Hiring System Health™

The standard way organisations track hiring is by outcome: how long did it take, did we make an offer, did the person accept, did they stay?

You cannot fix what you cannot see. Most hiring systems are invisible to the people running them.

These are lagging indicators. By the time they register a problem, the problem has been running for weeks or months. The search that took 22 weeks failed at week 4. The hire that left at 18 months was wrong at offer stage. The outcome is the last thing to change. The system failed long before the outcome did.

Hiring System Health reframes the question. Instead of measuring outcomes after the fact, it tracks the health of the system in real time, across three dimensions, each with observable indicators, each capable of producing an early warning before the outcome fails.

The three dimensions

```

PIPELINE HEALTH

↓

[Is the funnel filling?

Are the right candidates entering?

Is response rate holding?]

PROCESS HEALTH

↓

[Are stages moving?

Are decisions being made?

Where is velocity dying?]

DECISION HEALTH

[Are evaluations calibrated?

Is bias being managed?

Is the right person in the room?]

```

A mandate can fail at any one of these dimensions while appearing healthy in the others. A pipeline that fills with the wrong candidates looks healthy until the shortlist collapses. A process with fast stage movement looks healthy until every candidate drops off at offer. A decision that is made quickly looks healthy until the hire fails at six months.

Health requires all three dimensions, tracked continuously.

Dimension 1: Pipeline Health

Pipeline health measures whether the right candidates are entering the funnel at the right volume and rate.

Volume: Is the number of candidates entering the pipeline consistent with what the market should produce for this role? A VP of Engineering search in a major market should be generating 40–60 substantive conversations within the first three weeks. Significantly below this signals either a sourcing problem (not reaching the right people) or a positioning problem (reaching them but not converting the contact into a conversation).

Quality: Are the candidates entering the funnel actually relevant to the brief? Volume without quality is a processing burden that masks a targeting failure. The right metric is not how many candidates were contacted; it is how many of the candidates contacted were genuinely in scope.

Response rate: What percentage of outreach is generating a response? Response rate is the most sensitive early indicator of a positioning problem. A declining response rate, across channels and over time, signals that the message is not connecting, that the candidate pool has seen this type of outreach too many times, or that the opportunity is not being framed compellingly. Response rate that drops below 15% is a warning signal. Below 10% is a failure signal.

Decay rate: Is response rate stable or declining week over week? A stable rate at a low level is a different problem from a rate that started well and is falling. Decay signals that the initial outreach worked on the most accessible candidates and that the remaining pool requires a different approach.

Where this dimension fails: When the search is measured by volume of outreach rather than quality of response. When response rate is not tracked. When the same message is sent to the same pool multiple times without modification. When sourcing continues at volume rather than pausing to fix positioning.

Health check question: What is our week-3 response rate, and is it higher or lower than week 1?

Dimension 2: Process Health

Process health measures whether candidates are moving through stages at appropriate velocity, and whether the right decisions are being made at each gate.

Stage velocity: How long does each stage take? The specific acceptable duration varies by stage and by role, but the principle is consistent: every day a candidate spends waiting for a next step is a day in which a competing opportunity can emerge, interest can cool, and the candidate's perception of the company's operational effectiveness is being updated in a negative direction. For senior candidates, a response time of more than 72 hours between stages is a signal that the company is not prioritising the search.

Drop-off by stage: Where are candidates leaving the process? Drop-off at early stages is a brief or positioning problem. Drop-off at interview stages is an evaluation or candidate experience problem. Drop-off at offer stage is a compensation, scope, or relationship problem. Each location produces a different diagnosis and a different fix.

Decision lag: How long between interview completion and decision communication? Decision lag is the most common and most underestimated driver of candidate loss at late stages. A candidate who has completed three rounds of interviews and is waiting ten days for a decision is actively reconsidering their interest, accepting other interviews, and updating their view of what it would feel like to work for this organisation.

Hiring manager availability: Is the hiring manager present in the process at the frequency the search requires? A search where the hiring manager is available for one conversation per week cannot close in less than eight weeks, regardless of how good the sourcing is. Availability is a resource constraint that most searches don't model explicitly.

Where this dimension fails: When velocity is not tracked per stage. When drop-off data is not collected. When decision lag is normalised because the hiring manager is busy. When candidates are left in limbo between stages without proactive communication.

Health check question: What is the average number of days between stage completion and next-step communication, and how many candidates dropped off at each stage?

Dimension 3: Decision Health

Decision health measures whether the evaluations being made inside the process are calibrated, bias-managed, and structured to produce an accurate assessment of the candidate.

Calibration: Are the people making decisions using the same criteria? Calibration failure is the most common decision health problem: different interviewers are evaluating different things, with different implicit standards, and no structured way to aggregate their assessments. The result is a committee conversation that replays everyone's individual impressions without converging on an evidence-based view.

Bias indicators: Which candidates are generating strong subjective reactions, positive or negative, early in the process, before structured evaluation? Strong early reactions are usually driven by familiarity (this person is like people I've worked with and trusted), presentation (this person communicates in the style I find impressive), or pedigree (this person went to the right school or worked at the right company). These are bias signals, not capability signals.

Evaluation structure: Is there a structured scorecard? Are criteria defined before the first interview? Is the evaluation completed independently before committee discussion, or is the committee discussion doing the evaluation? The order matters: group discussion before independent evaluation produces anchoring and conformity, not independent assessment.

Reference quality: Are references being used as a genuine intelligence source? A reference call that consists of "would you rehire this person?" is not a reference call; it is a formality. A reference call that investigates the specific conditions in which the candidate has thrived and struggled, the quality of their relationships with peers and reports, and the specific ways their performance evolved over time: that is intelligence.

Where this dimension fails: When there is no structured scorecard. When interviewers debrief together before completing independent assessments. When reference calls are conducted after the decision has already been made. When "culture fit" is the deciding criterion at the final stage, the stage at which bias is most likely to operate.

Health check question: Can every person involved in the final decision articulate, in writing, the specific evidence that drove their assessment, independent of what anyone else in the room said?

Running a health check

The health check is not a retrospective. It is a real-time diagnostic, run at the end of each week during an active search.

Three questions, one for each dimension:

1. Pipeline: What is this week's response rate, and what does the trend look like?

2. Process: What is the average stage velocity, and where is drop-off occurring?

3. Decision: Are evaluation criteria consistent across interviewers, and are assessments being made independently?

A search that can answer all three questions with data, not impression, is a search that is being managed. A search that cannot is a search that is hoping.

Most executive searches are hoping. The ones that close in 50 days are managing.

See also: Compounding Failure Loop™, Failure Prediction System™, Hiring SLO Framework™

Sources

Google: Site Reliability Engineering, Monitoring Distributed Systems

McKinsey: Attracting and Retaining the Right Talent

SHRM: HR Metrics and Analytics, Using Data to Drive HR Strategy