Vendor Scorecards: How Buyers Measure Supplier Performance

A vendor scorecard turns a supplier relationship into a reviewable record. It normally measures delivery, quality, responsiveness, commercial accuracy and the control of product information. The useful version does more than rank suppliers. It shows which event moved an order, who owned the next action and what must change before the next purchase order.

The difficult part is not assigning a red, amber or green status. It is agreeing the definitions before an order is late. A supplier cannot own a missed ex-factory date that was reset after a late fabric approval, just as a buyer cannot act on a score that calls every delivery late without recording the approved delivery change. The same discipline applies to defect rates. Count the inspected units, use agreed defect definitions and separate a contained issue from one discovered after shipment.

This guide sets out the calculations behind on-time delivery and defects. It works alongside a time-and-action calendar, which controls decisions before the event.

What should a vendor scorecard measure?

A vendor scorecard should measure the outcomes that affect the order and the behaviour that predicts them. Delivery and quality are the core results. Sample response, document accuracy, communication and corrective-action follow-through explain whether those results can be relied on in the next season.

Keep the measures few enough to review. Five to seven metrics are normally enough for a quarterly supplier review, with an order log holding the detail behind each result.

Measure	What it records	Evidence to retain
On-time delivery	Whether goods met the agreed delivery milestone	Original date, approved revisions, booking or handover record
Quality performance	Defects found against the agreed inspection basis	Inspection report, quantities checked and defect classification
Sample performance	Whether samples arrived to the agreed brief and milestone	Sample request, comment sheet and dispatch record
Commercial accuracy	Whether the confirmed cost and assumptions remained controlled	Approved costing, change log and purchase-order revision
Responsiveness	Whether enquiries and production issues received a usable response	Issue log with owner, request date and closure date
Corrective-action closure	Whether an agreed action prevented a repeat issue	Root-cause record, action owner and effectiveness check

Set a clear purpose for each metric. Delivery supports allocation and calendar planning; quality supports improvement. Communication must not hide a serious quality or delivery problem through a high average score.

Use the same definitions for comparable suppliers. Category-specific notes belong in the order file, not in a score created to make unlike situations look identical.

How is on-time delivery calculated?

On-time delivery is the percentage of orders that meet the last mutually approved delivery milestone. The calculation is:

Calculation	Formula	Example
On-time delivery rate	Orders delivered on or before the approved date ÷ total eligible orders × 100	18 on time ÷ 20 eligible orders × 100 = 90%
Average lateness	Total days late across late orders ÷ number of late orders	12 late days ÷ 3 late orders = 4 days
Date adherence by quantity	Units delivered on time ÷ total eligible units × 100	Useful where one large order has greater programme impact

The choice of milestone must be explicit. One buyer may measure ex-factory readiness, another cargo handover and another delivery to a nominated warehouse. These are different events with different owners. A supplier score should not record a freight handover as a production delay if the goods were released on the agreed ex-factory date and the later movement sits outside the supplier's control.

Record the baseline date at purchase-order confirmation. If a later change is approved, retain both the original and revised dates, the reason, the approver and the date of approval. The revised date becomes the scorecard date only when the change was accepted before the milestone passed. A request raised after a date has been missed is evidence of a late order, not a retrospective calendar correction.

Delivery measures need an eligibility rule. Cancelled styles, orders not yet due and split deliveries should not be folded carelessly into one percentage. Score each agreed tranche separately when it has its own date and quantity.

How should defect rate be calculated?

Defect rate is the number of defective units divided by the number of units inspected, expressed as a percentage. The calculation needs a defined inspection stage and a rule for counting a unit with more than one defect. Without both, two inspection teams can report different rates on the same lot.

Quality measure	Formula	Use it for
Unit defect rate	Defective units ÷ units inspected × 100	A clear view of how many garments were affected
Defects per hundred units	Recorded defects ÷ units inspected × 100	Tracking the total fault burden in a lot
Major-defect rate	Units with a major defect ÷ units inspected × 100	Escalating faults that affect saleability or function
Rework rate	Units repaired ÷ units received or produced × 100	Seeing the operational cost of correcting the lot

For a supplier scorecard, count a garment once in the unit defect rate even if it carries two faults. Count each fault in defects per hundred units. This prevents a few badly affected garments from making it appear that more units were defective than were inspected. The inspection report should still describe every fault and its classification.

Agree the classification before bulk inspection. A defect classification records whether a fault is critical, major or minor according to the agreed standard and the garment's intended use. A loose thread may be minor in one context; a seam failure can be major because it affects wear or saleability. Do not use a scorecard to redefine a defect after the result is known.

The score should also distinguish detection point. An issue caught and corrected during production deserves a corrective-action record, but it is not the same delivery risk as a defect found during final inspection or after receipt. Keep in-process findings, final-inspection results, post-shipment claims and rework as separate fields. Pre-shipment inspection and AQL explains why a sample inspection result is an acceptance decision for the inspected lot, not a guarantee that every garment was examined.

Which delays belong to the buyer and which belong to the supplier?

