Review an active recruiting partnership on the quality of the work and the responsibilities each side controls. Look at calibration, candidate context, communication, and process reliability alongside outcomes.
A placement result matters, but it does not explain every strength or weakness in the partnership. Hiring-manager delays and changing requirements also affect the search.
A partner contribution that submission counts would miss
In a Refery engagement, a recruiting partner challenged the combination of skills in a full-stack brief. Their point was specific: deeper specialist work overlapped another seat, so requiring both areas in this hire was constraining the search.
Refery took the question to the founder, received agreement that a broader profile could work, and reported that answer back to the partner. This was a useful contribution before counting another submission.
On the scorecard, it belongs under role understanding and learning. The evidence is the requirement identified, the client's decision, and the answer returned to the search. It is not a claimed conversion-rate improvement.
Read the anonymized case study.
Use a balanced review
| Dimension | Evidence to review |
|---|---|
| Role understanding | Introductions reflect the current mandate and requirements |
| Assessment quality | Relevant claims have context; unanswered questions are identified |
| Candidate alignment | Interest, working conditions, and role expectations are discussed |
| Responsiveness | Updates and changes reach the right people |
| Process reliability | Introductions, consent, duplicates, and status are handled consistently |
| Learning | Feedback produces a visible adjustment in the search |
Use examples from the agreed workflow. Avoid turning the review into a collection of private candidate material distributed beyond the hiring team.
Separate controllable work from shared outcomes
A partner can control the quality of an introduction and the accuracy of its context. They cannot independently resolve an unapproved role budget or schedule interviews that the hiring team will not staff.
Review your side as well: clarity of the brief, quality of feedback, decision speed, and stability of the mandate.
Illustrative finding: A fictional partner's introductions are relevant, but candidates wait for unresolved internal decisions. The useful action is to fix decision ownership, not demand more profiles.
Use ratings with explanations
A simple rating such as “working well,” “needs adjustment,” or “insufficient evidence” can help organize the discussion. Attach an example and an action to each conclusion.
Do not average away an essential issue. If candidate consent or accurate representation is a concern, investigate it directly rather than allowing strong performance elsewhere to cancel it out.
Avoid invented numerical benchmarks for how many introductions every search should produce.
Agree on one improvement cycle
Select the most important problem, define the change, assign an owner, and choose a review point appropriate to the search.
Examples include clarifying a must-have requirement, adding evidence of personal project ownership to introductions, or resolving feedback before new sourcing continues.
A review should change the work. Repeating the same scorecard without an action is administration.
Decide whether to continue
Use the accumulated evidence to continue, narrow, pause, or end an engagement according to the applicable terms. Communicate decisions clearly and preserve a coherent candidate process.
For choosing a partner initially, see the recruiting marketplace evaluation checklist. For correcting an early mismatch, use the first-slate calibration guide.