Back to Blog

    A Bad Hire Isn't a Cost. It's a Risk.

    In safety-critical work, hiring deserves the same discipline as engineering

    November 1, 20268 min read
    A warm plaster wall in angled sunlight with a single fine crack running through it and a small patch of red clay showing where the surface has flaked away.

    The companies building in defence, aerospace and other mission-critical sectors already know how to manage risk. They define requirements before they build. They verify against those requirements. They keep a record of why each decision was made.

    Hiring is rarely treated in the same way, but it should be.

    Two years to fill one seat

    Earlier in my career, I watched a business hire a new executive to lead its HR function. The initial search took several months. It was run largely by an agency, with several members of the executive team interviewing along the way.

    Within a couple of weeks of the new leader starting, the cracks showed. Their philosophy on how work should get done ran against the rest of the leadership team. By week five or six, it was clear their decision-making style simply wasn't compatible with how the business operated. They ended up departing only a few months into the role.

    The best part of two years passed between the start of the first search and the day their successor walked in. Months to find the wrong person. Months to realise it. Then a second full search.

    That was an HR role in a commercial business. Now picture the same failure in a systems engineering lead on a defence programme, or the person who signs off on a safety case. The delay is the smaller problem in these situations.

    Define the bar before you search

    No engineering team starts building without a requirements spec. Hiring deserves the same discipline: decide what the role must have, and separate it from what would be nice to have, before the first resume arrives.

    For safety-critical roles, four requirements tend to separate strong hires from the rest. Each one needs defining as behaviour you can observe, not a trait you hope for.

    Judgement under pressure. Good judgement isn't a single setting. It's knowing which kind of decision you're in. Some decisions are binary: safety, compliance, legal. You meet the bar or you don't, and risk tolerance should be close to zero. Most day-to-day decisions sit on a spectrum, with no legal line attached. There you want someone who takes the emotion out, invites views, and then disagrees and commits rather than waiting for a consensus that will never come. A candidate who treats every call as bendable is reckless. One who treats every call as rigid can't function. You're hiring for the ability to tell the two apart.

    Rigour. My test is simple: could this person defend their decision later, to a room, with confidence? Rigour isn't certainty or having hard data every time. A clear "it was a judgement call, and here's exactly why I leaned that way" passes. A muddled explanation doesn't.

    Honesty and accountability. Responsibility covers what you do yourself. Accountability covers everything in your span of influence, including indirect effects and the actions of people you lead. Most people are comfortable being responsible, because it's hard not to be. Far fewer can sit in accountability, because it means owning outcomes that weren't directly their doing. In a leadership hire, that gap matters enormously.

    Standards compliance. On the binary decisions, there is no creative interpretation. You want people who see standards as the floor, not an obstacle. In an interview, watch what happens when a case study tempts them to cut a corner. Do they hold the line where it is essential to do so?

    Everything else, from specific tools to sector experience, belongs on the nice-to-have list. Those things matter, but they can be taught or learned. The four above usually can't be fixed after the fact.

    Turn every requirement into evidence

    A requirement you can't verify isn't a requirement. Each must-have needs a method that produces evidence, not an impression.

    Use real problems, then keep asking why. Case studies built on problems the person would actually face are the best rigour test I know. The case itself is only half of it. The other half is an interviewer who keeps asking why, and applies a little pressure. Strong candidates can trace every decision back to evidence, a requirement, or a reasoned trade-off. Weaker ones get muddled.

    Ask for the redo, then probe it. Every candidate expects a question about failure and has a polished answer ready. I ask: what piece of work would you like a second chance at, and why? The real signal is in the follow-up. If you got the redo, what would you do differently, and how does that differ from the first time? Someone who has genuinely learned gives a specific, causal answer. Someone reciting an anecdote stays vague.

    Read accountability across the whole conversation. There's no single question that reveals it, because candidates know what you want to hear. Instead, listen for where accountability lands across every answer. Does the person own what went wrong, or does blame drift outward to "we" and "they"? Context matters. Sometimes "we" is honest, and sometimes the cause really was outside their control. You're looking for the pattern, not a pronoun.

