An Intake Meeting Produces a Decision Rule, Not a Wish List

What should the meeting end with?

A one-page document the hiring manager has read and corrected, carrying four things: the outcomes, the requirements, the evidence and the rule. Alignment is not one of the four. Alignment is what people feel at the end of a good conversation, and it lasts about as long as the memory of the conversation does, which is usually until the first shortlist arrives.

What goes on the page, in the order it gets written:

1. Outcomes. Three to five things this person is accountable for by the end of quarter two, in the manager's words. "The monthly close runs without me in it" is an outcome; "supports the finance team" is a duty and tells the loop nothing. 2. Requirements. Four or five, each already matched to the stage that will produce evidence for it. Anything that cannot be matched is a preference, and preferences go in a separate paragraph so nobody mistakes one for a bar later. 3. Evidence. For each stage, what the interviewer will come back with in writing: a sentence about what the candidate did, said or produced, in the detail a rating cannot carry. It is the line most intake agendas leave out. 4. The rule. How the evidence combines into a decision, and who breaks a tie by name.

The evidence line is worth defending against the manager who thinks it is bureaucratic. Text-mining post-interview notes on 7,650 candidates hired at a large Chinese internet technology company, researchers found that the number of job-related capabilities an interviewer named in their notes was positively related to the candidate's later job performance and promotions and negatively related to turnover ¹. The effect is small, roughly 2 percent of a performance measure per standard deviation, it is correlational, and it could only be measured on people who were hired. But notes that name the capabilities the job called for carried signal that a rating alone did not, and what the intake page defines is exactly which capabilities the notes should be naming.

Write the decision rule while the manager is in the room

A decision rule says how the pieces of evidence combine and what happens when they conflict. At its plainest: all four requirements demonstrated is a yes, three of four is a discuss unless the gap is the requirement the role turns on, two or fewer is a no, and if the panel splits the hiring manager decides. Write it in the meeting, in that much detail, or it will be invented at the debrief by whoever speaks last.

The reason to insist is that the debrief is not a neutral place to invent one. In Kuncel and colleagues' meta-analysis, the average correlation with job performance was .44 when applicant data were combined mechanically by a formula and .28 when the same data were combined by expert judgment, and the authors classify group consensus meetings as expert judgment: what makes a method holistic is that evidence is combined by insight rather than by a rule applied the same way every time ². Their derivation puts the cost at up to a 25% reduction in correct hiring decisions in a moderately selective scenario.

That comparison rests on nine studies for job performance, the 25% is derived from those validities and was never measured at an employer, and the consensus meeting is a category in the authors' taxonomy that they did not test on its own. Hold it loosely. Two things in it are not loose. A mechanical rule can be as plain as adding up a written scorecard, which the paper says outright. And none of this argues against holding a debrief: the meeting is how evidence gets surfaced and challenged, and what falls down is only the final step, when a conversation ends in a shared feeling instead of a stated rule. What a debrief decision rule looks like written out is the downstream half of this page.

The tie-breaker is the part managers resist and the part that pays. Naming it costs one sentence at intake and saves the week where three people re-litigate a candidate because nobody agreed in advance who owns the call.

Why deferring the bar until the resumes arrive stopped working

Because the pile no longer varies the way it used to. The old deferral had a logic: define the role loosely, look at fifteen applications, and let the bar become obvious by comparison. That worked while applications differed from each other in ways that tracked the candidates. It works far less well once the posting is the brief an application gets assembled against, and the requirement list comes back to you confirmed.

So the comparison that used to reveal the bar now mostly reveals which candidates read the posting closely. The bar has to be set before the pile arrives, which means intake is carrying a job it was previously able to postpone, and the meeting has not changed to absorb it.

The evaluation side is where processes are thinnest to begin with. The US Office of Personnel Management's practical guide defines a structured interview by three properties: every candidate gets the same questions in the same order, every candidate is rated on a common scale, and the interviewers agree in advance on what an acceptable answer looks like ³. The first two are built by whoever writes the scorecard. The third is a decision about the job, and there is no meeting other than intake where it can be made.

That guide is federal practice guidance from 2008, written for US government hiring, and it binds no private employer. It earns its place here for the definition rather than the authority: agreeing in advance what an acceptable answer looks like is the intake page's work, and it is the component most likely to be skipped, because skipping it feels like flexibility right up until the debrief.

The intake page is where the evaluation side gets specified, because it is the only meeting in the process that happens before anybody has a candidate to argue about. Two inputs make it much better and neither needs another meeting on the calendar: fifteen minutes with the person currently doing the work, and an honest read on what AI actually does in this role day to day before writing a requirement about it. The second one increasingly determines whether the requirement list describes the job as it is or as it was two years ago.

Run it in forty minutes, then send the page back to be corrected

Forty minutes of the manager's time, one page written inside two hours, then sent back with a specific request: correct anything wrong, and reply. A bar the manager has not read is a bar the recruiter invented, and it will be disowned at exactly the moment it matters, which is the first strong candidate who fails on the requirement nobody re-read.

The length should vary with how new the role is, and only that. A role the manager has hired for three times needs twenty minutes and a page that mostly says what changed. A first-of-kind role needs the forty minutes and a deliberate second session after the first five screens, because the first five screens are what reveal which requirement was imaginary. Book that second session at the same time as the first, or it will not happen.

The time is available. SHRM's 2026 benchmarking brief, built on data from over 4,600 organizations, puts median time-to-fill at 39 calendar days for nonexecutive positions ⁴. That is a median across organizations that answered a survey rather than across hires, from a US HR membership association, and time-to-fill has no shared definition across sources, so it is not a benchmark to compare your own number against. It is enough to make the point: against a search measured in weeks, forty minutes at the front is not where the time goes.

Test this against a search that has already closed. Ask two people from that loop to state the decision rule from memory. If the answers differ, or if one of them is a description of the ideal candidate rather than a rule, the meeting produced a wish list. The repair is one page and one reply, and the argument it prevents is usually the argument about who decides what qualified means.

See the benchmarks >