Workforce Availability Intelligence: Fit, Qualified, Rostered, On-Site
- Maturity
- L4
- Domain
- Plan & Cost
- Analytics
- predictive
- Workforce availability rate
- ~91%
- Sponsor
- Chief Operating Officer
- Confidence
- High
The situation
Shift by shift and site by site, will IronPeak have a fit-for-work, qualified, rostered, on-site crew sufficient for planned production — and where will availability fall short before it happens?
The recommendation on the table
Stand up site-shift availability forecasting with a pre-shift continuity-risk flag
Production losses anticipated and mitigated days ahead, not discovered when the roster fails.
Trade-offRequires roster, attendance, travel, qualification and fatigue data joined reliably at shift grain — the portfolio's hardest integration.
The evidence
IronPeak loses production not because it is under-headcounted but because rostered crews turn out to be unavailable — fatigued and stood down, on leave, travel-disrupted, weather-stranded, or lost to contractor churn — and at a remote site there is no nearby bench. The Mining flagship, MIN-02 forecasts availability — fit-for-work, rostered, qualified, on-site — against production requirements at the roster/shift grain, flags shortfalls before the shift, and optimises rosters and mobilisation. It found swing-change gaps, structural availability erosion from churn and fatigue, and remote sites that cannot mobilise fast enough to fill forecast gaps.
Workforce Availability Intelligence
Match fit-for-work, rostered, qualified, on-site availability to production at the site-shift grain.
| Day | Night | Swing-change | Weekend | |
|---|---|---|---|---|
| Pilbara | 1 | 2 | 3 | 2 |
| Highlands | 1 | 2 | 2 | 2 |
| Lithium-W | 2 | 3 | 3 | 2 |
| Underground-E | 1 | 2 | 2 | 1 |
Key takeawayShortfall intensity (3 = frequent). Swing-change shifts run ~9% below required available crew.
- Structural (churn + fatigue + roster)40%40%
- One-off (travel/weather)60%60%
Key takeaway~40% of one site's availability loss is structural — fixable, not bad luck.
Key takeawayThe remote lithium site runs one disruption from a production loss.
Key findings
Workforce availability — fit-for-work, rostered, qualified, on-site — runs around 91% against plan, with swing-change shifts ~9% below required crew, yet it is the binding production-continuity variable and is almost unmeasured today.
What we can’t claim
About 40% of one site's availability loss is structural (churn, fatigue, roster design) rather than random disruption, and the most remote site can mobilise only ~68% of forecast gaps in time. The inconvenient truth is that establishment, capacity and availability are three different numbers: a site fully staffed on paper can be red on availability, and the fixable losses are the roster and churn ones, not the weather.
Recommendations
Stand up site-shift availability forecasting with a pre-shift continuity-risk flag
high priorityProduction losses anticipated and mitigated days ahead, not discovered when the roster fails.
Trade-off
Requires roster, attendance, travel, qualification and fatigue data joined reliably at shift grain — the portfolio's hardest integration.
Redesign the rosters that structurally manufacture availability gaps
high priorityRecovered structural availability and fewer one-disruption-from-loss sites.
Trade-off
Roster change is industrially sensitive and must be done with the workforce, not to it.
Analytical framework
How we reached this
Predictive availability matching — forecast fit-for-work, rostered, qualified, on-site availability and match it to production at the site-shift grain, protecting continuity.
ConfidenceMedium-High
Analytical framework
How we reached this
Predictive availability matching — forecast fit-for-work, rostered, qualified, on-site availability and match it to production at the site-shift grain, protecting continuity.
Methods applied
Statistical techniques
Algorithms
Data sources
Outputs generated
Why this confidence
Roster, leave and attendance data are operational and tractable, giving good forecast accuracy on stable rosters; bounded below High by contractor-churn volatility and travel/weather disruption, shown as widened bands. Fitness and qualification are hard gates.
The reasoning
Business context
The flagship and clearest expression of Mining's thesis: availability — not headcount, not capacity — as the binding production-continuity variable. Availability = Fit-for-Work + Rostered + Qualified + On-Site; a person counts toward a shift only if all four hold. Sponsored by the COO because availability directly determines continuity. It consumes fatigue signals from MIN-03 but does not own fatigue.
Expected value
The largest value pool in the Mining portfolio — avoided production losses from availability gaps, lower premium mobilisation cost, and improved continuity at fragile sites. IronPeak measures equipment availability obsessively and workforce availability barely at all, yet the workforce is the binding continuity variable.
Workforce landscape
Swing-change shifts at two sites run ~9% below required available crew (a roster-design gap); ~40% of one site's availability loss is structural (churn and fatigue); the most remote lithium site can mobilise only ~68% of forecast gaps in time; high-contractor sites forecast to ~81% vs. ~92% on stable rosters.
The analytics journey
Level 4, predictive. Forecasts site-shift availability with confidence and matches it to production under a hard fit-for-work and qualification gate. Honest that contractor churn and travel/weather disruption add variance, shown as widened bands; fitness and qualification are never traded.
Under the hood
Interpretable, roster/calendar-aware time-series forecasting projects availability; linear programming optimises rosters and mobilisation under a fit-for-work and qualification constraint; coverage gaps are classified roster-float / mobilisation / unfillable. Availability is modelled as the conjunction of four gates sourced from four systems — fitness/fatigue, rosters, credential register, and site access.
Confidence & evidence
Why you can rely on this
The inconvenient truth
About 40% of one site's availability loss is structural (churn, fatigue, roster design) rather than random disruption, and the most remote site can mobilise only ~68% of forecast gaps in time. The inconvenient truth is that establishment, capacity and availability are three different numbers: a site fully staffed on paper can be red on availability, and the fixable losses are the roster and churn ones, not the weather.
Method
Confidence is a deterministic read of KPI strength, target and benchmark coverage across this project — shown on an illustrative reference dataset, computed the same way it would be on live data.
Take this further