Vendor Scorecards That Procurement Teams Actually Use
Every procurement team has built a vendor scorecard. Most have built it twice — because the first one died in a spreadsheet after two quarters of manual updates. The difference between scorecards that drive real decisions and scorecards that decorate QBR slides comes down to three things: where the data comes from, how often it updates, and whether anything happens when a score drops.
Why spreadsheet scorecards die
The failure mode is always the same: the metrics live in systems (delivery dates in the ERP, quality issues in emails, invoice accuracy in AP), but the scorecard lives in Excel. Someone has to re-key everything monthly. The someone changes jobs, the quarter gets busy, and the scorecard quietly stops. Worse — because the data was hand-gathered, vendors dispute it, and the conversation becomes about the data instead of the performance.
The four metrics that carry the weight
- On-time delivery % — promised date (PO) vs actual date (GRN). Purely mechanical, indisputable, and the single best predictor of operational pain.
- Quality / rejection rate — quantity rejected at receipt vs delivered. Captured at GRN, not from memory.
- Invoice accuracy — share of invoices that clear 3-way match without exception. Vendors who bill wrong cost you AP hours every month.
- SLA / responsiveness — for service vendors: response and resolution against contracted SLAs, tracked per ticket, not per anecdote.
Add commercial competitiveness (quote position in RFQs) if your sourcing volume justifies it. Resist the urge to score twenty dimensions — a scorecard nobody disputes on four metrics beats a comprehensive one nobody trusts.
Automation is the difference between a report and a system
Every metric above is a by-product of transactions you already record: POs, GRNs, invoices, tickets. When procurement runs on one platform, the scorecard computes itself — every receipt updates delivery stats, every match result updates invoice accuracy, every ticket updates SLA compliance. That's how ProcurePulse's vendor management approaches it: scores that auto-update from operational data, visible to the vendor through their own portal, so the QBR starts from agreed facts.
Consequences: the part most programmes skip
A score without a consequence is a decoration. Working programmes wire scores into decisions: escalation when a score breaches threshold (auto-flagged, not remembered), allocation when sourcing splits volume between vendors, and renewal — the scorecard is the first exhibit when a contract or AMC comes up for renewal. If your renewal conversations don't open with the scorecard, the scorecard isn't real yet.
Starting practically
Don't launch with all vendors — start with the top 20 by spend plus any vendor tied to production-critical operations. Publish the four metrics, let vendors see their own scores, and run one full quarter before attaching consequences. The goal of the first quarter is trust in the numbers; the goal of every quarter after is better vendors.