Latest / Hiring & Firing with Fexingo: Recruitment, Performance, and Workforce Decisions / How a Failed Hire Cost One Startup 2 Million Dollars
Transcript
- Lucas: There's a SaaS startup in Austin — about fifty people, Series A, decent traction — that made one bad senior hire last year and it cost them roughly two million dollars. Luna: Two million dollars for one person? That's almost their entire annual burn for a team that size. Lucas: Right. And they didn't pay that person two million in salary. The salary was like two-twenty plus equity. But the total cost came from everything else — the lost productivity, the missed deadlines, the two other employees who quit because of the friction. Luna: Okay, so walk me through the specifics. What was the hire? Lucas: They hired a VP of Engineering. The founder ceo was formerly a product person, not technical, so he relied heavily on the interview loop. The candidate had a great resume — FAANG pedigree, solid references, nailed the behavioral questions. Luna: Classic pattern. Great on paper, falls apart in practice. Lucas: Exactly. Within the first month, the VP started making architectural decisions that the senior engineers pushed back on. He dismissed their concerns, went ahead anyway. By month three, the team's velocity dropped by about forty percent because they were re-doing work he'd green-lit. Luna: And that's where the two million comes from? Lucas: Partially. The CEO did a back of the envelope calculation: the VP's salary and severance — about eighty thousand. But the bigger numbers were the engineering hours wasted on the scrapped features — roughly six hundred thousand in payroll cost for work that never shipped. Then the missed product launch delayed a partnership deal worth about eight hundred thousand in committed revenue. Luna: So the original revenue number just disappeared because they couldn't deliver on time. Lucas: That deal fell through completely. And then two senior engineers — people who'd been there since the seed round — gave notice because they didn't want to keep fighting bad technical decisions. Their replacement cost and the knowledge loss, call it another four hundred thousand. Luna: That's brutal. But I have to ask — shouldn't the CEO have caught this during the interview process? There must have been red flags. Lucas: That's the thing — traditional interviewing is terrible at predicting performance. Behavioral questions are rehearsable. References are curated. Even a technical screen can miss how someone actually works on a team. Luna: So what did they do differently afterward? I assume they changed their process. Lucas: They made three changes. First, they introduced a 'pre-mortem' on the hiring brief. Before even sourcing candidates, the CEO and the existing engineering leads wrote down: 'If this hire fails, what are the most likely reasons?' They came up with things like 'doesn't listen to the team' and 'over-engineers solutions.' Luna: So they defined failure modes upfront instead of hoping the interview would reveal them. Lucas: Exactly. Second, they replaced the standard coding challenge with a work-sample test. They gave the final candidate a real, messy problem from their own backlog — not a LeetCode problem — and asked them to spend two hours sketching an approach and explaining their trade-offs. Luna: That's harder to game. You're evaluating how they think about actual constraints. Lucas: Right. And third — this is the one I think is most important — they put a thirty-day review with explicit kill criteria. At day thirty, the CEO and the team would sit down and decide: is this person making the team faster or slower? If slower, they'd part ways immediately, with a pre-agreed severance. Luna: That's uncomfortable but honest. It treats the first month as an extended tryout rather than a done deal. Lucas: Exactly. And it forces the manager to actually evaluate performance in real time instead of hoping things improve. The CEO told me that after they implemented that, they let go of one other new hire at day twenty-eight. Cost them maybe forty thousand in severance instead of two million. Luna: That's a massive ROI on process. So the two-million-dollar hire was essentially a tuition fee for a better hiring system. Lucas: That's one way to frame it. But it's a tuition fee most startups can't afford. The broader point is that hiring is a financial risk, not just an HR problem. And yet most companies spend more time deciding on a new vendor than they do on a director-level hire. Luna: Speaking of financial risk — we've been talking about the cost of a bad hire, and I think it's worth mentioning that this show is intentionally ad-free. We don't run commercials or sponsored segments because we want to keep these conversations focused on what actually matters. If you find value in that, and you want to support the work, there's a simple way: buy me a coffee dot com slash fexingo. No pressure, no perks, just a way to keep the show independent. Lucas: Yeah, Luna, that's exactly right. We hear from listeners that the ad-free format makes a difference — you can share an episode with a colleague without it feeling like a commercial break. So if that resonates, the link is buy me a coffee dot com slash fexingo. And now back to the numbers. Lucas: So after that experience, the CEO started tracking a metric he calls 'hire impact lag' — the time between when a new person starts and when their net contribution becomes positive. For the bad VP, that number was never reached. For a good hire they made later, it was about sixty days. Luna: That's a useful concept. Most companies just track time to fill or cost per hire, not whether the hire actually improves things. Lucas: And those metrics can be misleading. If you hire fast and cheap but the person drags down the team, you've saved a little money upfront and lost a lot later. The U.S. Department of Labor has a rough rule of thumb that replacing a single employee costs one-third of their annual salary. For a director-level role, that climbs to two hundred percent when you factor in lost productivity and training. Luna: So for a VP making two hundred thousand, you're looking at four hundred thousand in replacement costs — and that's before you account for the kind of cascading damage we saw in the Austin startup. Lucas: Exactly. That's why I think every company, especially startups, should run a post-mortem on every significant hire — good or bad. What did we learn about our process? What signals did we miss? What signals worked? Luna: And the goal isn't perfection — it's to reduce the variance. You'll never eliminate bad hires entirely, but you can make them less catastrophic. Lucas: Right. The Austin startup now has a hiring playbook that includes the pre-mortem, the work-sample test, and the thirty-day review. They've made eight hires since the VP disaster — two didn't work out, but they caught both within the first month. Total cost of those two failures: about seventy thousand. A far cry from two million. Luna: So the lesson is: invest in process upfront, or pay for it later. I think that's a good note to end on. Lucas: Yeah. And next time, we'll talk about how to actually conduct a thirty-day review without making it feel like a firing squad. That's for next week.