The Vendor Scorecard: Scoring Contractors on Something Real

A vendor scorecard is only worth building if it answers one question: is this contractor getting better or worse? Any scorecard that cannot show a direction of travel is a report card, not a management tool. It sits inside the wider property management vendor management cycle.

The number itself is the least interesting part. What matters is which categories moved, which items repeated, and whether the contractor's own inspection findings agree with yours. Get those three, and a renewal conversation becomes arithmetic instead of opinion.

Key takeaways

• Score categories you can verify yourself: quality, response, compliance, reporting, safety behaviour. Never score "communication" or "partnership" as standalone categories.
• Weight the categories so a critical quality failure cannot be masked by good paperwork. Quality should carry the largest share.
• Four or five categories and a 1–5 scale per category is enough. Long scorecards do not get completed.
• Grade on a fixed cadence, at the same sampling size each time. A score from a light month is not comparable to a score from a heavy month.
• The scorecard is a conversation tool. Send it before the meeting and let the contractor correct factual errors.
• Track the number, but quote the items. "Restrooms scored 2 in three consecutive inspections" persuades; "we rate you 3.4" does not.

What a scorecard is actually for

Three jobs, and nothing else.

1. Detect drift. A site can lose half a grade over two quarters without a single complaint, because complaints only arrive when something crosses a tenant's personal threshold.
2. Compare fairly. With one fixed method, you can compare the same contractor across months and compare contractors across sites.
3. Carry consequences. A trend is the only defensible basis for an improvement plan, a service credit, or a renewal decision.

If your scorecard is not doing one of those three, it is consuming time you should be spending in the building.

Why single scores fail

A single number hides the thing you need. A site scoring 3.8 overall could be excellent on floors and consistently poor on restrooms.

Worse, a single number invites gaming. If the contractor knows only the total matters, they protect the easy categories and accept losses on the hard ones. Floor care is visible and easy to lift. Restroom detail work is tedious and behind a door. Without a visible category breakdown, effort migrates to whatever moves the total.

Two habits fix both problems. Publish the category scores, not just the total. Show the previous two periods beside the current one.

Choosing categories that can be verified

A category counts only if you can observe it during a normal inspection cycle. These five work in most commercial buildings.

Quality carries the largest weight deliberately. A contractor with immaculate paperwork and a dirty building is failing, and the weighting should say so.

Self-reporting deserves its own line. A vendor who inspects and shares finds problems before you do, and that behaviour predicts everything else.

Deliberately excluded: "communication", "partnership", "flexibility", "value". These are impressions, not measurements. Score the observable consequence — the same discipline as assessing vendor cost against value rather than headline price.

Building the vendor scorecard

Keep it to five categories with a defined 1–5 scale. Define what each number means, or two managers will score the same building two points apart.

Define "target level" in the contract so the scale has an anchor. Most commercial office scopes set target cleanliness at APPA level 2, with level 1 for entrances and restrooms and level 3 acceptable in low-traffic storage areas. Whatever the anchor, it must be written down.

A worked scorecard

Here is the same contractor across three quarters. This is the format to use — categories, weighted score, trend, and open items.

The total fell from 4.20 to 3.25 in six months, and the category detail tells you why. Responsiveness dropped because resolution times started slipping. Self-reporting dropped again. Quality slid one point and stayed there.

That is a different story from "the score went down." It points at supervision capacity, because responsiveness and self-reporting decay together when a supervisor's site count grows.

Weighted totals should be calculated, not argued. Fix the weights in the contract or in the review pack template so the arithmetic cannot become the debate.

Scoring cadence and sampling

Frequency and sample size have to be fixed, or the scorecard measures your month rather than their performance.

• Monthly — full scored inspection of the whole site. Same checklist, same areas, same time of day where practical.
• Quarterly — averaged monthly scores plus a deeper sample covering high dusting, floor care and back-of-house.
• Annually — rolling four-quarter average feeding the renewal decision.

The sampling rule protects you in a dispute. Inspecting three of twenty restrooms in one month and eighteen in the next produces two incomparable numbers. Fix the sample: every restroom once a month, or a rotating six of twenty with full coverage each quarter.

Time of day matters as much as sample size. An inspection at 6am before the shift has started scores an empty, unsettled building. An inspection at 2pm scores the building tenants actually use. Pick one and stay with it.

Critical items should override the average

Some findings should not be averaged away. A blocked fire exit, an unsecured chemical store, a restroom with no soap during operating hours, or a documented access breach all sit outside the arithmetic.

Critical items get their own rule: any unresolved critical finding in the period caps the total score, regardless of the category averages. A cap of 2.0 for the period is common and defensible.

