Build a Hiring Funnel from Scratch

Proof required

Submit for at least 3 roles: (1) job descriptions — each must include: specific problem the role solves, 2–3 concrete 90-day outputs (e.g. 'first API design reviewed and approved by engineering lead'), required evidence (what past work or take-home exercise would demonstrate readiness), compensation range; (2) an evaluation framework for each role: the 3–4 criteria being assessed, how each stage of the hiring process (screen call, technical exercise, panel interview, reference check) assesses which criterion, and the decision process (who has input, who has final decision, how tie-breaks work); (3) a sourcing plan: which channels will be used for each role and why.

What gets checked

• JD outputs are specific and verifiable — '90-day output: design the API schema for the payments module, reviewed and approved by the platform lead' is specific; 'contribute to the team' is not verifiable
• Evaluation framework assigns each interview stage to specific criteria — 'everyone interviews' is not an evaluation framework; 'stage 3 panel interview assesses system design competence (engineering lead) and cross-functional communication (PM lead)' is a framework
• Compensation range is included in the JD — hiding compensation until an offer is made wastes candidate and interviewer time; the JD must include a range

Common mistakes

• JDs with a list of requirements rather than outputs — a list of 15 skills required produces applications from candidates who match the list, not candidates who can produce the outputs; the JD must describe what the person will accomplish, not what they must have on their resume
• Evaluation frameworks that require agreement on a candidate's 'cultural fit' — subjective fit judgements produce inconsistent decisions and introduce bias; each evaluation criterion must be specific enough that two different interviewers would assess it the same way
• Starting sourcing before the evaluation framework exists — sourcing before the decision criteria are agreed produces inconsistent interview questions, panel disagreements, and decisions driven by whoever interviewed last

Resources

What a verifier looks for

• JD output specificity: ask the submitter to read aloud the first 90-day output for one of the roles — is it specific enough that the candidate would know exactly what they are expected to produce by day 90?
• Evaluation framework criterion assignment: ask the submitter which interview stage assesses communication competence for one of the roles — if the answer is 'all of them', the framework lacks assignment clarity
• Compensation range: confirm that each JD includes a compensation range — if it does not, ask why and whether the submitter is prepared to share it with candidates at the beginning of the process

You'll sign in first, then come straight back here.

Proof required

Submit for each of the 3 roles: (1) the funnel metrics: number of applications, screens completed, technical exercises sent and completed, panel interviews, reference checks, offers made, and outcomes (accepted/declined/role closed); (2) an evaluation consistency log: for each hire or final candidate, the scores or notes from each evaluation stage against the pre-defined criteria — showing that the same criteria were applied consistently; (3) one debrief summary per role: the key points from the panel debrief, the final decision and the criteria it was based on, and any criteria that were disputed or unclear.

What gets checked

• Funnel metrics cover all stages — a hiring funnel with only application and hire data is not a complete funnel; each stage's conversion must be reported
• Evaluation consistency log uses the pre-defined criteria — if the final decision for one role was made on criteria not in the evaluation framework (e.g. 'they felt right for the culture'), the log must note this as a deviation
• Debrief summaries are specific — 'the panel agreed on the candidate' is not a debrief summary; 'the panel agreed that the candidate's system design exercise showed strong first-principles reasoning (criterion 2) but their communication in the panel was unclear on trade-off decisions (criterion 3)' is a debrief summary

Common mistakes

• Reporting only successful hires — a hiring funnel that reports only the roles where offers were accepted omits the roles where the process broke down; all 3 roles must be reported regardless of outcome
• Criteria drift during the process — a common failure is starting the process with one set of criteria and ending with a different set after a strong candidate is seen; the evaluation consistency log must reflect the original criteria, not retroactively adjusted ones
• Panel interviewers who did not know their assigned criteria — if interviewers were not briefed on which criteria they were assessing, the evaluation framework was not followed; the debrief notes should show that each interviewer assessed their assigned criteria

Resources

What a verifier looks for

• Criteria consistency: ask the submitter to show the evaluation notes for the best and second-best candidate in one role — do both sets of notes assess the same criteria?
• Process completion: confirm that all 3 roles reached a final decision (offer made, role closed, or hire decision made) — a funnel that was paused mid-process does not count as a completed funnel
• Debrief quality: ask the submitter to describe the most contentious debrief discussion — what criteria were disputed, and how was the disagreement resolved

You'll sign in first, then come straight back here.

Proof required

Submit: (1) a hiring retrospective: for all 3 roles — JD accuracy (did the role turn out to match the JD after 90 days?), evaluation framework effectiveness (which criteria were most predictive, which were hard to assess consistently), funnel health (which stage had the highest drop-off and why), and offer outcome; (2) a lessons-learned document (300 words minimum): what the evaluation framework did well, which criteria produced inconsistent assessments, and how you would redesign the JD or framework for these roles next time; (3) documentation of a real-time Q&A session with a named reviewer who has ≥3 years of hiring management experience and a verifiable professional profile — the documentation must include specific challenge questions and your responses; the reviewer must not be a direct report or someone who participated in the hiring process.

What gets checked

• JD accuracy is assessed post-hire — the retrospective must note whether the 90-day outputs in the JDs matched what the role actually required in the first 90 days; if they did not, the retrospective must say so
• Lessons-learned identifies at least one criterion that was hard to assess consistently — no evaluation framework gets all criteria right on the first run; naming a criterion that produced inconsistent scores or led to panel disagreements is more credible than claiming the framework worked perfectly
• Reviewer is external to the hiring process — a reviewer who participated in any of the 3 hiring processes cannot give an adversarial view of the framework or the process; the reviewer should encounter the evidence fresh

Common mistakes

• Hiring retrospective without 90-day JD accuracy check — if hires were made within the submission window, the retrospective can note that 90-day accuracy will be assessed at a later date; claiming JD accuracy without 90-day data is not acceptable
• Lessons-learned that only credits the evaluation framework — every first-run evaluation framework has criteria that are hard to assess; a lessons-learned with no improvement suggestions is either dishonest or insufficiently observed
• Reviewer who is evaluating based on the submitter's reputation rather than the evidence — a reviewer who knows the submitter well and trusts their judgment without challenging the evidence does not meet the adversarial standard

Resources

What a verifier looks for

• 90-day JD accuracy: ask the submitter whether any of the 90-day outputs in the JDs turned out to be wrong after the hire started — if the answer is 'no, everything was accurate', probe whether the outputs were specific enough to have been falsifiable
• Criterion consistency: ask which criterion produced the most disagreement in panel debriefs — a confident answer with an example shows genuine retrospective analysis
• Process deviation: ask whether there was any stage in any of the 3 hiring processes where the pre-defined evaluation framework was not followed — and what prompted the deviation

You'll sign in first, then come straight back here.