Rate your segment
Move every slider to reflect this team1 = easily replaceable · 3 = notable knowledge loss · 5 = revenue owner / single point of failure
1 = under half stay · 3 = roughly two-thirds stay · 5 = over 90% stay. Kept out of the score — used to test the model against your judgement.
Why this score
Priority moves
Rate the items to generate your prioritised retention plan.
Retention economics
Replacement cost defaults to 0.75× salary; the defensible range is 0.50–2.00× (SHRM: 50–200%, scaled by role level), before the cost of lost institutional memory. Preventable share defaults to 65% — deliberately conservative against Work Institute’s 2025 exit-interview base, which places roughly three-quarters of exits in employer-controllable categories. Adjust both to your reality.
Methodology & weighting How the Index is built ↓
The Index reads three layers. Stay-Ability is a weighted average of seven drivers (segment-specific weights). Flight-Pull is the mean of four external sub-factors. Criticality multiplies through to a Retention Priority Index that ranks where to act first.
Weighting
Weights are expert-judgement defaults, calibrated to African talent markets and flagged for empirical recalibration after pilot data. They are not derived from stated-reason frequency. The active profile is shown in the readout; full re-weighting is available in the enterprise engagement.
Flight-Pull · japa sub-model
Migration risk is real only when eligibility, motivation and mobility co-occur, so it uses a geometric mean — one missing precondition drags the score down:
Retention Priority Index
External pull is weighted above the internal stay-gap (0.6 / 0.4) because it is the harder lever to move; the 0.2 floor keeps low-impact roles from zeroing out. These are calibration parameters.
Bands
| Stay-Ability tier | Flight-Pull band | RPI band |
|---|---|---|
| Exposed < 2.50 | Low < 2.50 | Monitor 0–24 |
| Fragile 2.50–3.24 | Moderate 2.50–3.49 | Plan 25–49 |
| Resilient 3.25–3.99 | High ≥ 3.50 | Act 50–74 |
| Magnetic ≥ 4.00 | Escalate 75–100 |
v2.0 additions
Stay-intent anchor. Turnover intention is the strongest single predictor of actual exits in the meta-analytic literature (ρ ≈ .50, explaining roughly a third of turnover behaviour). It is deliberately kept out of the composite — intent is an outcome, not a driver — and is used instead as a divergence check between your judgement and the model. Stated intentions fail to translate into action about half the time, so divergence in either direction is information, not error.
Indicative risk window. Behavioural exit signals typically surface 3–9 months before a resignation. The window shown maps your priority band onto that horizon; it is indicative, not a forecast.
Preventable share. Exit-interview evidence (Work Institute, 2025; 120,000+ interviews) places roughly three-quarters of departures in employer-controllable categories. The economics default of 65% is deliberately conservative and editable.
Self-serve mode is single-rater: results are a hypothesis to test, not a measurement. Statistics cited elsewhere (Gallup, SHRM, Work Institute) are directional and refreshed annually (last: Jul 2026). Calibrate weights and thresholds to your segment before client use.
Get your full retention-risk breakdown
We’ll send your scored snapshot, the segment-specific 90-day moves, and a route to expert support. No spam — just your results.