrective status would hide the fact that three were fixed and checked. Target met does not mean fully resolved.

1. 01Treating N/A as Pass. Inflates the score with standards nobody judged.
2. 02Treating unanswered as Pass. Rewards incomplete inspections.
3. 03Over-weighting too many items. If half the checklist is “high priority”, nothing is.
4. 04Changing weights mid-period without preserving history. Makes this month’s score incomparable with last month’s.
5. 05Letting critical failures disappear into the percentage. 96% reads as success even when the one standard that mattered failed.
6. 06Rewriting the original score after corrective action. Erases what was actually found.
7. 07Using rating labels inspectors cannot distinguish. Measures the inspector rather than the site.

A note on averaging averages

Inspection A has 10 applicable points and scores 100%. Inspection B has 100 applicable points and scores 80%.

Naive average: (100 + 80) ÷ 2 = 90%

Combined weight: (10 + 80) ÷ (10 + 100) = 90 ÷ 110 = 81.8%

Neither number is “the” answer — it depends on whether a portfolio figure should treat each inspection equally or each point of applicable weight equally. The point is not to average percentages mechanically without knowing what sits beneath them.

A technically clever model is useless if Supervisor A routinely scores the same condition higher than Supervisor B. Calibration doesn’t need to be elaborate:

• two supervisors inspect the same area independently
• compare results item by item
• discuss each disagreement and find its cause
• rewrite the standard or rating anchor that allowed it
• add example photos where words aren’t enough
• repeat after major standard changes and when new supervisors start

Most disagreements trace back to wording. The standard-writing guide covers calibration exercises in more detail.

Simple Pass / Fail

Ten equally weighted, applicable standards in an office inspection. Eight pass, two fail.

8 ÷ 10 × 100 = 80%

80% alone tells the client there were problems. The two lines tell the cleaner what to fix and the supervisor what to re-check.

Weighted

Unweighted, this would be 3 of 4, or 75%. The washroom carries 5 of the 12 applicable points, so its failure alone takes the score to 58.3%. If the contract really does rank washroom presentation above everything else, that is the honest result.

N/A

The washroom has no urinals, so their weight of 3 disappears from the applicable total. Counting them as a pass would have given 10 ÷ 13 = 76.9% — credit for a standard that was never tested.

Target met, work still open

Using the fictional sample client report: the inspection scored 91.3 against a target of 90, so the target was met. Four issues were found. Three corrections have been verified; one is outstanding. All four statements are true at once — the score describes the inspection, the target compares it with the agreed standard, and the corrective status describes what has happened since.

These combine. Weighted Pass / Fail with a handful of critical controls covers a great deal of commercial cleaning.

1. 01Write observable inspection standards. Everything else depends on this.
2. 02Decide what Pass and Fail mean. In writing, per standard where needed.
3. 03Decide whether intermediate ratings add real value. Apply the 4-versus-3 test.
4. 04Identify genuine weighting priorities. A few higher weights, agreed in advance.
5. 05Define N/A rules. When a standard doesn’t apply, and who may say so.
6. 06Separate critical controls. Report them alongside the score, not inside it.
7. 07Set and test the target. Against representative inspections, not guesswork.

Then run calibration inspections before treating the scores as meaningful trends.

Scoring rules evolve. Weights get adjusted, rating mappings change, checklist wording improves, targets move with a new contract. Each past inspection should keep the standards, weights and target that applied when it happened. Recalculating last year’s inspections under this year’s weights rewrites history, however well-intentioned.

CleaningQA freezes a snapshot of the checklist with each inspection for this reason.

Comparing scores over time

A move from 88 to 94 looks like improvement. It probably is — if the scoring logic, checklist scope and applicable items stayed comparable. If the weighting or checklist changed materially in between, the trend needs a note explaining that, and a single month’s movement rarely proves much on its own.

Separate the figures rather than compressing them into one mysterious quality number. An illustrative client view, using the fictional Northstar sample:

Each figure answers a different question the client may ask. See the full sample client report for how findings and evidence sit beneath them, or commercial cleaning quality control for the wider process.

Free to copy — no signup. Fill in one line per rule and keep it with the contract so everyone scores the same way.

Need ready-made standards to score? Start with the commercial cleaning checklist, and read how to conduct a commercial cleaning inspection for running it on site.

What is a good cleaning inspection score?

There is no universal percentage. A 90% target on a short, lightly weighted checklist means something different from 90% on a long, heavily weighted one. Agree a target per contract after running representative inspections against the final standards.

How is a cleaning inspection score calculated?

In a weighted Pass / Fail model, add the weight of every passed standard, divide by the total weight of every answered, applicable standard, and multiply by 100. Failed standards earn nothing but stay in the denominator. N/A and unanswered standards are left out.

Should N/A count as a Pass?

No. N/A means the standard does not apply to that space, so it should be removed from the calculation. Counting it as a Pass inflates the score with work that was never judged.

What happens to unanswered items?

They are excluded until someone answers them. Counting them as Pass inflates the score; counting them as Fail penalises an inspection for being incomplete. Report completion alongside the score instead.

Should some cleaning items carry more weight?

Only where they genuinely matter more to the inspection outcome — for example, washrooms or an entrance a client sees every day. Decide weights before inspecting, and never adjust them afterwards to reach a preferred result.

Can you use a 1–5 cleaning inspection rating?

Yes, if every level has a written, observable description and inspectors are calibrated against it. Without those anchors, a five-point scale mostly records how strict each supervisor is.

What happens if a critical item fails?

Show it separately from the score. If your organisation wants a critical failure to override the inspection result, write that rule down explicitly rather than hiding a penalty inside the percentage.

Should the inspection score change after corrective work?

No. The score records what was observed at inspection time. Corrective actions record what happened afterwards. Report both.

A useful cleaning score makes the condition of a site easier to understand, not harder. If a supervisor can’t explain how the score was produced, what failed and what still needs correcting, the system is too complicated.