Supplier scorecard calculator

Supplier scorecards, in plain terms

Why supplier scorecards exist

Purchasing decisions get made on price and lead time. The consequences show up months later, in your own scrap, your rework hours, your expedited freight, and your customer complaints. A scorecard closes that gap. It takes evidence you already generate, receiving inspection records, promise dates against receipt dates, open corrective actions, the emails you had to send twice, and turns it into one number per supplier per period that you can compare, trend, and act on.

The second reason is that it takes the personality out of the conversation. “You have been letting us down” is an argument. “Your on-time delivery was 88 percent this quarter against the 95 percent we agreed, here are the seven receipts” is a fact, and facts are what a supplier can actually work on. The third reason is that an auditor will ask how you decide who stays on your approved supplier list, and a scorecard is the shortest honest answer.

What ISO 9001 clause 8.4 asks for

ISO 9001:2015 clause 8.4.1 requires you to determine and apply criteria for the evaluation, selection, monitoring of performance, and re-evaluation of external providers, based on their ability to supply in line with your requirements. It also requires you to retain documented information on those activities and on any actions arising from the evaluations. That is four distinct duties: define criteria, apply them at selection, monitor performance over time, and re-evaluate.

Notice what the clause does not say. It does not name a metric, a formula, a weighting, or a passing score. Auditors do not check your arithmetic against a standard, because there is no standard arithmetic to check it against. They check that you defined criteria in advance, that you applied them consistently, that you kept the records, and that something actually happened when a supplier fell short. A scorecard done properly satisfies all four in one artifact.

Clause 8.4.2 is the other half. It says the type and extent of control you apply to an external provider should reflect the effect that provider has on your ability to deliver conforming product. This is what makes a score useful rather than decorative: a strong score can justify moving a supplier toward dock to stock, and a weak one justifies tightened incoming inspection, source verification, or a hold on new orders. The score is not just a report card, it is the input to a control decision.

Sector standards build on the same clause. AS9100D adds explicit expectations around maintaining a register of external providers with their approval status and scope, and reviewing their performance periodically. IATF 16949 requires supplier performance monitoring against defined indicators, including delivered product conformity, disruptions at the receiving plant, delivery schedule performance, and any special status customer notifications. ISO 13485 requires the evaluation and selection criteria to be proportionate to the risk the purchased product carries, with monitoring and re-evaluation and records to match. Different wording, same shape: define, apply, monitor, re-evaluate, record.

Metrics that are actually worth tracking

The test for any metric is simple: can you pull it the same way every period, from records you already keep, without a spreadsheet archaeology project? If not, it will quietly stop being measured by the third quarter. These earn their place.

• Lot acceptance rate. Lots accepted divided by lots received. It is the most durable quality metric because receiving inspection already produces it, and it works whether you buy ten castings or ten thousand fasteners.
• Defect PPM. Defective pieces per million received. Worth adding when you buy in volume, because lot-level pass or fail treats a lot with one bad part the same as a lot that was half scrap.
• On-time delivery. Receipts that landed inside the agreed window, divided by total receipts. Define the window once and never quietly change it. Decide up front whether you measure against the original promise date or the latest revised one, and whether a delivery two weeks early counts as a miss (in many plants it does, because it becomes your inventory problem).
• Corrective action performance. How many formal requests you had to raise, how quickly containment arrived, whether the response identified a real root cause or just an apology, whether it closed on time, and whether the same failure came back. The count is the easy part to score, and the handling is the part that tells you the most.
• Documentation compliance. Certificates of conformance, material test reports, and calibration certificates arriving complete, correct, and with the shipment. This is invisible until an audit, and then it is the whole day.
• Responsiveness. Subjective, and worth keeping anyway, because it predicts how the next problem will go. Label it as subjective, keep its weight modest, and write down why you gave the rating you gave.

