Attribution Without Guesswork
Matching sourcing data to funnel outcomes with audit-ready attribution
Recruitment analytics gets frustrating when “channel performance” is based on incomplete tracking. This article shows a practical method to connect sourcing signals to the stages candidates actually reach, so you can defend decisions in board rooms and move metrics in weeks, not quarters.
Why sourcing attribution breaks
Most teams record where candidates come from (requisition sources, referral origin, job board, agency), but they don’t consistently carry that information forward. The result is a funnel that looks complete while attribution remains speculative.
- Source drift: referrals and internal transfers overwrite original origin.
- Stage timing gaps: status changes happen outside the analytics timestamps you rely on.
- Identifier mismatch: different systems use different candidate IDs, so outcomes can’t be reliably joined.
The audit-ready approach: define keys, then map events
Attribution without guesswork is less about clever math and more about disciplined data contracts. Start by defining how a candidate, a submission, and a funnel stage are identified. Then map sourcing events to the earliest attributable candidate record, and preserve it through the funnel.
-
1) Pick a stable candidate identity
Use your ATS candidate ID as the primary key whenever possible. If you must reconcile across systems, standardize a crosswalk table once, then reuse it for every reporting run.
-
2) Establish a single “origin” field with rules
Create an origin field that’s set once using a clear precedence rule: for example, referral beats generic “website,” agency beats import, and internal reclassification must retain the original first-seen source.
-
3) Use event timestamps that match funnel decisions
Stage movement should use the timestamp of the decision or status change, not a later “sync” time. When teams can’t rely on timestamps, attribution confidence drops sharply.
-
4) Compute outcome rates at every stage
Don’t collapse outcomes into one conversion rate. Instead, compute stage-to-stage rates (submission → qualified, qualified → interview, interview → offer) by origin. This is where you learn whether a channel brings fit or simply brings volume.
A practical checklist for “data you can stand behind”
Before trusting a dashboard, validate the joins and the attribution coverage. Your goal is to know what percent of candidates have a usable origin and an outcome you can verify.
Coverage
Origin present for candidates who reached at least the first measurable stage.
Completeness
Every stage outcome uses a consistent timestamp definition across job families.
Consistency
Origin categories roll up cleanly without one-off labels or silent renames.
Reproducibility
Re-running a date range produces the same attribution counts and stage outcomes.
Where matching unlocks faster decisions
Once you can reliably match sourcing data to funnel outcomes, you can stop debating anecdotes and start running experiments. You can see whether dropout concentrates at specific stages for each origin, whether interview-to-offer differs by sourcing channel, and which recruiters or teams are improving selectivity without shrinking candidate supply.
This is the foundation for candidate sourcing ROI that accounts for downstream outcomes rather than top-of-funnel counts.