ATTENTION, FOUNDERS WHO JUST MET A ‘CULTURE FIT’ OVER COFFEE!
PEOPLE WHOSE ENTIRE DILIGENCE PROCESS WAS “WOW, WE REALLY CLICKED”! PEOPLE ABOUT TO BUY A LAPTOP, A COMPANY EMAIL ADDRESS AND A NEW SET OF EXPECTATIONS FOR SOMEONE WHOSE LAST NAME THEY LEARNED TWENTY MINUTES AGO!
Are you tired of not finding out whether a hire works until after the email address has been issued?
Have you left a coffee genuinely certain — the good kind of certain, the kind that hums — and then been completely unable to say what the person would actually be doing on a Wednesday?
Then congratulations.
You qualify for the Five Number One Rules of Hiring Someone You Will Still Want to Hire — five gates, five Number Ones, and no coffee-shop chemistry meter.
“Wait,” you’re shouting into the applicant-tracking system, “they can’t all be Number One!”
That’s exactly what Big Numbering wants, while it sells a dazzling interview as a substitute for knowing what the job is.
Hiring failures are specialists.
A role mission can’t turn a friendly conversation into comparable evidence.
A scorecard can’t independently verify a candidate’s story.
A glowing reference can’t prove future performance.
And no selection process can audit its own prediction unless somebody goes back and looks after the start date.
Five gates. Five different escape routes for wishful thinking.
So here they are.
Five rules.
Five Number Ones.
No substitutions. No “but they seemed sharp.” No hiring a horoscope with a LinkedIn profile because the vibe turned up in clean shoes.
RULE #1: WRITE THE MISSION BEFORE YOU MEET THE PERSON
Introducing ROLE-FLAVORED-PEOPLE™, the recruitment system that starts with a candidate and then searches the building for a problem shaped exactly like them.
Includes the bonus sequel: “WE CAN FIGURE OUT WHAT THEY DO LATER!”
Write the role mission before anyone sources a single candidate. One or two sentences. State the business result that explains why the role exists — not a shopping list of duties.
“Grow enterprise-segment ARR from $4M to $9M in 24 months” has a direction, a destination and a clock.
“Manage enterprise sales” has a beige cardigan and no point of contact with reality.
Then derive three to eight outcomes from that mission. Each one specific, measurable, dated.
“Hire and ramp four AEs to full quota by month 9” is an outcome. “Build the team” is an attractive fog bank.
The mission supplies the reason. The outcomes make its promises inspectable.
Only then list competencies — five to ten that this role genuinely requires. Published reference lists run to about fifty stock items, and that’s a shelf to pick from, not a menu requiring every applicant to become a fifty-tool Swiss Army executive.
Do it in the other order and you get the golden hammer in a blazer: existing expertise quietly substituting for an honest look at what the job needs. Kaplan called it the law of the instrument; Maslow gave us the version about everything looking like a nail.
Hire for an outcome before admiring a person-shaped solution.
A role without a mission doesn’t have a hiring process. It has a guest list.
RULE #1: SCORE EVIDENCE AGAINST OUTCOMES
NOW AVAILABLE: TRAIT-PARFAIT 7000™!
Layer confidence, “energy,” and one very polished anecdote into a tall glass. No outcome required. No evidence written down. Just spoon directly from the feeling.
Turn the mission, outcomes and competencies into a scorecard before the interview.
The scorecard is not an after-dinner spreadsheet built to justify a decision already made. It’s the standard against which the team records what evidence each candidate supplied, for each outcome.
The order does real work here.
Start from generic competencies and you produce a familiar document that holds up nothing at all: leadership, communication, ownership, and possibly “strategic,” set in a font large enough to frighten the finance team.
Start from the role outcome and you’re forced into the question that matters. What has this person actually done that resembles the result now required — under what conditions, and by when?
That’s also how you separate an assertion from evidence. A candidate can say they built a team. The scorecard asks whether they hired and ramped the specified number, hit the result, and did it on the timeline.
It guards against the debrief drifting, too. Parkinson’s committee spent forty-five minutes on a bicycle shed because everybody felt qualified to comment. A hiring debrief can do exactly that with a handshake.
