How Berkeley Group transformed apprenticeship hiring with blind recruitment and CV-free applications.

Going into the 2024 peak season, Northwind’s talent team knew the maths. To staff up to peak, they needed to make roughly 400 warehouse operative hires across two sites. To do that with a 38% 90-day attrition rate, they actually needed to hire closer to 550. To process 550 hires through their existing funnel, they needed to phone-screen around 1,800 applicants — and even then, hiring managers reported that fewer than half of new starters were “obviously a good fit” by week two.

The phone screen had been the standard for years. A 15-minute call, a set of behavioural questions, a yes/no on whether to invite to interview. The script was tight, the screeners were experienced, and the result was a steady churn that nobody quite liked but everybody had stopped questioning.

The TA director had run the numbers in early summer. “If we run last year’s process again, we’ll spend roughly £180,000 on first-year attrition costs across these two sites alone. And we’ll still have the same conversation in January about why so many people quit in their first month.”

ThriveMap had been running with Northwind’s permanent operative hiring for nine months by the time peak planning began. The assessment — a 12-minute structured walk-through of typical warehouse scenarios — had already shown a meaningful effect on retention in permanent roles. The question for peak was whether the same approach could carry the volume.

The team made three deliberate choices about the seasonal version. First, they kept the assessment short — 10 minutes median completion — to protect application rates during a competitive labour market. Second, they made the assessment the first stage of the funnel, replacing the phone screen entirely, so candidates encountered the realistic preview before investing time in a phone call. Third, they let the assessment do the screening — the hiring team interviewed only candidates above a fit-score threshold, and the threshold was calibrated to deliver roughly the same number of interviews per week as the previous process.

Across the seasonal hiring window, the team measured three things they cared about: 90-day attrition, interviews-to-hire ratio, and hiring manager satisfaction. By week six of the campaign, all three had moved meaningfully.

The interviews-to-hire ratio moved from 3.1 to 1.8 — the team was still making the same number of hires, but they were getting there through fewer interviews because the candidates they did interview were more likely to be the right ones. Hiring manager satisfaction with new hires (measured at 30 days) moved from 3.4 to 8.7 on a 10-point scale.

The effect the team hadn’t expected was how many candidates would screen themselves out. Across the campaign, 22% of candidates who started the assessment chose not to complete it — most of them at the point where the realistic preview made clear what the role actually involves. The candidates who completed and proceeded had effectively self-declared that the work was for them.

This changed the conversation about candidate experience too. The candidates who’d previously quit in week three had effectively been pulled into a job they wouldn’t have chosen if they’d known what it was. Self-selection wasn’t just a numerical effect on the funnel — it was a quieter ethical improvement on a process that had been spending people.