Latest / Hiring & Firing with Fexingo: Recruitment, Performance, and Workforce Decisions / How a Two-Day Paid Trial Replaces Four Rounds of Interviews
Transcript
- Lucas: So Luna, I was looking at a case study last week about a company called Nextera Analytics — about two hundred employees, B2B data analytics. And they were completely fed up with their hiring process. Luna: Fed up how? Like high turnover or just slow? Lucas: Both. Their early attrition — people leaving within the first six months — was running at about twenty-five percent. And their time to hire was averaging forty-five days across four interview rounds. They knew something was broken. Luna: Four rounds is a lot for a company that size. But what did they try? Lucas: They ran an experiment. They replaced the final two interview rounds for three roles — a data engineer, a product manager, and a sales account executive — with a two-day paid project. For each role, they brought in the top three candidates and asked them to do a real, but scoped-down, project over two days. Luna: Paid? How much? Lucas: They paid a flat one thousand dollars for the two days. Not a huge investment compared to what they were spending on wasted recruiter time and bad hires. And they designed the projects to mirror actual first-week tasks. Luna: So what happened? Did it work? Lucas: Within twelve months, early attrition dropped from twenty-five percent to eight percent. Cost per hire fell by roughly forty percent because they cut out those two interview rounds and the associated coordination overhead. And manager satisfaction with new hires — measured on a one-to-five scale — went from two point eight to four point two. Luna: That's a huge jump. But I'm wondering — how do you design a project that doesn't favour people who just have more free time or better equipment? Lucas: Great question. Nextera was careful. They capped the time commitment to about twelve hours total, spread over two days. They provided all the data sets and tools in a sandbox environment, so candidates didn't need their own expensive software. And they had clear rubrics — not just 'did it look good' but specific criteria around problem-solving, communication, and technical accuracy. Luna: Was there any pushback from candidates who didn't want to invest two days? That's a lot of time. Lucas: They had about a fifteen percent decline rate. Some people said it was too much. But Nextera framed it as a paid audition — you're not working for free, you're showing us what you can do. And the ones who declined were often candidates who were already juggling multiple offers or weren't seriously interested. So in a way, it self-selected for genuine engagement. Luna: That's a good point. So the trial is also a filtering mechanism. But what about bias — if the project is designed by a team that's not diverse, does it end up favouring people with a similar background? Lucas: Absolutely a risk. Nextera had each project reviewed by a panel of three people from different teams — including someone from outside the department. And they anonymized the deliverables as much as possible, no names or backgrounds. They also tested the project on a few internal employees first to make sure it wasn't accidentally relying on specific industry jargon or tools that only one school teaches. Luna: That sounds like a lot of upfront work. How long did it take to set up? Lucas: About three weeks per role the first time. But after that, they could reuse and tweak the project for similar roles. And the payoff came fast. Within six months, they had the data to show that the project was a stronger predictor of on-the-job performance than any of the interview rounds they replaced. Luna: I mean, it makes sense intuitively. Seeing someone actually do the work has to be more informative than asking 'tell me about a time you resolved a conflict'. Lucas: Exactly. And there's research backing this. A study from the Journal of Applied Psychology found that work sample tests — they call them 'job performance simulations' — have a validity coefficient of about point five four, which is significantly higher than unstructured interviews, which are around point two zero. So Nextera's experience isn't an outlier. Luna: So why don't more companies do this? Is it just inertia or is there a real barrier? Lucas: I think inertia is a big part. The standard interview process is comfortable — everyone knows how to do it. But there are also legitimate concerns: legal risk if the project discriminates, the time cost of designing and evaluating the project, and the candidate experience of asking for a big time commitment. Nextera addressed most of those with the paid, structured, panel-review approach. Luna: If this is working so well, what's the downside you've heard from companies that tried it and stopped? Lucas: The most common complaint is that it slows down the pipeline when you have many candidates. If you're hiring ten people a month, running a two-day project for every finalist means you can only process a few at a time. But Nextera argued that it's better to hire the right person slowly than the wrong person quickly. Also, they used the project only for the final three candidates, so earlier stages were still fast. Luna: That makes sense. And the numbers speak for themselves. I'm sure listeners are thinking about whether they could try this in their own teams. Lucas: If today was actually useful to you, the way these episodes stay ad-free is listener support — you can buy me a coffee dot com slash fexingo. Every little bit helps us keep digging into these concrete case studies. Luna: Yeah, seriously. We put a lot of work into finding specific examples like Nextera, and knowing it helps someone's hiring process is what keeps us going. Lucas: So to bring it back to the design — one of the smartest things Nextera did was to measure not just the output but how candidates handled feedback. During the two-day project, they gave each candidate a midpoint check-in with one piece of constructive feedback. Then they evaluated how the candidate incorporated that feedback. That turned out to be a stronger predictor than the final deliverable itself. Luna: Ah, so they're assessing learning agility, not just technical chops. That's clever. Lucas: Exactly. They called it the 'feedback iteration score.' Candidates who didn't just take the feedback but actively asked clarifying questions and adjusted their approach — those people became the top performers. And that's something you can't get from a standard hour-long interview. Luna: I bet that also helps with culture — if you need people who are coachable, this screens for that naturally. Lucas: Yes. And it reduces the risk of hiring someone who interviews well but can't adapt. Nextera found that candidates who scored high on the feedback iteration score had a ninety percent retention rate after eighteen months, compared to sixty percent for those who scored low. Luna: That's a massive difference. So essentially, they're using the trial to simulate the first week on the job — including the learning curve. Lucas: Right. The project itself was designed to have a few intentional ambiguities — incomplete requirements, conflicting data points — to see how candidates handle uncertainty. That's the stuff that actually kills people in the first ninety days, not whether they can recite the STAR method perfectly. Luna: I love that. It's like the interview becomes the job preview as much as the assessment. And the candidate gets a real taste of what they'd be doing. Lucas: Exactly. Nextera even had a few candidates who self-selected out after the trial — they realized the work wasn't what they expected. That's a win for both sides. Better to find out over two paid days than after two months of onboarding. Luna: So what's the one takeaway for someone who wants to try this on a smaller scale, say for a single role? Lucas: Start with one role that's high-volume or high-impact. Spend a week designing a four-hour project — not two days initially. Pay the candidates a fair stipend. Use a simple rubric. Evaluate it on a pilot basis with three to five candidates. And track the same metrics: early attrition, time to hire, and manager satisfaction. You'll have your own case study within a quarter. Luna: And if it works, you scale it. If it doesn't, you've only lost a little time and money. That sounds like a smart experiment for any hiring team. Lucas: Yeah. The companies that are leading in this space — like Buffer and Automattic — have been doing paid trials for years. Nextera just proved it works for smaller, less well-known companies too. Luna: It's a great reminder that sometimes the best hiring innovation is just letting people show you what they can do. Lucas: Exactly. And if you try it and see a similar drop in attrition, we'd love to hear about it. That's the kind of story that makes this show worth doing.