And use the same scorecard for everybody. Don’t add an outcome because a charming candidate happens to have achieved it. Don’t quietly drop one because your favourite hasn’t.
A scorecard is a promise to judge the work, not the feeling the conversation left behind.
If the criterion appears after the favourite appears, it isn’t a criterion. It’s a receipt.
RULE #1: ASK EVERY CANDIDATE THE SAME QUESTIONS
From the manufacturers of “LET’S JUST CHAT” comes CONVERSATION-PLUS™ — the thrilling process where every interviewer follows whichever thread feels most natural, and every candidate gets compared against a completely different weather system.
Run a structured interview. Identical sequence of job-relevant questions for everyone, evaluated against the scorecard you already wrote.
One good version walks chronologically through roughly the last ten years. At every job: what were you hired to do, what accomplishments are you proudest of, what were the low points, who did you work for and how would they rate you, and why did you leave?
That sequence isn’t a personality test in a necktie. It makes evidence comparable.
Without it, one candidate gets an hour on achievements while another gets a free-form tour of hobbies, office dogs, and the one conference where everybody agreed with them. Then you compare them, confidently, as though you’d learned the same things.
The validity numbers make this too large to wave off as style.
The 1998 Schmidt and Hunter table put structured interviews at .51 and unstructured at .38. A 2021–2022 re-analysis by Sackett and colleagues revised those to .42 and .19, noting most prior estimates drop by .10 to .20 once a range-restriction correction problem is addressed.
The estimates moved. The ranking didn’t, and the gap got wider. Structured interviews came first among the individual procedures reported.
This is where cargo culting turns up in a company hoodie — the panels and the calendar invitations, without the fixed questions or the common criteria that made any of it work. Feynman’s warning was about scientific form kept while the disconfirming discipline was dropped.
Fair comparison begins before the first answer, when the questions become the same.
Chemistry can be a data point. It cannot be the measuring instrument.
RULE #1: VERIFY THE STORY WITH STRUCTURED REFERENCES
BEHOLD REFERENCE-FAIRY EXPRESS™: type “excellent communicator” into a form, receive a perfumed cloud of praise, and never once ask whether the account agrees with anybody who sat near it.
Satisfaction not guaranteed. The adjectives, however, are magnificent.
Use structured references as a consistency check. That framing matters, and I’ll come back to why.
In the TORC approach, the candidate is told before the offer that they’ll arrange calls between the hiring manager and former managers, peers and subordinates from recent roles.
Ask every reference the same questions. Biggest strengths. Development needs. Performance on a 1–10 scale. Would you rehire.
Then compare those accounts against the chronological interview. Somebody who can’t produce willing references — or whose references contradict the story — has supplied a signal that no amount of warm final conversation can manufacture away.
Now the caveat, because this gate gets oversold constantly.
Schmidt and Hunter put reference checks at .26, among the weaker predictors in that table. And that estimate rests on studies conducted before the current legal climate — by 1998, 19 of 50 U.S. states had good-faith-reference immunity statutes with 9 more considering them. A current employer may confirm only dates and title.
TORC doesn’t cure any of that. What it changes is the mechanism: candidates arrange references who have some incentive to cooperate, rather than asking a legal department to volunteer candour to a stranger.
So this is not an independent prediction of future performance at anything like the strength of a structured interview. It’s a test of whether the story survives contact with other witnesses. Sell it as more and it will disappoint you.
One doctoral thesis following the method across six companies and roughly 1,000 hires reported mis-hire rates falling from 69.3% to 10.5% — a single unpublished dissertation about a proprietary method, which is worth exactly what that description suggests.
A reference does not predict the future. It tests whether the past has one version.
The point isn’t to find a perfect witness. It’s to stop hiring on an unchallenged monologue.
RULE #1: AUDIT THE HIRE AT 30, 60 AND 90 DAYS
AND NOW: SET-AND-FORGET EMPLOYMENT™!
Make a decision once. Defend it forever. Treat every mismatch as an exciting opportunity to redefine “working out.”
Use the first 90 days to audit the hiring decision, not just to manage the person.
