How to Measure Cleaning Contractor Performance

If the first sign of poor cleaning is a complaint, the measurement happened too late.

Cleaning contractor performance should be measured against agreed outcomes at the site. This normally includes inspection results, response to reported problems, completion of corrective work and the evidence behind each result. Attendance and task completion can help explain a problem, but they do not prove that the cleaning outcome was achieved.

I think this distinction is important. It is easy to measure activity because activity produces convenient numbers. A cleaner clocked in. A checklist was submitted. A supervisor visited the site. These records are useful, but a full set of records can still describe a building that was not cleaned well.

The scorecard needs to begin with the result the client is paying for.

Start with the cleaning outcome

An outcome describes what should be true after the work is completed.

“Clean the washroom every evening” is an instruction. “Floor is dry and free from visible dirt after the evening service” is an outcome that someone can inspect.

For each important area, write down:

What acceptable cleaning looks like
When the result should be assessed
How it will be scored
What evidence is required
What happens when it fails

The timing matters more than it first appears. A lobby inspected immediately after cleaning is different from the same lobby after the morning crowd has entered. If the inspection cannot happen near the cleaning time, the buyer and contractor should agree how normal re-use of the area will be treated.

Singapore’s National Environment Agency makes the same point in its 2026 Guide on Specifications for Outcome-based Cleaning Contracts. The guide recommends outcomes that are specific, measurable, achievable, relevant and time-bound. It also says that immediate assessment after cleaning is ideal, while reports and photographs can help verify completion when immediate inspection is not practical.

Five parts of a useful contractor scorecard

The exact KPIs should come from the contract and the building. An office, hospital, shopping centre and residential estate do not have the same priorities. However, a useful scorecard usually needs the following five parts.

1. Cleaning outcome

Inspect the areas that matter and record whether the agreed standard was met.

A simple pass or fail method works when the outcome is unambiguous. A rating scale can be more useful when the contract needs to distinguish between acceptable work, a minor defect and a serious failure. The important thing is to define the scale before the inspection, not while the parties are discussing a failed score.

Do not hide every area inside one site-wide average. A clean meeting room should not cancel out a failed washroom. Keep critical areas visible on their own.

2. Response to requests and complaints

Measure the time between a reported problem and the contractor acknowledging it. Then measure the time until the problem is resolved.

These are different events. A quick reply saying “we are checking” is not the same as a completed corrective action.

The record should show:

When the issue was reported
Its priority
Who accepted responsibility
When work started
When it was completed
What evidence closed it

3. Corrective action and repeat defects

A failed inspection item should lead somewhere. It may create corrective work, a supervisor review or a change to the cleaning method.

Track whether the correction was completed and whether the same defect returns in the same area. A high closure rate can look reassuring while the same problem is being closed every week. Repetition is often the more useful signal.

4. Evidence completeness

Evidence should be connected to the item it proves. A photograph in a group chat may have a time and sender, but it can still be unclear which site, inspection or complaint it belongs to.

For evidence that matters to the contract, keep the location, time, person, inspection item and follow-up work together. Use photographs where they add useful context. Requiring a photograph for every routine action usually creates a large archive without making the result easier to understand.

5. Review and follow-through

A scorecard is not complete when the total is calculated. The buyer and contractor still need to review exceptions, agree corrective actions and record any disagreement.

The monthly review should answer:

Which outcomes failed?
Which failures repeated?
What was corrected?
What remains open?
Does the scope or inspection method need to change?

This is also where the parties should record context that a number cannot explain, such as renovation work, unusual foot traffic or an area that was inspected long after cleaning.

A scorecard template

The table below is deliberately simple. Add contract weights and payment bands only after the measures and evidence are agreed.

Measure	Calculation	Evidence	Review question
Cleaning outcome	Satisfactory inspected items divided by applicable inspected items	Completed inspection, item results, relevant photographs	Which areas failed and why?
Request acknowledgement	Requests acknowledged within the agreed response time divided by applicable requests	Request timestamp and acknowledgement record	Were urgent requests recognised quickly?
Corrective work completed	Corrective actions completed within the agreed resolution time divided by applicable actions	Work order, completion time and closing evidence	Which actions remain open?
Repeat defect rate	Repeated defects divided by all recorded defects for the review period	Inspection and issue history grouped by area and defect	Which problems are recurring?
Evidence completeness	Records containing the evidence required by the contract divided by applicable records	Inspection, work order and attachment history	Can both parties verify the result?

