Ask the head of a mining group’s project office what one week of delay costs on the mill expansion, and then on the tailings facility raise. The answer is usually a rough figure for the month, not the week. Mining is precise about what lies in the ground: grade, tonnes, recovery, price per ounce. It is far less used to pricing time. Yet November is when many groups settle next year’s capital plan, and every line of it is a bet on weeks that nobody has priced.
Portfolio tools should help here. Enterprise project portfolio management tools, usually called EPPM software, list every study, expansion and shutdown with its budget and status. The harder test for any EPPM software is whether it shows which projects wait for the same few engineers, and what each week of that wait costs.
The loss starts in the plan
McKinsey analysed 80 mining projects from the past decade and found real schedule delays averaging about 25 per cent and real cost overruns of about 40 per cent. It traced roughly two thirds of the delays and overruns to poor initial assessments, and one third to poor execution. That points at the plan. A plan that assumes the right engineers will be free in the right month is an assessment too, and few plans test the assumption.
The people who make the assumption true are a thinning group. EY’s Canadian mining practice wrote in February that over half of the workforce in the United States and Canada is expected to retire within a decade, leaving gaps in critical roles such as planning, process engineering and sustainability. Every new study and expansion draws on that same layer.
Not every week costs the same
Portfolio reports tend to treat delay as a straight line: one week late, one week worse. Mining work does not behave that way. Some work has float, and a week of slip costs almost nothing. Other work is pinned to dates that do not move: a plant shutdown booked months ahead with contractors and parts, a wet season that closes earthworks, a shipping slot, an offtake commitment.
Take a simplified example, with invented numbers. A group has two projects waiting for the same two senior engineers. The first is a throughput upgrade at a mill, adding 3 per cent to output. At a copper price of US$10,000 a tonne, a mine producing 150,000 tonnes a year earns about US$29 million a week, so each week of delay on the upgrade costs about US$0.9 million of extra output.
The second is the design package for a tailings storage raise, which must be approved before the wet season shuts earthworks. Slip a week with float to spare and it costs a week. Slip it past the start of the wet season and the raise is not ready when storage runs short. The mill then has to slow, and every week of slowdown puts a share of that US$29 million at risk.
Both projects go on the list as critical. Put a price on each week and the order is clear.
Each site plans for itself
Every asset team plans its own projects, and each plan assumes the planning and process engineers will be free when needed. Each plan is defensible. Laid side by side, they often ask for well over 100 per cent of the same people in the same quarter. Nobody sees the total until a queue forms at the engineers’ desks.
Putting the price next to the date changes what a group can do about it. When dates move with real engineer capacity, the cost moves with them, and a group can see the penalty coming while there is still time to prevent it. On a sample portfolio of late projects and overloaded teams, a $330,000 loss became nearly $1.3 million of profit. Nothing about the work changed. The decisions did.
Spend where the week is expensive
Mining is rightly interested in AI, and AI tools that prepare first-pass technical reviews or study data can pay back. They pay back when they sit in front of the expensive week. Free a tenth of a bottleneck’s time with first-pass reviews and the portfolio can deliver up to a tenth more with the same people. Anywhere else, the tool saves effort that nobody was waiting for.
Before buying, ask three questions about the work you plan to hand over. Is anything waiting on it? Does the wait cost real money? Can a machine do it well?
Three steps before the 2027 capital plan is signed
First, name the desks every project waits on. Look at last year’s late projects and ask whom they waited for. It is usually the same few names.
Second, write a price next to each date. Use the figures the group already holds: output per week, shutdown and contractor bookings, shipping slots, contract penalties. A rough number beats a colour.
Third, give the scarce hours to the expensive week, and price it again every week. Forecasts move, so the price should move with them.
Mining prices the tonne to the cent. The week an engineer keeps a project waiting is the cost it writes down least.
Dr Albert Ponsteen is CEO of Epicflow, a project portfolio platform that engineering organisations including Russell Mineral Equipment, a mining equipment manufacturer, and Fokker Services use to plan their projects and people. His PhD research at the Open Universiteit examined resource allocation in multi-project environments.
