Building a Vendor Performance Scorecard That Actually Bites

Fill rate answers whether you got what you ordered. Take the quantity received against the quantity ordered on each line, over the period, weighted by value or by line count depending on what you care about. It is the single most predictive measure of whether a supplier will cause a stockout, and it is the one most often missing because purchase orders get closed short without the shortfall being recorded anywhere.

On-time-in-full is stricter and more useful. A delivery counts only if the full quantity arrived by the promised date. Partial deliveries and early deliveries both fail, which surprises people until they think about the storage and expiry consequences of a supplier dumping six months of stock in March to hit a target. Measure against the delivery date on the order, and if your orders do not carry a committed delivery date, fix that before you build anything else.

The remaining three are quality rejection rate at goods receipt, documentation compliance, and responsiveness. Rejections should be counted by line and by value, with the reason coded. Documentation compliance asks whether batch, expiry, licence details, certificate of analysis and cold-chain records arrived complete and correct. Responsiveness is measured on quote turnaround and complaint closure time, both of which you can pull from dated records rather than impressions.

If building the scorecard requires a new register, it will not survive. Everything above is already implicit in your stores transactions, and the work is usually to close three gaps rather than to start collecting from scratch. The first gap is that purchase orders do not carry a committed delivery date, so nothing can be measured as late. The second is that short-closed orders lose the shortfall. The third is that rejections are recorded as a note rather than as a coded reason.

Fix those three at the transaction screen and the scorecard becomes a report. Make the delivery date mandatory on order creation, require a reason code when an order is closed short, and give the receiving clerk a short dropdown of rejection reasons rather than a free-text box. Each of these adds a few seconds to a transaction that already happens and removes hours of monthly reconstruction.

Then run it as a report, quarterly, on the same day each quarter. A stores system that holds the order, the receipt and the rejection against one record can produce the whole scorecard without anyone assembling data. What it cannot produce is judgement, which is why the report is the input to the review meeting rather than the output of it.

Composite scores are convenient and quietly deceptive. A supplier with excellent responsiveness and a poor quality rejection rate can score the same as one with the reverse profile, and those two suppliers pose completely different risks. Publish the component scores alongside the composite, always, and set minimum thresholds on the components that matter most so a serious failure cannot be averaged away by good performance elsewhere.

Weight the components by item criticality rather than uniformly. For a critical drug where a stockout has clinical consequences, fill rate and on-time-in-full should dominate and a supplier who is cheap and slow should score badly. For a bulk non-clinical consumable, price competitiveness and documentation may reasonably carry more weight. One weighting across a whole supplier base tells you less than three weightings applied to three supplier segments.

Be honest about attribution too. A supplier scored down for late delivery on an order you released two days before the required date is being blamed for your planning. Before the review meeting, filter out orders with an unreasonably short lead time and count them separately, because that count is a measure of your own indent discipline and it belongs in a different conversation.

Accreditation frameworks expect medicines and consumables to be procured from approved sources and expect that approval to be reviewed rather than granted permanently. An assessor asking about your supplier process is looking for three things: how a vendor gets onto the approved list, what evidence supports that they remain qualified, and what happens when they do not. A scorecard supplies the second and third.

So make the list conditional on performance rather than a static register. Entry requires the qualification documents — licences under the Drugs and Cosmetics Rules where applicable, device licences, GST registration, quality certifications relevant to the category, and the certificate of analysis practice you require. Continuation requires a scorecard above a stated threshold and current documents. Both conditions should be visible on the vendor master record.

Document expiry is the quiet failure. A drug licence or a device licence that lapsed eight months ago sits unnoticed on a vendor master because nothing checks it. Put an expiry date field against each document, run a monthly report of documents expiring within ninety days, and make renewal a purchase-blocking condition rather than a reminder. HealUDoc vendor records can hold licence validity alongside the purchase history so the block is enforced at the transaction rather than discovered at an audit.

De-listing a supplier is a decision that can be contested, particularly in government and trust-run institutions where a supplier may seek relief against what they characterise as arbitrary blacklisting. The process that survives is procedurally boring: criteria published in advance, measured against records, communicated at each stage, with an opportunity to respond before a final decision and a defined duration rather than a permanent bar.

Sequence it. A score below threshold triggers a written performance notice with the underlying data attached. Continued failure triggers a formal show-cause with a stated period for reply. The reply goes to a committee rather than an individual, the committee records its reasoning, and the outcome is communicated with the duration of the exclusion and the conditions for reinstatement. Skipping the show-cause step is the single most common reason a de-listing is set aside.

Distinguish performance de-listing from integrity de-listing, because they warrant different processes and different durations. A supplier who is chronically late is a performance problem and may return after demonstrating corrective action. A supplier who has supplied spurious or misbranded product, or falsified documentation, is a different category entirely and involves reporting obligations to the drug control authority that sit outside your procurement process altogether.

Hold the review with the same attendees each quarter: materials, pharmacy, a clinical representative, finance and quality. Twenty suppliers is too many to discuss; take the five worst, the two most critical regardless of score, and any supplier whose score moved sharply in either direction. The rest are noted, not discussed. A meeting that reviews everything decides nothing.

For each supplier discussed, the output should be one of four things: no action, a performance notice, a change in order allocation, or a move to the de-listing process. Recording an outcome of continue to monitor for the fourth consecutive quarter is how scorecards become theatre. If a supplier has been below threshold for a year and nothing has changed, either the threshold is wrong or the committee is avoiding a decision, and both are worth naming out loud.

Close the loop with the supplier. Send the sheet, invite a response, and give a decent supplier the chance to explain a bad quarter that had a real cause. Some of what your scorecard shows as supplier failure will turn out to be your own indent timing, your own acceptance delays holding up their payment, or a specification the market cannot meet at the rate you awarded. Those findings are worth more than the de-listings.