What to leave out: anything the supplier cannot influence, anything you would have to reconstruct by hand, and price. Price belongs in the sourcing decision, not in a quality scorecard. Mixing it in makes a cheap supplier look good at exactly the moment their quality is telling you otherwise.

The math on this page

Nothing here is hidden. These are the exact formulas the calculator uses.

Dividing by the sum of the weights rather than by 100 does two things. It keeps the result on a 0 to 100 scale even when your weights do not add up to 100, and it lets the calculator drop a component you have no data for and renormalize the rest, instead of silently scoring an empty category as zero. If a component is excluded, the result says so.

One thing to be clear about: there is no industry-standard supplier scoring formula. The 40 / 30 / 15 / 15 split, the ten-point deduction per corrective action, and the A to F bands are common conventions, and that is all they are. ISO 9001 puts the criteria in your hands on purpose, because only you know what a late shipment or a bad lot costs your operation. Set the numbers deliberately, write them into your supplier quality procedure, and be able to explain why you chose them.

Turning a score into a decision

A score becomes arbitrary the moment you decide what it means after you see it. The fix is to define the bands and their consequences in advance, in writing, before anyone is scored. Then the outcome is not a judgment call, it is the application of a rule you published. A workable structure:

• Approved. Routine monitoring continues at the normal cycle. Send the scorecard anyway, so a good result is on the record too.
• Approved with an improvement item. Performance is acceptable but one component is dragging. Name it, agree a target and a date, and check it next period.
• Conditional. Request a documented corrective action plan with named owners and dates. Increase the control you apply in the meantime, and shorten the scoring cycle. Start qualifying an alternate source if this supplier is single-sourced.
• Restricted. No new business pending management review. Tightened or source inspection until an agreed number of clean periods.
• Disqualified. Removed from the approved supplier list, with the record of why. This is the last step, and it needs an approved alternate lined up first, otherwise the decision will get reversed by the production schedule.

Two safeguards keep this honest. First, write down who can override an outcome and on what basis, and require the reason to be recorded. Overrides will happen, because sometimes there is genuinely only one supplier who can make the part. An override with a documented rationale is a controlled decision. An undocumented one is a hole in your system. Second, check the denominator before you act. Two late deliveries out of four is 50 percent on-time performance, and it is also not a trend. Set a minimum activity threshold below which a supplier is reported but not graded.

How often to score

Match frequency to risk and volume rather than scoring everyone on the same clock. Monthly suits high volume or single-source critical suppliers, where a problem compounds fast. Quarterly is the common default and usually the right one: enough transactions to be meaningful, frequent enough to catch a slide. Annual is fine for low risk, low spend, catalog suppliers, and it also satisfies the re-evaluation duty in clause 8.4.1 for everyone on the list.

Layer event triggers on top of the cycle. Re-score immediately after an escape that reached your customer, a repeat of a nonconformance you already raised, a change of ownership or manufacturing location, a lapsed certification, or a new program award that increases your exposure. And keep the method stable: if you change the weights or the deduction scheme, note the change on the scorecard itself, otherwise the trend line becomes fiction.

Keeping the scorecard fair

A scorecard that suppliers consider unfair stops producing improvement and starts producing arguments. Four habits prevent most of that.

• Attribute misses correctly. A delivery that was late because your purchase order landed late, or because engineering changed the drawing mid-build, is not the supplier’s miss. Record a reason code at receiving so this is a fact and not a debate months later.
• Respect severity. One catastrophic lot that stopped your line is not equivalent to four cosmetic rejects, but a raw acceptance rate treats them the same. If severity matters to you, add a severity multiplier and state it, rather than adjusting scores by feel.
• Show your work. Send the supplier the score and the raw counts behind it. A number without its evidence invites a fight. It also catches your own data errors before they harden into a trend.
• Keep subjective inputs small and labeled. Responsiveness is a real signal, and it is also the input most likely to drift with whoever is having a bad week. Cap its weight, name the rater, and require a sentence of justification.