Run the line where it should live.
Find and lock in the operating points that lift yield, cut waste, and lower energy per unit — using the data you already have, the engineering you already trust, and trials small enough to run without stopping production.
The metrics worth optimising — and how the data sees them.
Yield
Find the operating points that produce more good parts. Statistical process control on the metrics that matter, with the operating context to explain the variance — not just flag it.
Waste
Every dollar of waste leaves a fingerprint in the data. Bucketed against the parameters that drive it, your team can see what to push on — and what to leave alone.
Energy
Tie every unit of energy to a product, a batch, and an operating regime. Benchmark lines against themselves, lines against lines, and shifts against shifts.
Quality drift
Process-side analytics see drift hours before the inspection report does. Detect, contain, and route to engineering — without the batch ever leaving the line.
Driver ranking that survives engineering scrutiny.
A correlation isn't a cause. We pair statistical attribution with engineering review so the variables you push on are the variables that actually move the outcome — not noise that happens to track it.
- Driver ranking by partial dependency, not raw correlation
- Operating-regime segmentation so seasonality isn't mistaken for cause
- Engineering review of each candidate before it gets trialled
- Sensitivity bands so the trial size is calibrated to the signal
Baseline · Model · Trial · Lock in.
Optimisation that respects the line. No year-long studies, no high-risk setpoint changes — small, reversible moves with measured impact, each one earning its place in the recipe.
Baseline
Three weeks of clean operational data with the variables and outcomes the team agrees matter. No optimisation yet — just an honest picture of where the line lives today.
Model
Engineering-aware models of the process. Setpoints, raw materials, ambient, and equipment state correlated to the outcomes that drive cost and quality.
Trial
Test-and-learn cycles run on the line with engineering oversight. Small, reversible setpoint moves with measured impact — not a year-long capex programme.
Lock in
New operating points written into recipes, standard work, and control limits. SPC rules updated; outcomes monitored continuously.
Better yield, less waste, lower energy — locked into the recipe.
A continuous chemical line gained 6 points of yield from three setpoint moves.
Twelve weeks from baseline to lock-in. Three reversible setpoint trials, each measured for two production cycles. The third move was rejected on engineering review; the first two are now the recipe.
Strongest as part of the chain.
Move the metric. Lock in the move.
A 30-minute working session with a process engineer — bring one line and one outcome, and we'll show you what the four-phase method looks like for it.