PBH Talent Hub →
Sourcing Channel Analysis · Performance-based Hiring

Relationships matter when it comes
to better hiring

The sourcing channel determines the quality of your candidate pool before a single resume is reviewed. Relationship-based channels arrive pre-screened — by people who have seen the work firsthand. Stranger-based channels rely on filters that can barely tell the strong from the weak.

Why relationship channels work

When you know the person, you already have evidence on all four quadrants

Hiring strangers forces you to infer all four dimensions from credentials and keywords — signals with almost no predictive power. Hiring people you know means you already have direct evidence on each quadrant before the process begins.

Quality of Hire oval — Use Performance-based Hiring to Predict Quality of Hire. Four quadrants: Fit (top-left), Motivation (top-right), Ability (bottom-left), Adaptability (bottom-right), surrounding Performance / Quality of Hire at center. Success = Ability × Fit × Motivation to the n × Adaptability
Fit — you've seen how they work with others, their pace, their culture alignment. Not guessed from a personality test.
Motivation — you've observed their drive, initiative and grit under real pressure. Not estimated from an interview question.
Ability — you've seen their technical skills and problem-solving on actual work. Not inferred from a resume keyword.
Adaptability — you've watched them handle ambiguity and change in real situations. Not predicted from a cognitive test.

The more quadrants you have direct evidence on, the higher the implied diagnostic ratio — and the higher your confidence in the hire before the process even begins.

The single metric that defines sourcing channel quality
The Diagnostic Ratio
P(Signal | Strong) ÷ P(Signal | Weak)
Hit rate on strong candidates ÷ false alarm rate on weak ones
What it measures

How much more likely a signal is to appear when the candidate is genuinely strong versus when they are not. A ratio of 1.0 means the signal fires equally on both — it tells you absolutely nothing. The higher the ratio, the more the signal discriminates between strong candidates and weak ones.

Why it defines sourcing channel quality

The diagnostic ratio of a sourcing channel is determined by the quality of evidence available about the candidate — not by how good your screening tools are. Watching someone do the job for two years is evidence with a ratio near 8–10x. A keyword match is evidence with a ratio near 1.1x. No screening tool applied afterward can close that gap.

The ratio for each evidence type
Direct management 8–10x
Trusted referral 4–5x
Achiever language on resume 5.6x
Keyword / credential match 1.1x
Cold inbound / no signal ~1.0x
Sourcing channel Prior Diagnostic ratio Post-screen confidence What the "screen" actually is
Relationship-based — direct evidence on ability, fit, motivation & adaptability
Internal promotion
Someone you've managed directly
75%
8–10x
96% confidence
Direct observed evidence across all four performance quadrants. The hiring decision is largely a confirmation.
Re-hire / boomerang
Former employee, performance on record
70%
7–9x
94% confidence
Historical performance record across ability, fit and motivation. No inference needed.
Prior direct report
Managed them at another company
68%
6–8x
93% confidence
Firsthand management evidence. Saw the work, the growth trajectory, the response to challenge.
Prior colleague (peer)
Worked alongside, observed output
58%
5–6x
89% confidence
Strong evidence on ability and fit. Motivation and adaptability observed but less directly.
Employee referral
Vouched by someone trusted inside
50%
4–5x
83% confidence
One trusted person's judgment — not your own observation. Still far above any credential screen.
Talent pipeline
Nurtured before the role opens
40%
3–4x
73% confidence
Relationship built on performance criteria. Self-selected based on the work, not the credential.
Mixed — some signal, but engagement barrier applies
Targeted direct sourcing
LinkedIn / SeekOut / JuiceBox — performance-based message
25%
~2.0x
40% confidence
Best candidates evaluate the message quality. Generic outreach gets deleted. Response rate 5–10%.
Stranger-based — screened on credential proxies, no prior evidence on performance
Job posting
Inbound / job boards — open to all
15%
1.1x
17% confidence
Open to anyone. Strong candidates already employed often don't apply — the best self-select out.
Cold outreach / blind inbound
No prior connection or signal
10%
~1.0x
10% confidence — no change
A 1.0x diagnostic ratio means the screen adds zero information. Confidence after screening = confidence before.
Relationship channels
avg post-screen confidence
88%
Stranger channels
avg post-screen confidence
14%
Confidence gap
after all screening
6.3×

The sourcing channel is the screen. Everything else is noise reduction.

Most companies try to solve hiring quality by improving their filters — better keyword screens, cognitive tests, personality assessments, AI interviews. The data shows this is the wrong lever. These tools screen strangers using signals that appear at nearly the same rate in strong candidates and weak ones alike. When a signal can't tell the difference, it adds almost no confidence to the hiring decision.

This is what the diagnostic ratio measures — how much more likely a signal is to appear in a strong candidate than a weak one. A ratio of 1.0 means the signal fires equally on both groups and tells you nothing. A ratio of 5x means it is five times more likely to flag a strong candidate. Keyword filters and credential matching have diagnostic ratios near 1.1x. Watching someone do the work for two years has a ratio near 8–10x. That gap is the whole story.

Relationship-based channels bypass the credential proxy problem entirely. When a manager has seen the work firsthand, when a colleague vouches for performance, when a re-hire returns with a track record — the screening has already happened, on the job, against real outcomes. The pool arrives pre-qualified. That is why an internal promotion enters the process at 75% prior confidence and exits with 96% post-screen confidence — before a single structured interview is conducted.

Stranger-based channels start at 10–15% and keyword screening moves them to 17%. The same structured interview that confirms a known candidate at 97% can only get a stranger-sourced candidate to 26% — because the prior the interview receives is so much weaker. A good interview cannot rescue a bad pool. Fix the pool first.

Performance-based Hiring · Talent Hub

Tools built on the
methodology behind this analysis

The Performance-based Hiring Talent Hub gives you the frameworks, job description templates, screening tools, and interview guides to put this approach into practice — across every sourcing channel.

Visit the PBH Talent Hub →