The plant had everything it needed on paper. A world-class ore body. A purpose-built processing facility. A team that understood the chemistry and the geology. What it did not have was a production profile that matched the investment.
The ramp-up to target throughput had been slower than planned. Output was erratic, swinging between strong days and poor ones with no predictable pattern. The variability was not just a production problem. It was a credibility problem: every missed forecast eroded confidence in the operation.
There was also a capability gap the team could feel but had not been able to close. The operation generated plenty of data, but the skills and routines to turn that data into operational decisions were not yet in place. People were reacting to what had happened rather than planning for what should happen next.
The conversation on site was centred on the processing plant’s metallurgical performance: recovery rates, throughput variability, the gap between actual and nameplate capacity. Reasonable targets for attention. But they were symptoms, not causes.
What we saw was a system designed for commissioning, not for sustained performance. Planning was not integrated across functions. Scheduling was informal. There was no consistent rhythm connecting what happened on the floor each shift to the decisions in the daily management meetings. And the performance measures people were managing against were not connected to the levers they could actually influence.
The plant did not need a technical fix. It needed an operating model. One of the onsite metallurgists put it simply: “They helped us debunk the myths.” The myths were the assumptions about what was limiting performance. Most pointed at the process. The reality pointed at how the process was being managed.
We worked in two connected phases with skills development running throughout. The first was a rapid improvement sprint: identifying the real constraints, developing the BI team’s analytical capability, and implementing focused initiatives around flow stability and recovery optimisation. Rather than trying to improve everything, we concentrated effort on the constraints with the highest leverage.
The second phase was the full operating model. We built the management system from the ground up: work validation and planning routines, an integrated schedule giving every function a shared view of priorities, a tactical resourcing process, and a structured performance review loop. We defined levels of work for every role and built day-in-the-life routines that made the operating model visible in each person’s calendar, not locked in a document on a shelf.
Coaching and support ran throughout both phases. We worked alongside supervisors and managers on the floor and in the daily meetings until the routines were embedded and the team was confident enough to run them independently.
Within the first three months, plant output improved by approximately 60%. The sprint had delivered the initial step-change by removing the most immediate constraints and giving the team a clear, data-driven focus.
As the full operating model took hold, the improvement deepened. Processing plant output increased by approximately 68% from baseline, with variability dramatically reduced. The operation moved from erratic performance to consistent daily throughput at a level the plant had not previously sustained.
The shift that mattered most was in how the team operated. The BI team was now running the data analysis and feeding insights into the daily management process. Supervisors were planning with a discipline and confidence they had not had before. The plant was no longer dependent on individual heroics to hit its numbers. It had a system, and the people who ran it understood why it worked.
Most operations stall in the gap between commissioning and sustained performance. They have the asset, the people, and the technical knowledge. What they lack is the operating rhythm, the integrated planning, and the clarity of roles that allow performance to become reliable rather than occasional.
The question for any operation underdelivering against its design capability is not whether the plant can do it. It almost certainly can. The question is whether the organisation around the plant is set up to let it.
