Equipment maintenance is 30 - 50 per cent of the operating costs for a typical mining company. While this is substantial, its importance is beyond its significance as a cost heading. Modern mining requires large amounts of investment in capital infrastructure and equipment to produce a product sold at a low unit cost. To stay competitive, such a business has to focus on full realisation of its capital investment, ie sustained production at high levels. Maintenance plays an essential role in achieving this goal. Mining companies typically use a mix of breakdown or corrective maintenance, planned maintenance, and some condition-based maintenance. The share of breakdown maintenance has been coming down through more effective planned maintenance and better condition monitoring tools. Planned or preventive maintenance (PM) is playing an increasingly important role in today's mining maintenance. This is the suite of inspection and repair/replacement actions performed at scheduled time intervals as part of a preplanned activity. The activities are scheduled well in advance and, usually, the current wear-state of the machine is irrelevant. For example, the replacement of a major gear may have been planned 18 months in advance; replacement is carried out when the time comes regardless of the actual condition of the gear. The PM costs are significant and include direct cost of parts and labour as well as the opportunity cost of lost production while the machine is under maintenance. It is important to identify the optimum PM effort because over-maintenance should be avoided as much as under-maintenance. A new concept called severity-based maintenance is introduced in this paper. We have seen that measuring equipment ageing in mining merely by counting the engine hours is a very imprecise way of planning and scheduling maintenance because the damage accumulated in one hour varies enormously from operator to operator and with changing pit conditions. We recommend continuous monitoring of duty on all major equipment. This is feasible without having to fit out the machine with an army of sensors because most of the physical sensors required by dutymeters are already installed and maintained as part of the existing performance monitors. While the investment is modest, the potential benefits are huge and can be grouped under three headings:Immediate benefits will come from providing operators with direct feedback on the effect of their behaviour on machine health and training them to use this feedback to correct their operating style.The equipment can operate closer to the limits of capacity, producing more without unduly sacrificing the availability by monitoring and controlling the severity of the operation, ie the duty on the machine.Longer term benefits accrue from establishing correlations between actual time to failure data and measured equipment duty and using these correlations to optimise maintenance effort as well as planning new equipment purchases and warranty conditions.