How Berkeley achieved 100% graduate retention by optimising their pre-hire assessment

Time to hire.
Application conversion.
Assessment completion rates.
Offer acceptance.

These metrics are easy to measure and easy to improve. They also tell us very little about whether hiring actually worked.

The real test of hiring success happens months later, when new employees decide whether to stay.

Across early careers hiring, this remains one of the biggest unresolved challenges. ThriveMap’s State of Assessment Market data shows that 66% of entry level employees have left a role because the reality of the job did not match what they expected. Despite this, only 29% of talent acquisition teams consistently track early attrition outcomes.

Retention, not recruitment efficiency, is the true measure of hiring quality.

Berkeley Group’s early talent hiring transformation demonstrates what happens when retention becomes the primary optimisation goal rather than a downstream HR metric.

Three years ago, Berkeley partnered with ThriveMap to redesign graduate and apprenticeship hiring around a Realistic Job Assessment introduced early in the recruitment journey. The objective was not simply fairer selection or faster hiring. It was expectation alignment.

Construction presents a unique challenge. Many applicants have little direct exposure to site environments before applying. Recruitment data showed that 44% of successful hires had never previously been on a construction site, meaning candidates were often making career decisions based on incomplete understanding of the role.

Traditional hiring attempts to solve this after employment begins through onboarding or engagement initiatives. Berkeley instead moved realism forward into the hiring process itself.

The ThriveMap assessment simulates day to day responsibilities, operational pace, collaboration demands and working environments before interview. Candidates experience the realities of the role and can opt out early if the job does not align with expectations.

This approach shifts assessment from screening candidates to helping candidates make informed decisions.

Candidate feedback reflected this immediately. Applicants described the process as a genuine level playing field focused on capability rather than background, while others highlighted that experiencing realistic scenarios helped them understand what working in construction would actually involve.

Initial outcomes were strong. Female representation increased significantly and hiring teams reported greater confidence in selection decisions. More importantly, the September 2024 graduate cohort achieved 83% retention, already outperforming industry averages where early career attrition commonly exceeds 40 to 60 percent.

Many organisations would consider this a successful endpoint.

At Berkeley, it became input data.

ThriveMap’s role as an assessment partner extends beyond implementation. The assessment was continuously refined using candidate feedback, hiring manager insight and six month probation outcomes. Where new joiners reported unexpected aspects of operational pace or collaboration, scenarios were strengthened. Where performance indicators suggested stronger predictors of success, competency weighting was recalibrated.

The hiring system learned from real workforce outcomes.

By the time the September 2025 graduate intake entered Berkeley, candidates encountered a more accurate and transparent representation of the role shaped by previous cohorts’ lived experience.

The result was measurable.

The 2025 graduate cohort currently maintains 100% retention.

The improvement from 83% to 100% did not come from hiring fewer candidates or tightening selection thresholds. It came from improving expectation alignment before employment began.

Candidate Insight	Industry Benchmark	Berkeley Outcome
Roles accurately represented	28% typical perception	93% clarity reported
Early retention	40–60% attrition common	100% retained (2025 cohort)

Most recruitment technology focuses on efficiency metrics such as reducing time to hire or increasing funnel conversion. These improvements optimise process speed but rarely address the underlying cause of early attrition: mismatch between candidate expectations and job reality.

ThriveMap’s approach treats assessment as a continuously optimised alignment system. Realistic job assessments allow candidates to understand working conditions, responsibilities and challenges before committing to the role.

When candidates choose to continue, they do so with informed intent.

Retention becomes a hiring outcome rather than an HR recovery exercise.

This aligns with broader ThriveMap research showing that job realism remains one of the most underused levers in hiring. Fewer than five percent of organisations currently use assessments to educate candidates about the realities of the job itself, despite mismatch being the leading driver of graduate attrition.

Berkeley’s experience demonstrates what changes when hiring is optimised against retention rather than speed.

Assessment evolves intake after intake.
Candidate understanding improves.
Confidence in accepting roles increases.
Early attrition declines.

Over time, the assessment moves from a selection tool to predictive hiring infrastructure.

The implication for talent acquisition is significant. Hiring success should not be measured by how quickly roles are filled, but by how long new employees stay and succeed once they join.

When realistic job assessments are continuously refined using retention and performance data, hiring stops being a transactional process and becomes a system capable of improving workforce stability year after year.

In labour markets facing persistent skills shortages, retention is no longer an outcome of hiring.

It is the purpose of it.