Write down the exclusions for every calculation. For example, a cancelled request or an inspection item marked not applicable should not quietly enter the denominator.

Pass or fail, rating scale, or AQL?

There is no single scoring method for every cleaning contract.

Pass or fail is easy to understand. It works well for a specific requirement such as whether soap is available or a bin is below an agreed fill level.

A rating scale records degrees of performance. It can show the difference between a minor defect and a condition that requires immediate action, but the descriptions for each rating must be clear enough for two inspectors to use consistently.

Acceptable Quality Limit (AQL) defines a tolerance for faults across the inspected sample. NEA’s 2026 OBC guide gives a simple example: an AQL of 20% requires satisfactory outcomes in at least 8 out of 10 inspections. This is a contract example, not a universal target. The buyer and contractor still need to select a tolerance that fits the area, risk and service requirement.

Singapore Standard SS 694:2023 covers the measurement of cleaning performance for commercial premises, residential estates and retail food and beverage premises. Where a contract references this standard, the inspection design should follow the purchased standard rather than a summary from a blog post.

How often should cleaning performance be reviewed?

Inspection frequency should follow the building’s risk, usage and contract. A busy washroom and a low-use store room should not be sampled in the same way simply because they are in the same building.

NEA’s 2026 guide recommends daily inspections by cleaning supervisors and monthly joint inspections involving the service buyer and provider. That is useful guidance for Singapore OBC arrangements, but the final protocol should still be written into the specific contract.

Frequent inspection is not automatically better. If the team collects more results than anyone can review or act on, the system becomes a recording exercise. Start with critical areas, inspect consistently and close the loop on failures.

Who should own the score?

The contractor can run operational inspections, but the client should not depend only on a score created by the party being measured. Joint inspections, client sampling or an agreed independent audit can provide balance.

Both sides should be able to see how the score was derived. If an inspection result changes, keep the previous result and the reason for the revision. A disputed score should not be replaced without a record.

This becomes particularly important when the score affects payment. NEA describes a performance management system that includes performance metrics, inspection protocols, documentation, trend analysis, joint reviews and corrective actions. It also cautions that liquidated damages should be proportionate and should not overlap with tiered payment structures in a way that creates a double penalty.

Where attendance fits

Attendance is a useful operational measure. It can explain why an area was missed and it helps a contractor find coverage problems.

But attendance is normally an input. If the contract is outcome-based, a fully attended shift does not prove that the required outcome was achieved. In the same way, a good inspection result should not erase a separate contractual failure to provide required coverage.

Keep the measures separate and use each for the question it can answer.

How CleanLog records contractor performance

CleanLog connects the score back to the work that produced it.

Inspections record the item result, time, location, inspector and any required photographs. A failed item can create a work order, which is then tracked from report to completion. Attendance, requests and corrective work remain attached to the location instead of being reconstructed at the end of the month.

For performance-based contracts, the contract’s KPIs, weights, targets and bands can be configured with CleanLog OBC Compliance. The monthly scorecard is calculated from the underlying records, reviewed by the client and contractor, and signed by both sides. Signed versions are locked so a later revision creates a new version rather than replacing the agreed record.

The purpose is not to produce more numbers. It is to let both parties move from a score to the inspection, request or corrective action behind it.

A practical starting point

Choose one important area at one site. Write down what acceptable cleaning looks like, when it should be checked, what evidence matters and what happens after a failure. Ask the client and supervisor to inspect it separately.

If they reach different results, the first problem is not the software or the score. The outcome needs a clearer definition.

Once both sides can use the same standard consistently, add it to the wider scorecard. This takes longer than copying a list of KPIs from another contract, but it creates a measurement system that people can actually trust.

If you are working through a performance-based cleaning contract, talk to us about the outcomes and evidence the contract needs. We can help you decide what should be measured before you configure the scorecard.