Your Job Requirements Changed. Your Candidates Were Judged Against the Old Version.

The silent rewrite that corrupts a shortlist

A hiring manager opens a search for a product analyst who can write SQL, explain experiments and work in Berlin twice a week. After six weeks, a new stakeholder asks for enterprise pricing experience and full-time office attendance. Recruiters update the job description, but the candidate scorecards and rejection reasons still reflect the original brief. The shortlist now contains two different selection processes masquerading as one. ^↗

This is not simply a communication problem. It changes the population of people who would have applied, the evidence requested from those who did, and the meaning of every pass rate. If the team quietly adds a hard requirement, previously rejected candidates may deserve reconsideration while already-assessed candidates may need a narrowly targeted follow-up. ^↗

The decision to change a role is legitimate. The mistake is treating a material change as routine copyediting. A practical control is to give every requisition a dated criteria version, an owner and an impact decision before recruiters apply the new rules. This approach is consistent with the job-related, documented assessment strategy described by the U.S. Office of Personnel Management and the fairness considerations in EEOC selection guidance. ^{↗ ↗}

Classify changes by their effect on eligibility

Use three categories. Editorial changes fix wording or formatting without changing who qualifies. Evaluation changes alter how an existing competency is tested, such as replacing a generic interview question with a job-related scenario. Eligibility changes add or remove a mandatory credential, shift the workplace requirement, change compensation bands or move the role to another level. ^↗

An editorial correction can be logged and published. An evaluation change needs updated scoring instructions, interviewer briefing and a check on whether past scores remain comparable. An eligibility change should trigger a fresh decision about the applicant pool, advertisements, candidate communications and legal review where applicable. ^↗

Do not label a new requirement 'preferred' to avoid the change process if interviewers will treat it as mandatory. The real test is behavioral: would a capable candidate who passed yesterday fail today solely because the requirement changed? If yes, the team needs a visible change record. ^↗

A real example: the requirement that arrives in week seven

Suppose 140 people applied under version 1.0. Forty met the published essentials; twelve completed a work sample; four reached final interview. Version 1.1 adds a must-have skill that only two finalists were asked about. It would be invalid to treat 'not demonstrated' as 'does not have it' for the other two finalists. Absence of evidence is a request for assessment, not proof of deficiency. ^↗

The team should first ask whether the new skill is genuinely necessary on day one. If yes, publish a short supplemental assessment to all affected candidates at the same stage. Reopen prior rejections only where the new brief expands rather than narrows eligibility. Document why any candidates are excluded from reconsideration. ^↗

Meanwhile, report funnel conversion by criteria version. A combined chart can conceal the operational consequence of the change: the team may believe screening quality fell when the denominator actually changed. ^{↗ ↗}

The six-field change log

Record the original requirement, the proposed wording, business reason, change class, effective date and candidate impact. Add two approvals: the accountable hiring manager owns the business need; the recruiting lead owns the assessment and communication consequences. If compliance or employee relations are implicated, route that decision to the appropriate specialist rather than asking the recruiter to make a legal judgment. ^↗

Freeze the evaluation rubric for candidates already in a live assessment unless a documented, job-related reason requires a consistent supplemental assessment. Never retroactively rewrite a candidate's old score. Preserve the original evidence, add a dated new observation and show both in the decision record. ^↗

Communications should be specific: 'The role now requires three office days rather than two; please confirm whether this remains feasible' is more respectful and useful than 'Our needs have changed.' A material change may justify reposting the vacancy so the newly eligible audience can apply. ^↗

What to measure after the change

Track the share of active candidates affected, the number who need reassessment, the time to resolve the change, and the proportion of declined candidates who cite the revised terms. These are operational indicators, not a score of a recruiter's competence. ^↗

Audit three sample cases after each change: one candidate who was rejected before the change, one who was in process and one who applied afterward. Can a reviewer reconstruct which criteria applied and why the decision was fair? If not, the process is not yet controlled. ^{↗ ↗}

Start with the last ten requisitions that took more than sixty days. Ask whether the final offer matched the original published role. The gap between the two versions is often more actionable than another generic time-to-fill dashboard. ^↗

Run the operational check