Unique Contribution Rate: Measure What Each Sourcing Channel Actually Adds

The short answer

Most sourcing dashboards tell you how much a channel produced or which channel received credit for an outcome. Unique Contribution Rate asks a different question: what share of this source’s evidence-fit leads did none of the other tested sources surface? That makes it useful for detecting redundancy and identifying low-volume lanes that add genuinely different people.

Definition: Unique Contribution Rate

Unique Contribution Rate (UCR) is the percentage of evidence-fit leads surfaced by a source that were not surfaced by the comparison source stack in the same requisition-level test.

An evidence-fit lead is a lead the recruiter actually reviewed against the same job-relevant standard used for every other source in the test. Appearing in an unreviewed result list does not qualify.

UCR measures additive discovery. It should be calculated against a deduplicated lead set and reported with the requisition, role family, review cap, source order, and collection window.

UCR(source) = evidence-fit leads unique to that source / evidence-fit leads surfaced by that source

Example: if a source contributes 20 evidence-fit leads and 7 appear nowhere else in the comparison, its UCR for that test is 35%. That number describes additive discovery in that test, not overall source quality.

UCR complements three other views

Volume: how much did the source produce?

Volume is useful for capacity, but a large lead set can substantially overlap with other channels.

Outcome attribution: which source received credit?

Source-of-hire and related attribution fields are useful for outcome reporting, but attribution rules vary across recruiting systems and organizations. They do not directly measure cross-source discovery overlap.

Overlap: which source pairs are redundant?

Pairwise overlap shows whether two channels repeatedly surface the same people. UCR turns the comparison around and asks how much each source uniquely adds against the full tested set.

Marginal yield: does the next lane still add people?

When source order is controlled or rotated, marginal yield helps evaluate whether adding another lane continues to expand coverage. It pairs naturally with the Search Exhaustion framework.

Pre-registered hypotheses for the first multi-req run

These are expectations written before the controlled dataset exists. They are not results.

1. The largest licensed source will produce the highest evidence-fit lead volume but not necessarily the highest UCR.
2. ATS rediscovery will be low-volume but materially additive on at least some requisitions because prior internal leads are not guaranteed to appear in external search lanes.
3. Referral lanes will contribute a smaller volume and may show high uniqueness, but the result will vary materially by team and role family.
4. Public code or artifact search will contribute more on roles where job-relevant work is commonly public and less on roles where relevant work is rarely public.
5. Title-led and skill-led searches inside the same underlying platform will overlap substantially on some roles, but the overlap magnitude must be measured rather than assumed.

The study will publish contradictions, null results, sample sizes, source order, and collection windows, not only the findings that make the framework look good.

How not to misuse UCR

• Do not rank recruiters with it. UCR describes source contribution inside a search design, not recruiter performance.
• Do not cut a source from one requisition. A lane can be redundant for one role family and essential for another.
• Do not treat uniqueness as quality. Pair UCR with submittal outcomes, response, downstream results, cost, and recruiter time.
• Do not reward questionable collection methods. Additive discovery does not justify scraping restricted systems, bypassing access controls, or using data with unclear provenance.
• Do not hide the raw counts. A 100% UCR from one evidence-fit lead is not comparable to a stable rate from a substantial sample.

Where this connects to SourcingOS

The long-term product opportunity is to record which lane surfaced each lead, retain evidence provenance, and build the overlap matrix as the search progresses. The metric becomes useful when it is a byproduct of normal sourcing work rather than a separate spreadsheet project.

Combine UCR with the five Boolean query archetypes, Search-Path Scarcity, the Source Stack Coverage Worksheet, and the Search Exhaustion Calculator to measure whether the next lane is still expanding the market.

FAQ