    Design for failure modes. Safety engineers ask how a system could fail before it does. Ask the same of your process: which of our must-haves could a candidate fake, and which stage would catch it? If the answer is "none", you have a gap.

    Build in redundancy. Critical systems never rely on one sensor. Don't rely on one interviewer's read either. Each must-have should be tested by at least two independent methods, such as a case study and a structured question, or an interview and a reference.

    Calibrate: evidence beats vibes

    Most interview processes do assess something. The problem is that nobody can say what.

    When I ask hiring teams what they learned in a first interview, the honest answer is often "good vibes". That usually means real things: personality, communication style, general experience. But if you can't articulate what you assessed, you can't look back at it, and you can't decide where to drill in next. You've lost traceability, the same failure behind the two-year hire.

    An undesigned process breaks down in three predictable ways:

    • Wrong signals. The team assesses what's easy to see, not what the role requires.
    • Redundancy without independence. Three interviewers each run what is effectively a first interview, testing the same things. That's three opinions, not three verifications.
    • Unexplainable decisions. The debrief becomes a vote on feelings, and nobody can say why the hire was made.

    The fix is unglamorous. Assign each must-have to a specific stage and interviewer before the process opens. Brief interviewers on what good and poor evidence look like for their area. Have everyone score independently on a shared scorecard before the debrief, so the loudest voice doesn't anchor the room. Then run the debrief on evidence: what did you see, and what does it tell us against the requirement?

    A useful test: if you had to defend this hire to your board in twelve months, could you show what you looked at and why you decided? Win or lose, that's the standard.

    Integrity and disclosure: where interviews hit their limit

    Learning from mistakes and disclosing them are different behaviours. The redo question tests the first. In safety-critical work, the second matters more: the person who reports their own near miss, flags the error they could have buried, or says "I don't know" instead of bluffing.

    There's no interview question that reliably catches this. Ask directly and candidates will tell you what you want to hear. The best an interview can do is the accountability read described above, gathered across every answer.

    This is also where hiring overlaps with clearability. Security screening relies on candour about the past. A candidate who is open about uncomfortable history during your process is showing you the behaviour that screening depends on. One who is evasive about small things is giving you an early warning.

    Because interviews have limits here, references have to carry more weight. Treat them as a verification stage, not a formality:

    • Speak to people who managed the candidate, not just peers they chose.
    • Ask about specific situations: a mistake, a disagreement, a time they raised bad news.
    • Test what you heard in interview. If the candidate described owning a failure, ask the referee how it was handled.
    • Listen for hesitation as much as content.

    Rigour without slowing down

    Everything above sounds like it should make hiring slower. It doesn't have to. The best candidates have a job, commitments outside work, and usually other options. If your process asks for five more hours than a competitor's, they'll often choose the easier route, even if yours is better. That's not a character flaw. It's being human.

    Two principles keep the bar high and the process short.

    Make practical assessments part of your interview, not a separate stage. Don't set a generic take-home that disappears for review. Either give the case study in the room, or ask the candidate to prepare something they'll present and discuss. People will happily put in effort that is clearly preparation for a specific conversation. They resent unpaid homework that goes nowhere.

    Cap it at three or four stages, even for leadership roles. When a process runs to five or six interviews, it's rarely because the role demanded it. It's because nobody decided up front what each stage would test. Three interviews in, the team realises it hasn't assessed most of what matters, and starts chasing its tail. To the candidate, that reads as doubt about them, and the streamlined competitor wins.

    This is the point CEOs most often miss: speed and rigour aren't in tension. The discipline that makes you rigorous, deciding in advance what each stage is for, is exactly what makes you fast.

    Build your own scorecard

    No hire is risk-free, and no process gets a 100% hit rate. But you can cut the odds of failure dramatically by doing the right things up front, and you can always show what you did.

    The traits above tend to separate strong people in safety-critical roles. Even so, the right scorecard for your company depends on your mission, your values, and what matters to you. It will also differ by role. A head of sales and a detail-driven systems engineer are different profiles, and should be assessed differently.

    Your hiring process is either designed, or it runs on vibes you can't defend later. If you'd like help designing the version that fits your company, talk to Resonance.

    Is your hiring process designed, or inherited?

    We'll look at how you hire today and rebuild the parts that decide whether the next person works out.