Insights and Recommendations — HR Analytics¶
Executive Summary¶
Nexus Technologies' 16.1% attrition costs an estimated £8.3M per year — and it is not random. Three actionable drivers dominate: excessive overtime (3× attrition), early-tenure attrition (28% of new hires leave within two years), and the systematic underpayment of Sales Representatives (40% attrition, lowest pay). Targeting these three levers could plausibly cut attrition to 11% and save £2.5-3M annually.
Findings¶
Finding 1 — Overtime is the Single Biggest Driver¶
Employees who work overtime leave at 30.5% vs 10.4% for those who don't — a 3× difference, and the strongest correlation in the dataset.
This is the most important finding because it's actionable. Unlike age or marital status, overtime is something leadership directly controls through staffing, workload management, and culture.
Finding 2 — The First Two Years are Critical¶
Employees with 0-2 years tenure leave at 27.7%, vs 8.6% for those with 10+ years. More than a quarter of new hires don't make it past two years.
This points to two possible problems: poor hiring fit (recruiting the wrong people) or poor onboarding (good people who don't get supported early). Either way, the early-tenure window is where retention effort has the highest leverage.
Finding 3 — Sales Representatives are Underpaid and Overworked¶
Sales Reps have the highest attrition of any role (39.8%) and the lowest average pay (£2,626/month). The combination of high pressure, low pay, and likely high overtime creates a perfect storm.
Finding 4 — Low Pay and Promotion Stagnation Compound Risk¶
The lowest income quartile churns far more than the highest. Employees overdue for promotion (5+ years since last) also show elevated attrition. People leave when they feel financially undervalued and professionally stuck.
Finding 5 — Work-Life Balance Scores Predict Departure¶
Employees rating work-life balance as "Bad" (1) churn dramatically more than those rating it "Better" or "Best." This reinforces the overtime finding — it's not just hours, it's the felt experience of imbalance.
Recommendations¶
Recommendation 1 — Tackle Overtime (Highest Impact)¶
Action: - Audit which teams systematically work overtime and why (understaffing? unrealistic deadlines? poor planning?) - Set overtime thresholds that trigger a staffing review - Hold managers accountable for their team's overtime levels
Expected impact: If overtime workers' attrition dropped from 30.5% toward the 10.4% baseline, that alone could save ~80 employees/year — roughly £3M in replacement costs. Even a partial improvement is the single highest-ROI intervention available.
Recommendation 2 — Build a 90-Day and 1-Year Onboarding Programme¶
Action: - Structured onboarding with clear milestones - Manager check-ins at 30, 60, 90 days and 6 months - Early mentorship pairing - Exit interviews specifically for sub-2-year leavers to understand why
Expected impact: Reducing 0-2 year attrition from 27.7% to 20% would retain ~30 additional employees/year.
Recommendation 3 — Review Sales Representative Compensation¶
Action: - Benchmark Sales Rep pay against the market — are we below it? - Review the comp structure (base vs commission balance) - Investigate the role's workload and overtime
Expected impact: Sales Reps are 83 people churning at 40%. Halving that rate retains ~16 reps/year and stabilises the sales pipeline.
Recommendation 4 — Address Promotion Stagnation¶
Action: - Flag employees 4+ years without promotion for a career-path conversation - Create clearer advancement criteria and lateral-move options - Ensure high performers without promotion opportunities get retention incentives (stock, raises)
Expected impact: Retains experienced, high-value employees who are most expensive to replace.
What NOT to Do¶
Don't act on non-actionable correlations
Age, marital status, and distance from home correlate with attrition — but you cannot (and must not) make HR decisions based on them. Using "single employees churn more" to influence hiring would be discriminatory and illegal. Focus exclusively on factors the company can ethically change: workload, pay, development, and management quality.
Don't build individual "flight risk" lists for managers
A model that flags individuals as likely to quit can become self-fulfilling — managers disengage from flagged employees, accelerating their departure. Keep the analysis at the systemic level.
What I'd Analyse Next¶
- Manager-level analysis: does attrition cluster under specific managers? (The saying "people leave managers, not companies.")
- Exit interview text analysis: what reasons do leavers actually give?
- Compensation gap analysis: are there pay inequities within the same role/level?
- A predictive model — but used to identify systemic risk factors, not to surveil individuals.