This is the difference between a scorecard that reports and a scorecard that governs. Without an override, a contractor can average a 3.5 while leaving a hazardous condition open for three weeks.

CCOHS inspection guidance supports exactly this approach — its report template assigns each observation a priority (major requiring immediate action, serious requiring short-term action, minor requiring longer-term action) and separately notes whether an item is a repeat, because unresolved repeat items indicate that previous corrective actions were not effective. Adopting both conventions makes your scorecard legible to anyone who has run a safety program. See the CCOHS workplace inspection guidance for the report structure.

Scoring the vendor's own reporting

Self-reporting is the category managers leave out and then wish they had not. Score it on four checks:

1. Was a report received for the period, unedited?
2. Did it contain item-level findings rather than a general grade?
3. Were open items from the previous period carried forward?
4. Did the report arrive by the agreed deadline?

A vendor that scores well here is worth keeping even when quality dips, because you will hear about the dip from them rather than from a tenant. A vendor that scores poorly here is a black box.

Presenting results to the contractor

Send the scorecard and the underlying inspection reports at least three business days before the review meeting, with a clear invitation to correct factual errors in writing.

Structure the conversation around three questions:

• Which categories moved, and what changed operationally?
• Which findings repeated, and what is the root cause?
• What are the three agreed actions for next period, with dates and owners?

Three actions, not fifteen. A list of fifteen becomes a list of zero.

Watch for a specific failure mode: the contractor responds to a score drop by disputing the method rather than the finding. If that happens, the fix is transparency, not argument. Show the checklist, the photos, the sampling rule and the scale definitions. Most method disputes collapse once the contractor sees the same evidence you saw.

The meeting mechanics — agenda, evidence pack, action log — are covered in our guide to running a quarterly vendor performance review.

Common scorecard mistakes

Scoring what is easy to score. Paperwork is easy. Restroom detail at 2pm is not. Weight accordingly.

Changing the checklist mid-term. A new item in month seven breaks comparability with months one to six. Add items at renewal, not mid-term.

Scoring from memory. A quarterly score assembled from recollection is not evidence.

Averaging critical failures. The most common way a scorecard becomes meaningless.

Using the scorecard as a threat. If the only time a contractor hears a number is when it is bad, they will treat the process as adversarial and stop supplying honest self-reports.

Comparing across dissimilar scopes. A scorecard only compares fairly when the scope and target level match. Comparing a Class A tower to a light-industrial warehouse tells you about the buildings, not the contractors. Our guide to benchmarking vendors fairly covers how to normalise before comparing, and the same fixed method is what makes scores comparable when you are managing vendors across a portfolio.

Where NeatScore fits

The scorecard is only as good as the inspections behind it. NeatScore lets you build area-specific checklists with critical items flagged, score each item on the APPA 1–5 scale, attach photos per finding, and inspect offline in parkades and stairwells. Inspections sync when you are back in signal, and each one produces a dated, branded report.

Because the checklist is fixed, the sampling is consistent and the scoring is calibrated, the monthly numbers are comparable — which is the whole requirement for a trend line. Deficiency tracking keeps repeat findings visible so the "repeat item" flag is a fact rather than a memory.

Start a free 14-day trial or see pricing.

Frequently asked questions

What should a vendor scorecard measure?

Measure things you can observe: quality of work against the checklist, responsiveness measured by resolution times, compliance document currency, whether the vendor performs and shares its own inspections, and safety and conduct on site. Leave impressions such as communication and partnership out, and score their observable consequences instead.

How often should I score a vendor?

Score monthly so the record is built from inspections rather than recollection, roll the monthly scores into a quarterly average for the review meeting, and use a rolling four-quarter average for renewal decisions. Monthly scoring also keeps the sampling consistent, which is what makes the numbers comparable over time.

Should the vendor see their scorecard?

Yes, before the review meeting, at least three business days ahead, along with the underlying inspection reports. Give them a written route to correct factual errors. A scorecard the vendor never sees is just internal paperwork, and it will not change a single behaviour on site.

How do I stop scores drifting upward over time?

Fix the scale definitions, keep the checklist and sampling rule constant, and cap the total when a critical finding goes unresolved. Score inflation usually comes from an undefined scale and from a manager who likes the contractor. Reviewing the categories with a second person once a quarter helps as well.

What is a good vendor score?

There is no universal figure, because the target depends on the scope and the building class. What matters is the direction and the repeat rate. A stable 3.8 with no repeat findings is healthier than a 4.4 built on one good month, and any downward trend across two consecutive quarters warrants a conversation regardless of the absolute number.