The scorecard should allocate a delay to the event that caused it, using the critical path and the decision record. This is the safeguard against punishing a supplier for a buyer's late approval. It also prevents a supplier from treating an earlier buyer delay as a blanket excuse for a later production failure.

Event	Usual scorecard treatment	Record needed
Buyer approves a lab dip after its agreed date	Buyer-caused delay; reforecast the affected material milestone	Approval request, due date, approval date and revised plan
Supplier submits the wrong lab dip or submits late	Supplier-caused delay	Submission request, submission date and correction record
Buyer changes the approved artwork	Buyer change; assess cost and calendar effect before approving the revision	Revised artwork, change request and approved date change
Supplier misses an agreed fabric or trim booking	Supplier-caused delay	Booking plan, confirmation and recovery action
A sample comment changes the garment specification	Attribute timing to the party that issued the change; track the new sample cycle separately	Consolidated comments, version number and sample instruction
Supplier identifies a construction risk early	Do not score as late if the risk is raised in time and the plan is re-agreed	Risk log, options presented and decision date

The crucial distinction is between a late approval and a change. A buyer can approve the requested lab dip late, which moves the material calendar. A buyer can also reject it and request a new colour direction, which changes the work itself. Both may affect delivery, though the scorecard should retain their different causes. The supplier's responsibility is to submit the correct item, flag the cut-off date and show the effect of a missed decision while there is time to act.

Use a delay code on every changed milestone: buyer approval, buyer specification change, supplier material readiness, supplier production, quality hold, shipping instruction or another agreed category. Add a short narrative only where the code cannot explain the event. This creates a record that can be reviewed without re-reading a long email chain.

No scorecard should erase shared responsibility. Allocate the primary cause to the event that first put the critical path at risk, then capture contributing causes in the order review.

How do you build a scorecard that changes decisions?

Build the scorecard from order-level facts, then roll them into a supplier view at a fixed review point. Begin with an agreed data sheet, a named owner for each input and a review cadence that matches the buying cycle. A quarterly review is often useful for trend analysis, while a live order review handles issues that cannot wait for quarter end.

Step	What to set	Why it matters
1. Define milestones	Baseline date, delivery event, approval cut-offs and split-order rule	Stops later arguments about what the measure means
2. Set quality basis	Inspection stage, defect classifications and rework treatment	Makes results comparable across orders
3. Create an exception log	Delay code, cause, owner, action and revised date	Separates a one-off disruption from a recurring weakness
4. Weight the score	Give delivery and quality enough weight to affect the outcome	Prevents easy service metrics from masking failure
5. Review trends	Compare periods, product types and recurring fault families	Directs development and production controls to the real issue
6. Agree actions	Owner, due date and proof of effectiveness	Turns the review into a change on the next order

Avoid a weighted total that conceals the reason for concern. Show the component scores beside the total and set an escalation rule for serious failures.

Use a normalised order file. The approved tech pack, sample comments, cost assumptions, material approvals and packing instructions should all point to their current version. When an order changes, the scorecard should identify the changed input and its timing. The tech-pack guide explains why version control protects both costing and production communication.

For a full-package programme, the scorecard may include material sourcing and development gates as well as garment production. In a contract programme, buyer-nominated material readiness may need its own buyer-side measure. The purpose is not to make every sourcing model use the same columns. It is to make the ownership of each milestone visible.

What makes a supplier score fair?

A fair supplier score uses pre-agreed rules, source evidence and a chance to correct the record. Share the order data before the review so factual errors can be identified before the score is final.

Fairness does not mean ignoring poor performance. It means measuring it at the right point. If a final inspection finds repeat major defects, record the quality outcome even when the supplier offers rework. If the rework moves the handover date, record that delivery effect as well. A recovery action is evidence of response; it does not remove the original event.

Credit early escalation when it protects the agreed milestone. The scorecard should reward the behaviour that keeps a problem from becoming a late shipment.

Short FAQ

Should a vendor scorecard use one overall score?

Yes, if the component scores remain visible beside it. Delivery, quality and corrective-action performance need separate thresholds because a high result in one area cannot cancel a serious failure in another.

Do late buyer approvals count against the supplier?

No. Record the approval date, the agreed cut-off and the approved calendar effect. Score the supplier only for its own submission, escalation and follow-through obligations.

What is a good on-time delivery target?

Set the target for the programme and measure every supplier against the same delivery milestone. The review is more useful when it also shows the number of late orders and average days late.

Can a supplier challenge a scorecard result?

Yes. Give the supplier the underlying order record and a defined period to correct factual errors. Keep the original evidence and document any agreed amendment.

The score that supports better sourcing

A vendor scorecard is credible when it preserves cause as well as outcome. On-time delivery must use the approved delivery date and distinguish production performance from buyer approvals or later freight events. Quality must show the quantity inspected, the defect definition and where the issue was detected.

Use the score to decide what changes next: a sharper approval cut-off, earlier material booking, a revised sample control, an added quality checkpoint or a corrective action with evidence of closure. When the measures are shared and the ownership is clear, the scorecard becomes a working supplier-management tool instead of a retrospective grade.