Performance-based Hiring · Diagnostic Ratio
The Evidence Meter — what actually moves the odds.
Every piece of evidence carries a Diagnostic Ratio (DR): how differently it falls for a strong performer versus a weak one. High-DR evidence moves the meter. Low-DR evidence doesn't — no matter how convincing it feels. Set a starting point, then stack evidence and watch.
Can this person do the job?
0255075100
noise
promising
mounting
win-win zone
The source sets two limits — a floor and a ceiling. A stranger starts near 8% and, doing everything right, tops out around 68%. A trusted referral runs ~30% → 78%. Someone you've worked with: ~45% → 85%. Switch channels and watch both the start and the ceiling markers move — a stranger can never reach a former colleague's confidence, however strong the evidence, because you can't verify a stranger the way you've watched a colleague work.
§ 01Starting odds · the sourcing channel
§ 02The evidence · click to stack it
High DR — moves the meter
Evidence about the work itself — what a weak performer can't fake. (Who vouches for them is already in the starting odds, so it's not repeated here.) Each one multiplies the odds.
Low DR — feels convincing, moves nothing
A strong and a weak candidate both produce these. Stack the whole pile — the needle barely twitches.
Evidence mounts; impressions don't. The first impression has a DR of about 1 — click it all day and the meter will not move. The only evidence worth gathering is the kind that tells a strong performer apart from a weak one.
How the meter works — Bayesian updating, bounded by what you can verify. The sourcing channel sets two limits, not one: a floor (starting odds — stranger ≈ 8%, referral ≈ 30%, worked-with ≈ 45%) and a ceiling (the most confidence that source can ever earn — stranger ≈ 68%, referral ≈ 78%, worked-with ≈ 85%). Each piece of evidence is a Diagnostic Ratio (DR) that raises the odds toward that ceiling; a DR of 1 can't move them. Two reasons the ceiling differs by source and the climb flattens: the evidence is correlated (track record, promotions, trajectory all measure the same capability, so each adds less than the last), and a stranger's evidence is mostly self-reported and unverified — you've never watched them work — so it can't carry you as far as direct experience can. That's why a stranger doing everything right still tops out below a former colleague, and why certainty is never reached.
Put the method to work
See where your own hiring odds stand, and get the tools that move them.
Stay in the loop
Leave your details and we'll keep you posted on new Performance-based Hiring tools and ideas.
You're in.
The Tools Library and Kaizen Audit links above are open to you now — dive in.