At 30 days: does the role match what the interview described? What surprised them? Where are the early knowledge gaps?
A large gap there is evidence about your scorecard and your interview. It is not an indictment of the person who believed you.
At 60 days: resources and fit. Workload, tools, training, and whether the daily work visibly connects to the original mission. This is where a mismatch between stated outcomes and actual work usually surfaces.
At 90 days: relationships, comfort raising problems, what they’d change. By then you’re closer to a retention signal than a hiring diagnostic.
Then go back to the original prediction. Did the mission describe the work? Did the outcomes identify the results that mattered? Does the evidence you recorded in the interview correspond to what you can see now?
Audit the process as relentlessly as the hire.
Because the thing you’re fighting is escalation of commitment, and it’s well documented. In Staw’s 1976 experiment, people who made the initial funding decision put more resources in after negative feedback than people evaluating the same scenario without having chosen it.
The mechanism is self-justification. Continuing protects a reputation long after it stops protecting the organisation.
Where direct observation is the only defensible test, a job tryout runs against a pre-established standard for a fixed period — Schmidt and Hunter report .44 operational validity. Supervisors erode it constantly, by lowering what counts as “minimally acceptable” rather than enforcing the bar they set.
Which is the same failure in a different room.
The hiring decision is not complete when the person starts. It is complete when the prediction meets reality.
Don’t lower the bar to save the decision. Save the next decision by keeping the bar visible.
BUT WAIT, THERE’S MORE!
“What about hiring the first salesperson?”
Same rules. Write the business result before admiring the network. Score comparable outcomes. Ask the same questions. Check the account. Audit the territory and results at 30, 60 and 90.
“What about an executive who needs confidentiality?”
Same rules. The reference mechanism changes and local law constrains it. The architecture doesn’t: mission, comparable evidence, fixed questions, corroboration, later audit.
“What about a junior or trade role where a portfolio says very little?”
Same rules. Direct observation matters most here — and the standard stays pre-set precisely so later discomfort can’t redraw it.
“What if the team has a very strong gut feeling?”
Same rules. Record the feeling. It still doesn’t replace an outcome, a common question set, corroboration, or observed results.
The details change.
The architecture doesn’t.
THE FIVE, WITHOUT THE LATTE
Write a one- or two-sentence mission before sourcing, derive three to eight dated outcomes from it, and list competencies last.
Build the scorecard before the interviews, and record evidence against outcomes rather than selecting traits once a favourite emerges.
Ask every candidate the same job-relevant questions, walking the same chronological ground.
Run structured references as a consistency check — and don’t oversell them as independent prediction.
Audit at 30, 60 and 90 days against the original scorecard, keeping the pre-set bar visible.
DO THIS BEFORE THE NEXT COFFEE
Open the role that’s about to be hired — or the one already producing a troubling fog of “we should probably talk.”
Write its mission in one or two sentences, as a result rather than a duty.
Derive three to eight outcomes with deadlines. Name competencies only after those exist.
Put the scorecard in front of every interviewer before anybody meets anybody.
Write the fixed questions now: hired to do what, proudest accomplishments, lowest points, manager rating, reason for leaving.
Decide how an answer becomes evidence against each outcome before a favourite candidate gets the chance to make your rubric mysteriously flexible. That flexibility feels like open-mindedness in the moment. It is not.
Write the structured reference questions. Decide, within your local law and practice, who can be asked.
Then put 30, 60 and 90 days in the calendar tonight, while it still feels unnecessary.
At each one, bring back the original scorecard and look for the gaps — role description, missing resources, mismatched work, results.
The aim isn’t to turn hiring into a courtroom drama staged around a stapler. It’s to replace charming uncertainty with evidence that survives comparison and time.
The person deserves a role described honestly. The team deserves a decision it can inspect. And the company deserves better than discovering its entire selection system was one enthusiastic latte.
For the low, low price of writing down what success means before somebody makes a good impression, the complete Five Number One system is yours.
No annual plan. No ceremonial culture deck. No tiny bronze interview bell.
Just five gates that refuse to cover for one another.
Operators are no longer standing by.
The operator is whoever has to work with them in March.