Weighted Scorecard Guide: How to Score Suppliers Fairly

Three quotes land on the same afternoon. One is clearly the cheapest, and it's the one your operations lead wants to sign. Then you reach the delivery schedule and see that supplier wants a lead time that puts your restart date at risk. The second bid is fine on delivery and thin on support. The third is strong on everything and the most expensive, which is the hardest position to defend in a room full of budget holders.

Price is measured in currency. Delivery is measured in weeks. Support is measured in expectations. Comparing them feels like comparing a weight to a temperature, so the conversation circles until someone says "let's just take the cheapest and manage the risk."

A weighted scorecard ends that loop, and it does it in public.

Table of Contents

• What a Weighted Scorecard Is and Where It Comes From

  • The three parts, in the order you build them

  • Where the idea comes from

• How Weights and Scores Add Up to One Total

  • Why weights exist at all

  • Three checks before you add anything up

• Choosing Criteria, Gates and Weights

• Scoring Evidence So Results Hold Up Later

  • A rubric that says what each score means

  • What to record against every score

• Testing Whether the Winning Supplier Holds

  • The procedure, in four moves

• How to Build a Weighted Scorecard Step by Step

What a Weighted Scorecard Is and Where It Comes From

A weighted scorecard is a grid you build before the bids arrive. Down the left side you list the criteria that matter for this purchase. Across the top you list the suppliers. Each criterion carries a weight, and the weights add to 100%. Each supplier earns a score on each criterion, usually between 1 and 5. Multiply each score by its weight, add the column, and every supplier finishes with one total you can rank.

A school report card works the same way. One subject counts double, another counts once. Two students with identical raw marks can finish with different grades because the school decided which subjects matter more. That stays fair as long as the weighting is published in advance, and suppliers deserve the same courtesy. The weights are fixed and dated before anyone sees a price, so no one can nudge them toward a favorite once the numbers are in.

The three parts, in the order you build them

• Criteria are what you judge: price, delivery, quality, service, financial stability, sustainability, implementation approach. Each one needs a one-line definition so two evaluators read it the same way.

• Weights are how much each criterion counts, expressed as percentages that total 100%. Settling them is where the argument lives. Moving 15 points from quality to price is a business decision, and the people who own the outcome should make it, not the person filling in the spreadsheet.

• Scores are how well each supplier meets each criterion on a fixed scale. They come from evidence rather than impressions, and every one of them needs a document behind it.

Where the idea comes from

The historical foundation is the Balanced Scorecard, introduced by Robert Kaplan and David Norton in a 1992 Harvard Business Review article. Their year-long research project examined 12 companies and combined financial measures with indicators for customer satisfaction, internal processes, and innovation and improvement, as set out in Kaplan and Norton's 1992 article. The approach spread widely through private, public and nonprofit organizations in the years that followed, which shows how far structured, weighted measurement travelled into ordinary management practice. Procurement took one lesson from it: price alone should not decide an award.

The whole exercise is compression. A weighted scorecard turns several judgments into one comparable number, and the weights state in public what the buyer cares about most. That public statement is also where the risk sits. Hand price too large a share of the total and a cheap bid can top the table even when it scores poorly on delivery, because a strong price score cannot repair a weak service record once the contract is signed. The three-supplier example below shows what that number looks like, and how easily the weighting, not the suppliers, decides who wins.

How Weights and Scores Add Up to One Total

Five criteria, three suppliers, scores from 1 to 5. The weights are fixed and total 100%.

Take Supplier A line by line:

• Price: 4 × 0.30 = 1.20

• Delivery: 5 × 0.20 = 1.00

• Quality: 3 × 0.25 = 0.75

• Service: 4 × 0.15 = 0.60

• Financial stability: 4 × 0.10 = 0.40

• Total: 3.95 out of 5

Supplier B runs to 3.85. Quality is its strongest line at 1.25, but 0.90 on price and 0.30 on financial stability pull it back. Supplier C also lands on 3.85. It takes the full 1.50 on price and 0.50 on stability, then hands most of it back on delivery, where a score of 2 earns just 0.40. Two suppliers tie, and the leader holds a lead of 0.10 out of 5. That isn't a comfortable gap.

The scores here are illustrative. In a live bid the price line comes from a published formula rather than an evaluator's opinion, and price, delivery time, technical compliance, service capacity and financial stability all arrive in different units. Weighting is what converts those mixed units into one composite score, while the grid keeps every individual score visible underneath.

