The typical reference call follows a pattern that most experienced hiring managers could recite in their sleep. You call the reference, exchange pleasantries, ask whether the candidate is hardworking and collaborative, confirm that you "would definitely recommend her," ask whether she is eligible for rehire, and end the call. The whole thing takes eight minutes. You mark the reference as completed in your ATS and move on to preparing the offer letter.
This process generates essentially no actionable information. It is a compliance ritual — something done so that the "references checked" box is ticked — rather than a genuine attempt to gather evidence about a candidate's likely performance. The problem is not that references refuse to share meaningful information. It is that unstructured reference calls, like unstructured interviews, do not ask questions that generate meaningful information.
Structured reference checks — with anchored behavioral questions, a consistent question set across all references for the same candidate, and a scored response format — are a different category of process. They take longer. They require more preparation. And they surface patterns that the interview loop itself, by design, cannot surface: how the candidate's behavior is perceived by someone who has managed or worked alongside them for months or years.
Why "Would You Rehire?" Is a Low-Signal Question
The rehire question has become a hiring industry shorthand for reference validity — if someone says yes, the reference is positive; if they hedge, the reference is soft. But as a signal, it is deeply confounded. References are almost always provided by the candidate. They have been pre-selected for positivity. In this context, a "yes" to the rehire question carries almost no information, because the base rate of "yes" among references that candidates choose to provide is very high — plausibly above 85% across professional roles.
The question you actually want answered is not "would you rehire her" but something more like: "Tell me about a situation where her work had a significant impact on an outcome your team cared about — and a situation where her approach created friction or required course correction." The structured version of that question will sometimes produce a "would not rehire" outcome from a reference who opened the call prepared to be positive, because the behavioral frame reveals something the candidate and reference both would have passed over in the abstract.
Employment verification (VOE) — confirming dates of employment, title, and whether separation was voluntary — is the foundational layer of a reference check and should be completed for all final-stage candidates regardless of the structured reference process. VOE typically occurs through a 180-day verification window process or through direct employer confirmation. It is distinct from a substantive reference call but should happen before, not instead of, meaningful reference conversations.
Building a Structured Reference Question Bank
A structured reference check uses the same behavioral question framework as the interview itself: behavioral, past-tense, specific, with follow-up probes designed to move the reference from vague summary to specific example. The difference is that the reference has access to a longer observation window and can speak to patterns — not just isolated incidents — across the full tenure of working with the candidate.
Useful behavioral question categories for reference calls include: how the candidate responded to critical feedback, how they performed under resource constraints or competing priorities, what the specific nature of their contributions was versus their team's (the attribution question), what they struggled with in their role, and — for leadership roles — what their direct reports thought of their management approach.
Each of these questions is designed to produce behavioral evidence rather than impressionistic summary. "She was really strong under pressure" is not behavioral evidence. "We had a Q3 where we lost two engineers mid-sprint and had to renegotiate three deliverables with clients. She ran point on all three client conversations and none of them escalated. I'd say she was better in crisis than in steady-state" — that is behavioral evidence. The structured reference question does not just ask about behavior in the abstract; it asks the reference to recall and narrate specific situations.
Scoring Reference Responses: The Anchored Approach
Most reference check processes stop at note-taking. The structured approach goes one step further: scoring reference responses against the same competency model used in the interview loop, using a simplified version of the BARS anchor framework.
This does not require the full 1-5 anchor set from the interview rubric. A three-point scale works well for reference scoring: the response demonstrates the competency clearly (3), the response is ambiguous or mixed (2), the response raises concerns (1). For each competency, the reference checker should record both the score and the specific evidence that drove it.
When multiple references are scored on the same competency, the pattern across references is more informative than any single reference's response. A candidate who receives a 3 on "handling critical feedback" from one reference and a 1 from another has a pattern worth exploring. A candidate who receives a consistent 2 on "cross-functional stakeholder communication" from three independent references is telling you something the interview loop may have missed.
A Scenario: Reference Checks That Changed a Hire Decision
An early-stage B2B software company was finalizing a hire for a Head of Customer Success — a role that required both client relationship management and internal process building. The candidate had performed exceptionally well through a four-stage structured interview loop, scoring above threshold on every competency. References were collected as a formality.
The hiring team ran structured reference calls with three references the candidate had provided: two former colleagues and one former manager. The interview loop had included a behavioral question on "navigating internal resistance to new process" — a critical competency given the role's requirement to build a new CS function from scratch. The candidate had provided a strong, specific behavioral response that scored 4 on the rubric.
In the reference calls, the former manager's response to an equivalent question was notably different. She described a situation where the candidate had introduced new process requirements without adequate buy-in from adjacent teams, which had created a "two months of cleanup" situation. She framed it constructively — "she learned from it" — but the behavioral description mapped to a consistent pattern across two of the three references. When the hiring team overlaid the reference responses on the scorecard, they decided to extend the loop with an additional reference and a more specific follow-up question on the process change scenario before proceeding.
The additional reference confirmed the pattern, though not in a disqualifying way — it provided context about the candidate's growth trajectory that the references were willing to discuss once asked specifically. The hire went forward, but with a structured onboarding plan that explicitly addressed the pattern the reference checks had surfaced. That is the practical value of a structured reference process: not necessarily changing hire decisions, but producing information that shapes how you bring someone on.
Operational Realities and Legal Considerations
Structured reference checks take time. A properly conducted behavioral reference call with probes and follow-up typically runs 25 to 35 minutes. Scoring and documenting the response adds another 10 to 15 minutes. For a senior role where three references are being checked, this is a meaningful time investment — and it should be made by the hiring manager or a senior team member, not delegated to a coordinator running through a checklist.
From a legal standpoint, reference checks are generally permissible when conducted consistently — asking the same substantive questions for candidates in the same role cohort reduces disparate treatment risk. Reference providers are generally protected from defamation claims when making good-faith responses about former employees, and many states have qualified privilege statutes that extend protection to reference givers who provide information honestly and without malice. That said, reference checks should focus on job-related performance dimensions and should not elicit information about protected characteristics — family status, medical history, or similar matters.
The reference check is the final stage of the evidentiary record. When it is conducted rigorously — structured questions, scored responses, patterns mapped against the interview scorecard — it closes the loop between what a candidate says about themselves and what the people who worked alongside them say about the same behaviors. That gap is often where the most useful information lives.
