Hiring · The full process

How to build a reliable hiring process.

Most hiring advice is a thin checklist or a discursive essay. This is a named, repeatable system that works for any role — the 4 Pillars of Talent Selection and the five interviews that actually predict performance.

By Lucas Price, founder of Yardstick. He scaled the sales organization at Zipwhip from under $1M to over $100M ARR, which was acquired by Twilio.

Why it is worth the work

Structure beats gut feel — measurably.

Hiring is one of the highest-variance things a team does, and most of it runs on gut feel. Leadership IQ’s study of more than 20,000 hires found that 46% fail within 18 months — fired, pushed out, or judged a hire the manager wouldn’t make again — while only 19% become unequivocal successes. And it’s rarely for lack of skill: the study pins 89% of those failures on attitude and fit — coachability, motivation, temperament — not technical ability, which is exactly what a structured process is built to surface. The good news is that structure measurably helps: in the most recent comprehensive meta-analysis, structured interviews are the single best predictor of job performance among all selection methods (Sackett et al., 2022). “Structured” is the load-bearing word — the technique only pays off when you actually hold the bar high.

Earlier in my career, roughly 40% of the people I hired didn’t work out. Building a structured process — the four pillars below — is what moved that to about 10% for me. That’s my experience as an operator, not a guarantee. But the difference between a 40% and a 10% mis-hire rate compounds into a completely different team.

Someone who is 10% better doesn’t stay 10% ahead — they ramp faster, learn faster, raise the work of everyone around them, and the gap widens every quarter. Structure is how you tilt the odds toward that person on purpose instead of by luck.

The framework

The 4 Pillars of Talent Selection.

Every good hiring process serves four pillars — the framework I refined while scaling Zipwhip’s sales organization from under $1M to over $100M ARR, and the one Yardstick is built around. If an interview loop isn’t doing all four, it’s leaking predictive power. They’re role-agnostic: the questions and competencies change from job to job, but these four don’t.

Consistent

Every candidate for a role answers the same core questions in the same stages.

Consistency lets you compare candidates to each other, not to whoever interviewed last.

Behavioral

Questions are tuned to what the role requires, and ask what a candidate did — not what they would hypothetically do.

Real past behavior predicts better than a hypothetical, and gets harder to fake as your follow-ups sharpen.

Scored

Every interviewer rates every candidate on the same scorecard, on the same scale, against the same competencies and goals.

Scoring puts every candidate on one scale. It adds objectivity to the evidence; it does not make the decision for you.

Assessed

Somewhere in the loop, the candidate does a version of the job — a work sample, a practical exercise, a role play.

A work sample is the most honest signal you get: behavior you watched instead of behavior they described.

The architecture

The five-interview architecture.

A hiring loop should be a short, purposeful set of interviews, each with a clear job. Design each one by asking the same three questions: what is this stage for, how do we prepare the interviewer, and what materials does it need? You scale the loop to the stakes — a high-volume role leans on screening, competency, and a work sample; a senior or hard-to-reverse hire is where the full chronological interview earns its time.

Screening

The hardest interview to design well — decide who is worth the loop's time by confirming the required skills, the right experience, and genuine interest and fit.

Competency

Deep behavioral interviews on the handful of traits that actually predict success in this role. The competencies differ by job; the method of probing them with real examples doesn't.

Chronological

A structured walk through the candidate's career. Its superpower is that the follow-ups are boring — and therefore unspinnable. Worth the most for senior and high-stakes hires.

Work sample

The Assessed pillar in action: the candidate does a representative slice of the job while you watch. The interview that predicts the most, and the one most teams skip.

Reference check

Done right, not as a rubber stamp: wait until you are offer-ready, ask structured fact-based questions, and go beyond the candidate's hand-picked list.

Behavioral questions

Writing questions that hold up.

The most common way an interview loop leaks signal is bad questions. Three fixes matter most.

Ask what they did, not what they would do.

Hypothetical “what would you do if…” questions lose predictive power fast as the role gets more complex — for complex roles their validity drops to around r = .30, while past-behavior questions hold near r = .51 (Huffcutt et al., 2004). Lead with a real situation and ask how they actually handled it.

Ask a few deep questions, not many shallow ones.

In a 45–60 minute interview, three well-followed-up questions beat a checklist of ten. Depth is where the signal is — a great follow-up chain is what separates a rehearsed story from a real one. Our framework for follow-up questions is the engine that makes this work.

Don’t bake the answer into the question.

“Tell me about a time you showed leadership” tells the candidate exactly what to perform. Ask for the situation first, then find the signal in their follow-ups — and skip the tired “tell me about a time” opener so they can’t pattern-match to a canned answer.

Scoring

Score every candidate on one scale.

A scorecard exists to do one thing: get every interviewer projecting against the same outcomes, so you’re comparing candidates on one scale instead of comparing whoever left the strongest impression.

Keep it to 3–5 competencies and 3–5 role goals on a single card. Role goals are what the hire needs to accomplish — written by asking, “in 3, 6, 12, and 24 months, what would make me say this hire was a success or a mistake?” Score on a 0–4 behaviorally anchored scale, where 0 means “not enough information,” not “bad.” Resist the urge to weight competencies — precise weights feel rigorous but are usually false precision.

