Compare the options
Rate each option from 1 (poor) to 5 (excellent). Higher must always mean better: for cost, rate affordability rather than entering a price. Weights may be any finite non-negative numbers; at least one must be positive.
Total = sum of weight × rating. The normalised score divides by total weight, keeping the result on the 1–5 rating scale. All joint leaders are highlighted. Unfinished or invalid entries pause the ranking.
Add a criterion
Add an option
New ratings begin empty, so an unassessed option cannot appear to be the winner.
Edit names directly in the table. Remove a criterion or option with its labelled button. This page does not save entries after a reload.
Make the trade-off visible
The starting example has weights of 0.3 for affordability, 0.4 for quality and 0.3 for support. Vendor A scores 3 × 0.3 + 4 × 0.4 + 5 × 0.3 = 4.0; Vendor B scores 3.6 and Vendor C 3.2.
Try a sensitivity check: raise affordability’s weight from 0.3 to 0.7. The three totals tie at 5.2 (normalised to 3.714). Raise it again and the ordering reverses. The preferred option depends on the trade-off you are willing to make.
Choose ratings that mean something
Define what 1 and 5 mean for each criterion and apply the same definitions to every option. Consider a weight as the value of moving from the bottom to the top of that criterion’s scale. Multiplying every weight by the same positive number changes the totals but preserves the ranking and normalised scores.
What this comparison assumes
This additive model allows a strong score on one criterion to offset a weak score on another. Check any must-have requirements separately, avoid counting the same benefit twice, and use criteria whose preferences can be assessed independently. These assumptions and sensitivity checks are discussed in the UK government’s multi-criteria analysis manual, especially chapters 5–6.
The table records your judgements; it does not supply evidence about the vendors or model uncertainty in ratings. A narrow lead calls for a closer look at the underlying assumptions.