Most significant sourcing decisions are made by a panel: a buyer, a technical expert, a user representative, sometimes quality, legal or finance. Each brings a different view, which is the point. Consensus scoring turns those views into one defensible result. It normally runs in two stages. First, each evaluator scores their assigned sections alone, against the same criteria and rubric. Second, the panel compares the spread and agrees a score for each criterion.
The order matters. If evaluators see each other's scores before submitting their own, the first or most senior score anchors everyone else, and the panel loses its independent views. Scoring independently first, then discussing, surfaces real disagreements. A useful practice is to flag criteria where scores differ by more than a set margin and ask the outliers to explain, since they may have noticed something the others missed, or misread the response.
Consensus is not the same as averaging. An average hides a disagreement that the panel should resolve; a consensus score records a decision. The chair, or the panel together, sets the agreed score and notes why. Keeping every individual score alongside the agreed one matters for audit, for debriefing unsuccessful suppliers and for challenges in regulated or public tenders, where the buyer must show that the decision followed the published criteria.
How iProcure handles evaluation consensus
iProcure's evaluator workspace is built around this process, and it is shown on sample data today. Each evaluator scores only their assigned sections, without seeing colleagues' scores first. The panel then sees where scores diverge and agrees one score on the record, with every individual score kept. The evaluation model the workspace scores against is live.