Insights — Process Control
May 2026 · 6 min read
Ask a mine control team if their dig-line deviation or payload distribution is "under control," and most will say yes — nobody's complained, the monthly average looks fine, and reconciliation hasn't triggered a review. None of that actually answers the question. An average can look stable for months while the underlying process swings wildly from shift to shift. The average just doesn't show you the swing.
A process can go quiet for reasons that have nothing to do with stability: a favorable ore zone, a run of experienced operators, a maintenance window that happened to land at the right time. Quiet means nobody noticed a problem this month. Controlled means the process behaves the same way, within predictable limits, regardless of who's on shift or what zone is being mined. Only one of those is something you can rely on.
A control chart plots individual measurements — dig-line deviation, payload per load, cycle time — against a center line and a pair of control limits calculated from the process's own historical variation, not from a spec sheet. The chart isn't asking "is this within spec?" It's asking a more useful question first: is this process behaving consistently?
Two signals matter most: a single point beyond the control limits (something unusual just happened — a special cause worth investigating), and a run of points showing a pattern — several in a row trending, or clustered on one side of the center line (the process itself may be shifting, even though every individual point still looks "fine").
Even a process that's perfectly in control — stable, predictable, no special-cause signals — can still be incapable of meeting the specification it needs to hit. Control tells you the process is predictable. Capability (measured through indices like Cp and Cpk) tells you whether that predictable process actually fits inside the tolerance the operation requires. A stable process centered outside spec is still a problem — it's just a consistent one, which is actually easier to fix than an erratic one.
The gap between "quiet" and "controlled" is exactly where reconciliation surprises come from. A payload distribution that looks fine on a monthly average can be masking a bimodal pattern — one shift consistently under target, another consistently over — that never trips an alarm because the average of the two looks acceptable. The same logic applies to HPGPS positional accuracy, cycle time, and ore-control tonnage: the number that gets reported monthly is rarely the number that would flag the real risk.
The fix isn't more inspection — it's the right chart, on the right variable, reviewed on a cadence the team actually follows. That's a small operational habit with an outsized effect on how many surprises show up at month-end.