If you don’t yet have a scorecard for the role — and most teams don’t, because you don’t have performance data on day one — that’s the normal starting point, not a failure. Yardstick’s AI scorecard generator is built for exactly this cold start: it drafts a first-pass scorecard you can react to. The draft won’t be perfect — it’s meant to be wrong in places and easy to fix — but reacting to a real draft is far faster than staring at a blank page. Generate a first-draft scorecard and start editing instead of starting from zero.

The decision

Decide as a team, without a score cutoff.

Here’s the counterintuitive part: more interviews stop helping surprisingly fast. Google’s internal analysis, published on its re:Work blog in 2017, found four interviews were enough to predict a hiring decision with 86% confidence — after that, each additional interviewer improved accuracy by less than 1%. Four good, independent interviews beat six rushed ones.

The word doing the work there is independent. Interviews only add signal if opinions form on their own — so no comparing notes before the debrief. In the debrief itself, the leader speaks last, so the room isn’t anchored to the most senior opinion.

And the most important rule: never turn the decision into a score cutoff. “Anything over a 3.5 is a hire” feels objective, but it hands the decision to a formula. The scorecard’s job is to add objectivity to parts of the process — to make sure every candidate was evaluated on the same evidence. The decision itself is a human one, made by a team looking at that evidence together. Structure raises the quality of the inputs; it never replaces the judgment.

Make it your own

Make it work for your specific role.

The four pillars and the five interviews don’t change from job to job. What changes is the content: the competencies you score, the questions you ask, and what a good work sample looks like. A sales hire and a first engineer need completely different questions — but the same structured process underneath.

That’s the point of a framework: you build it once and re-skin it per role. This page is the foundation — the role-specific playbooks start here and swap in the right questions, competencies, and work sample. The first one is live: how to hire salespeople runs this exact process with sales competencies, a discovery-call work sample, and stages tuned for a revenue role. Guides for executive, first, and engineering hires build on the same foundation.

Put it into practice

Build the process without starting from a blank page.

The hardest part of adopting a structured process isn’t believing in it — it’s the blank page. Writing a competency question bank, a scorecard, and an interview plan for every role is real work, and it’s why most teams stay unstructured.

That’s the specific problem Yardstick is built to solve. The four pillars aren’t a philosophy you have to operationalize on your own — they’re the product. Yardstick is a structured-interview ATS where your team creates job-specific interview plans, runs consistent interviews, and collects scorecards on one scale, with AI assistance to get you off the blank page. If you work with a coding agent like Claude Code or Codex, it can operate Yardstick through the yardstick CLI — drafting interview plans, questions, and scorecards for you to review — while every sensitive action waits for your approval. Agents prepare the work; you decide.

The AI generators are designed around the cold start: a first-draft scorecard or question set that’s useful precisely because it’s easy to correct. You’re not asking the model to be right; you’re asking it to give you something to react to.

Turn one framework pass into a role-ready interview kit.

Generate interview questions and a matching scorecard, then edit them to fit your role — or see how Yardstick runs the whole process in one system.

Generate interview questions Book a call

FAQ

Common questions about building a hiring process.

How do you build a hiring process?

Build it on four pillars: the same questions for every candidate (Consistent), tuned to the role and focused on real past behavior (Behavioral), scored by every interviewer on one scorecard (Scored), and including a work sample where the candidate does a slice of the job (Assessed). Run those across a short set of interviews — screening, competency, chronological, work sample, reference check — and decide as a team on the evidence, without turning the score into a cutoff.

What is a structured hiring process?

A structured hiring process asks every candidate for a role the same core questions, in the same stages, and scores them on the same scorecard — so you're comparing candidates on one scale instead of on interviewer impressions. Structured interviews are the single best predictor of job performance among selection methods (Sackett et al., 2022); the structure is what makes the interviews comparable and the evidence usable.

How many interviews should a hiring process have?

Around four independent interviews is the sweet spot. Google's re:Work analysis found four interviews predicted the hiring decision with 86% confidence, and each additional interviewer after that added less than 1%. What matters more than the count is that opinions form independently before the debrief.

What are the 4 Pillars of Talent Selection?

Consistent (the same questions for every candidate), Behavioral (questions tuned to the job, asking what they did), Scored (one scorecard and one scale for every interviewer), and Assessed (a work sample, not just conversation). They are the role-agnostic foundation of a structured hiring process — the questions change by role, the pillars don't.

How do you evaluate candidates fairly?

Use one scorecard for every candidate: 3–5 competencies plus 3–5 role goals, rated 0–4, where 0 means “not enough information.” Have interviewers score independently before comparing notes, resist weighting the competencies, and use the scores to make sure everyone evaluated the same evidence — not to auto-decide the hire.

What's the most common hiring mistake?

Making the decision a score cutoff. A scorecard adds objectivity to the evidence, but the best hiring decisions are human decisions made by a team looking at that evidence together. Handing the call to “anything over a 3.5 is a hire” throws away the judgment the whole process was built to inform.