Why weights exist at all

The technique sits inside multi-criteria decision analysis, which compares alternatives against several criteria at once rather than on a single metric. One documented weighted-linear-combination method uses a 7-point Likert scale to express preferences, derives criterion weights through an eigenvector method, then combines the weighted values into a ranking, as described in a published weighted-linear-combination method. The mechanics can be fancier than the version above. The audit requirement doesn't change: document the scale, the weight on each criterion, the evidence for every score, and the aggregation formula.

Three checks before you add anything up

• The weights total 100%. A set that adds to 92% shrinks every total and makes cross-supplier comparison meaningless.

• Every criterion uses the same scale. If quality is scored out of 10 and service out of 5, the weights stop meaning what the document says they mean.

• The scale and the weights were fixed before scoring began. If either changed after a price was opened, say so in writing and explain why.

Get those three right and the arithmetic is trivial. Almost every scorecard failure happens before the multiplication starts.

Choosing Criteria, Gates and Weights

The most expensive mistake isn't a poorly chosen weight. It's a requirement that should never have been scored at all. Official New Zealand procurement guidance warns that an unrealistically low-priced offer can achieve the highest overall weighted score even when the supplier cannot deliver at that price. A supplier that wins on paper and fails in delivery isn't a scoring error you can argue your way out of afterwards. UK guidance takes the same view, noting that solution deliverability and delivery risk belong in bid evaluation, not in a post-award conversation.

That's the reason to split your criteria into two groups before you weight anything.

The gate test: if you wouldn't sign the contract with this supplier whatever its score, the requirement belongs in a gate, not in a weighted line.

Gates cover legal registration, required insurance, a mandatory security certification and minimum capacity. A supplier that fails a gate is excluded and earns no points, which means strength on price can't rescue weakness on compliance. Scored criteria cover the tradeable attributes: service quality, implementation approach, delivery capability, sustainability. Those earn points and carry weight.

Keep the scored list to five to eight criteria. Every extra line dilutes the weight on the criteria that decide the outcome and pushes evaluators into scoring things they have no evidence for. A shorter scorecard with hard disqualifiers is usually more reliable than a granular one nobody can defend.

Price needs its own rule, written down before bids open. Score each valid bid as the lowest valid price divided by that bid's price, multiplied by the top score of 5. The cheapest bid earns 5. A bid 20% dearer earns about 4.2. Fix that formula in the document, and no evaluator can adjust it once the numbers are on the table.

Total cost of ownership, delivery risk and contract-change exposure belong in the model too, because a quoted price is not a cost. When the purchase is a managed service rather than a product, the delivery-capability criterion should also cover what leaving looks like; IT Cloud Global's guide to switching providers sets out the questions worth asking before signature, and its points make good evidence for that line. If you're still deciding which criteria deserve weight at all, this breakdown of what criteria to weight in a supplier evaluation template is a practical starting point.

Scoring Evidence So Results Hold Up Later

A score of 3 and a score of 4 sit next to each other on the page. Nine months later, when a losing bidder asks why its offer came second, only one of those numbers will have a document behind it.

Public-sector buyers face this directly. The OECD audit of public procurement requires criteria and weightings to be defined, transparent, linked to the contract, and applied objectively using only the published criteria. That standard is worth borrowing even in a private company, because the question it answers is the one your finance director will eventually ask.

A rubric that says what each score means

A rubric is the difference between a score and a preference. Give each level one meaning and one evidence test, so that two evaluators reading the same bid land on the same number.

What to record against every score

• The requirement as written, quoted rather than paraphrased.

• The rubric level awarded, plus the level below it, so a reviewer can see where the boundary sat.

• The source document, with page number or video timestamp.

• A one-line rationale in the evaluator's own words.

• A confidence level, high, medium or low.

• Partial status, kept as its own field rather than folded into the number.

Two habits make this work. Assign the same evaluator to every response to a particular question, so one person's reading of "adequate" applies to Supplier A and Supplier C alike. If three different people score the delivery line, you are measuring their habits, not the bids. Then treat partial compliance as a status, not a silent middle score, because a 3 that hides a gap will eventually be read as "nearly passed".

Consider a Quality award of 4 backed by an attached ISO certificate. The rationale should also note that the uptime report covers a single year, so the gap stays visible to whoever reads the file next year. Where a claim has no document behind it at all, the honest score is a 2 until you've checked, and guidance on sources of supplier information covers which independent records to pull. For a service contract the same bar applies: a contractor should be able to produce documentary proof of its certifications and past performance, and Solomon's Facilities Management on supplier vetting is an example of the kind of record a contractor puts in front of a buyer.

A score of 3 and a score of 4 look adjacent on paper. In an audit they are two entirely different claims.

