Unique Contribution Rate Calculator
Measure what each source or search lane adds that the others did not surface in the same controlled comparison. UCR is useful for overlap and stack decisions; it is not a stand-alone quality or ROI score.
| Source / lane | Reviewed candidates | Unique to this source | UCR | Optional cost | Cost / unique | |
|---|---|---|---|---|---|---|
| 27% | — | |||||
| 45% | — | |||||
| 73% | — | |||||
| 39% | — |
How to read the result
High UCR
A large share of the reviewed candidates from this source were not surfaced by the other sources in the same comparison. The source appears additive for this requisition and test design.
Low UCR
Most reviewed candidates also appeared elsewhere. That can indicate redundancy, but do not cut a source without considering role family, quality, speed, response, cost, and source-order effects.
Do not compare unlike tests
- Keep the requisition and search window fixed.
- Use comparable effort or reviewed-result caps per source.
- Rotate source order across repeated tests so the first source does not automatically look more unique.
- Dedupe on stable identity anchors, with human review when identity is uncertain.
- Report UCR by role family rather than blending unlike markets into one headline number.
The formula
UCR(source) = candidates unique to that source / reviewed candidates surfaced by that source
The denominator must use the same review definition across sources. Read the full methodology and test protocol before presenting results to leadership.