A weighted scoring model ranks options by scoring them against weighted criteria. It is the standard tool for vendor selection and the most commonly rigged artefact in procurement, because whoever sets the weights sets the winner.
The mathematics of weighted scoring is trivial; the governance is not. The model produces a defensible decision only when the criteria and weights are locked before anyone scores, the scorers are independent of the sponsor, and the losing option's strengths are recorded rather than scored away.
Last reviewed 3 July 2026 · Free and ungated
Before the scores are signed off, get confidential input from senior operators who have implemented the platforms on your shortlist.
Pressure-test your shortlistList the criteria that matter for the decision. Assign each a weight reflecting its relative importance, summing to 100 per cent. Score each option against each criterion on a fixed scale, multiply score by weight, and sum. The option with the highest total wins, arithmetically. Whether it wins legitimately depends entirely on when the weights were set and by whom.
The table below shows an illustrative comparison, scored on a 1–10 scale. Note the shape of the result: the vendor with the better demo loses on the weighted total, because implementation risk and cost carry real weight.
| Criterion | Weight | Vendor A score | Vendor A weighted | Vendor B score | Vendor B weighted |
|---|---|---|---|---|---|
| Functional fit | 30% | 8 | 2.40 | 6 | 1.80 |
| Total cost over five years | 25% | 6 | 1.50 | 8 | 2.00 |
| Implementation risk | 20% | 5 | 1.00 | 8 | 1.60 |
| Vendor viability | 15% | 7 | 1.05 | 6 | 0.90 |
| Support model | 10% | 8 | 0.80 | 7 | 0.70 |
| Total | 100% | 6.75 | 7.00 |
Weighted scoring rarely lies in the arithmetic. It lies upstream, in choices that look procedural.
Even an honestly run model only ranks the options against the criteria the team could think of. It cannot see the criterion nobody proposed: the integrator's A-team rotating off after month three, the vendor's roadmap deprioritising your industry, the true cost of the second integration everyone assumes will be easy. Those are learned from people who have lived with each vendor after the contract was signed, which is not information a scoring workshop generates.
The highest-value intervention is not re-scoring; it is challenging the criteria and weights before scoring starts, and pressure-testing the near-tie at the end. Selected senior operators who have implemented the shortlisted platforms can tell you which criteria turned out to matter in year two, which is routinely different from what mattered in the selection workshop.
The decision owner and the people who will live with the outcome, before any option is scored. After that the weights are frozen. Procurement can facilitate, but weights set by the function sponsoring a favourite vendor should be treated as a draft to be challenged, not a given.
On a 1–10 scale with honest scoring, differences under roughly half a point of weighted total are noise. Treat a narrow result as a tie and decide on factors the model handles badly (contractual flexibility, exit costs, reference customers in your sector) rather than pretending the decimals decided.
Handle hard constraints and price with care: blending price into the weights lets a cheap, weak option climb the table. A common discipline is to score capability first, then evaluate price among the options that cleared the capability bar, or to use total cost of ownership rather than headline licence price as the criterion.