Then lock the file. Keep the original criteria and weights as issued, and log every later amendment with its date and the reason for it. A reviewer who can see the whole chain of changes is a reviewer who can defend the award.

Testing Whether the Winning Supplier Holds

The total isn't the decision. It's a candidate for the decision, and the question you owe it is simple: would a reasonable change in priorities produce a different winner?

This is sensitivity analysis, and it takes twenty minutes. Rerun the same scores under different weights and watch the ranking.

In the second scenario, Supplier A holds at 4.05 while Supplier C reaches 3.95 and Supplier B drops to 3.65. The lead is still 0.10, so the order survives. Push price to 45% and the third scenario flips it: Supplier C reaches 4.10, Supplier A sits at 4.00, and the same scores now point at a different supplier. Nothing about the suppliers changed. Only the statement of what the buyer values changed.

Notice what happened to get there. Supplier C carries a price score of 5 and a delivery score of 2, which is exactly the profile the New Zealand guidance warns about: the cheapest offer, and the weakest promise on the one line that decides whether the contract works. Raise the price weight far enough and that weakness stops mattering to the arithmetic.

The procedure, in four moves

1. Take the two closest suppliers. They're the only ones that can realistically swap places.

2. Move each weight by five to ten points, keeping the total at 100%, and record which scenarios change the winner.

3. Change the cost assumptions too, not just the weights: a forecast price rise, a likely contract change, a failure cost you'd absorb if delivery slipped.

4. Write down each scenario and its result next to the scorecard, so the test is part of the record rather than a conversation someone remembers differently.

Decision rule: if the winner holds across every plausible scenario, award and file the test with the scorecard. If it flips, you have uncertainty, not a winner.

When it flips, don't switch suppliers and don't hold the original weights just because they were the ones you wrote first. Re-check the evidence behind the closest scores, call two references, and look harder at the risk lines. A narrow margin is a signal to investigate, never proof of superiority. Once the award is signed, the same discipline carries into the contract through a clear method for tracking and scoring vendor performance after signature.

How to Build a Weighted Scorecard Step by Step

Ten moves, in order, each with the reason it matters and the slip that catches teams.

1. Write the purchase need in one sentence, then list the must-haves underneath it. A clear need chooses its own criteria. The slip is starting from a supplier list and reverse-engineering a justification.

2. Set the gates. Legal registration, insurance, security certification, minimum capacity, delivery capability. A supplier that fails one is out, and no amount of price advantage should bring it back.

3. Choose five to eight scored criteria and define each in a single line. Vague criteria produce inconsistent scores, and inconsistent scores produce an award nobody can defend.

4. Assign weights that total 100%, fix the price formula, and date the document before bids open. This is the step that prevents the most common failure: weights re-tuned after a price is seen.

5. Apply the rubric so each level from 1 to 5 has one meaning and one evidence test. Without it, two evaluators score the same answer two ways.

6. Name one evaluator per question and log the source behind every score. An unsourced score can't be reviewed, and a score that can't be reviewed can't be defended.

7. Multiply, add, rank. Keep the arithmetic visible line by line, so a reader can check any total without rebuilding the sheet.

8. Record partial and missing answers separately from the number. A 3 that hides a missing certificate is worse than a 1 that names the gap.

9. Run the sensitivity test before any award. If the winner moves when a weight moves by ten points, you don't have a winner yet.

10. Save criteria, weights, scores, evidence and sign-off in one record, with every amendment logged and dated.

The finished scorecard is a single sheet: criteria down one side, weights beside them, three supplier columns, totals at the bottom, and a link from every score to its evidence. That sheet is the decision record, and it should be readable by someone who wasn't in the room.

Once the manual method runs cleanly under your own hands, a platform can carry the load. Procright scores each product against every line of your specification and links each score to the page, PDF or video behind it on the Professional plan and above, with buyers setting their own weights rather than accepting a fixed checklist. In a full sourcing run it drafts the specification, scans the market and returns a side-by-side comparison of candidates, so the same method you just built by hand becomes repeatable across purchases.

If a walkthrough helps, this short video covers the same sequence from requirement to award.

Start today with the smallest useful version. Write the gates and the weights into one dated document before any supplier sees the requirements, and you'll have the part that's hardest to fix later already locked.

Procright is built for the moment after you've written the spec and set the weights, when you need every score to point back at something a reviewer can open. It compares candidates side by side against your own weighted criteria, keeps the evidence attached to each answer, and produces one record instead of a folder of emails. Run a free test on your next purchase at Procright and see what your scorecard looks like when the sources come with it.