Pulse Survey × Turnover Prediction — Detecting Resignation Risk Early with 6-Axis Scoring
The Late Detection Problem
"I had no idea they were thinking about leaving."
This is what most managers say when they hear about a resignation. Exit interviews gather information after the decision is made. The two weeks' notice is tendered, the manager reflects on signals they missed, and the hiring cycle begins again.
Pulse surveys offer a fundamentally different approach: ongoing signal monitoring that detects resignation intent 1–3 months before an employee commits to leaving. The key is not just score levels, but the patterns of how scores move across multiple axes over time.
Why Pulse Surveys — Not Stress Checks
Japan's stress check system (Industrial Safety and Health Act, Article 66-10) explicitly prohibits employers from accessing individual stress check results without the employee's explicit request. Using stress check data for turnover prediction or risk analysis is not only ethically problematic — it is legally impermissible.
Pulse surveys operate under an entirely different framework:
| Factor | Stress Check | Pulse Survey |
|---|---|---|
| Legal basis | Industrial Safety and Health Act Article 66-10 (employer access prohibited) | Employer-administered voluntary survey (employer use permitted) |
| Purpose | Measure psychological load; primary prevention | Organizational and management improvement feedback |
| Result access | Individual only (group analysis permitted) | Aggregation, trend analysis, management response — all permitted |
| Frequency | Annual | Monthly, biweekly, or custom cadence |
This is not a workaround — it is the intended design. Pulse surveys measure the organizational and relational factors that drive voluntary turnover. Stress checks measure psychological load. They serve different purposes and operate under different rules.
The COCKPITOS 6-Axis Framework
COCKPITOS's pulse survey measures six dimensions that together map the full employee experience:
| Axis | Sample Question | Predictive Value for Turnover |
|---|---|---|
| Retention intent | "I expect to still be working here a year from now" | ⭐⭐⭐ Direct — highest predictive value |
| Workload | "My workload is manageable given my abilities" | ⭐⭐ Leading indicator for burnout-driven departures |
| Psychological safety | "I can share my opinions without fear of criticism" | ⭐⭐ Organizational trust signal |
| Manager support | "My manager supports my growth and development" | ⭐⭐⭐ Manager-driven departures move fastest |
| Colleague support | "My team members support each other well" | ⭐ Workplace isolation indicator |
| Growth opportunities | "I see opportunities to grow my career here" | ⭐⭐ Medium-term departure signal |
Score levels matter, but score movement and co-movement across axes is what drives prediction accuracy.
Three Resignation Signal Patterns
Pattern A: Acute Signal — Resignation Risk Within 3 Months
Indicators: - Retention intent drops 20+ points month-over-month - Retention intent and manager support both decline in the same month
What's happening: A triggering event has likely occurred — a negative performance review, being passed over for a promotion, a specific conflict with the manager. The employee's decision window is short, and the signal is already late-stage.
Response: - Schedule a 1on1 within the week - Do not open with the score: start with "Has anything changed for you recently?" - The goal is to understand what happened — curiosity, not confrontation - If the manager is part of the problem, involve HR directly
Pattern B: Accumulative Signal — Resignation Risk Within 6 Months
Indicators: - Workload scores below the 60-point baseline for 3 consecutive months - Psychological safety stays below 50 points for 3 months
What's happening: Chronic overload or persistent interpersonal tension is eroding the employee's ability to stay engaged. This is slower-moving than Pattern A, but equally serious — and more structurally addressable.
Response: - At the next 1on1, frame workload and team dynamics as the explicit topic - Look for structural causes: task allocation, meeting density, unclear role scope - Propose one concrete structural change before the next survey cycle
Pattern C: Career Dissatisfaction Signal — Resignation Risk Within 12 Months
Indicators: - Growth opportunity scores below baseline, sustained across multiple months - Retention intent is moderate (60–70) but growth is chronically low
What's happening: The employee has mentally decoupled from the organization's direction. They're not actively job searching, but they're no longer imagining their future at the company. A compelling external offer will convert them quickly.
Response: - Schedule a dedicated career conversation — separate from the regular 1on1 - Do not ask "where do you see yourself in five years?" — ask "what do you wish you were working on that you're not?" - Surface specific growth opportunities the organization can actually offer, not generic encouragement
Translating Signals into 1on1 Agendas
When a Pattern A or B signal appears, the next 1on1 should be structured around it:
| Step | Time | Content |
|---|---|---|
| Opening | 5 min | Open-ended check-in — no scores, no evaluation |
| Signal reference | 5 min | "Your workload rating dropped this month — what's been happening?" Showing the score is appropriate; interpreting it for them is not. |
| Root cause inquiry | 10 min | One level deeper: what specifically is driving the score? Volume, complexity, unclear expectations, or something relational? |
| Single commitment | 5 min | One concrete change before the next 1on1 |
What managers should avoid: - ❌ Challenging the score ("Why is this so low?") - ❌ Dismissing it ("Everyone feels overwhelmed during Q1") - ❌ Using scores as performance evidence — pulse surveys are management feedback, not evaluation material - ❌ Waiting for the next month's data before responding to a Pattern A signal
Organization-Level Risk Monitoring
Individual signals matter, but department-level trends reveal structural problems that individual 1on1s alone cannot address:
| Monitoring Metric | Review Frequency | Action Threshold |
|---|---|---|
| Department average retention intent | Monthly | Below 60 → flag for HR review |
| Individuals with multi-axis decline | Monthly | +2 vs. prior month → team-level response |
| Company-wide growth opportunity average | Quarterly | Below 55 → review career development programs |
Risk heat map: Visualize quarterly where Pattern A and B signals are concentrated by department. This tells HR where manager coaching investment is most urgent, and where team-level structural changes are needed.
What Pulse Surveys Cannot Detect
For calibrated expectations:
- Silent resigners: Employees who answer positively but have already committed to leaving will not appear in the data
- Sudden external offers: High-compensation poaching moves faster than survey cycles
- New employees (< 3 months): Insufficient baseline for trend analysis
Pulse surveys work best for the largest category of voluntary turnover: employees who are disengaged, dissatisfied, or burned out — but have not yet decided. The signal is there. What managers do with it is what determines whether it becomes a resignation.
Summary
| Signal Pattern | Detection Window | Primary Response |
|---|---|---|
| Pattern A (retention intent drops sharply) | Within 3 months | Immediate 1on1; curiosity-first |
| Pattern B (workload/safety sustained low) | Within 6 months | Structural review + focused 1on1 |
| Pattern C (growth dissatisfaction sustained) | Within 12 months | Career conversation + development planning |
| Department-level heat map | Ongoing | HR-led manager support |
Pulse surveys are not surveillance tools — they're a system for managers to ask better questions at the right time. The 6-axis framework exists because no single question captures why someone stays or leaves. It is the combination of scores, the direction of trends, and the management response that determines whether a signal becomes a resignation.
COCKPITOS provides pulse survey deployment, 6-axis trend analysis, and high-risk employee alert notifications integrated with 1on1 records — so the signal reaches the right person at the right time. Contact us via our free consultation form.