The financial impact of unexpected equipment failure can be costly, as AK Steel Corp. shared recently. The West Chester, Ohio-based steel maker reported that an unplanned outage at one of its blast furnaces in June would contribute approximately $12 million to the company's loss column for the third quarter, with additional losses related to the mechanical failure trickling into the fourth quarter.
While the AK Steel incident has reaped particularly expensive consequences, every day manufacturers across the globe experience equipment failures. Machines stop running, operate at less-than-optimal speeds, or don't perform as expected in some other way.
Given the obvious imperative for better reliability, what metrics should the corporate reliability professional review to drive improved performance?
It takes more than one or two. "Metrics drive behaviors," says Jeff Shiver, managing principal of People and Processes Inc., a maintenance consulting and training firm. "If you look at one metric, you will box yourself into a corner and drive behaviors you didn't anticipate, and not necessarily good ones."
Moreover, reliability is a function of more than just the maintenance department, says Shiver. Production operations and engineering also play roles, he points out. Given their influence, "we have to work together as a whole team to improve reliability. It is never a good thing if we silo functions."
Shiver, whose background includes maintenance management at Mars' Waco, Texas, confectionary plant, outlined 10 maintenance metrics to drive better decision-making. They include:
- Maintenance cost per unit of production. Manufacturers must determine the right balance between running equipment to failure and spending too much.
- Maintenance cost as a percentage of replacement asset value. A world-class percentage is less than 2%, Shiver says.
- Preventive maintenance compliance. "We want 100% compliance, but what typically happens is that only about 60% are done, and only 20% to 30% are done right," Shiver says.
- Planned versus unplanned maintenance. A world-class level is 90% planned.
- OEE (overall equipment effectiveness) or availability. Aim for 85% OEE and/or 95% availability. While neither measure is strictly maintenance-oriented, Shiver says maintenance can influence both of them.
- First-pass yield or first-pass good. Typically considered a quality metric, first-pass yield can be influenced by maintenance, hence its inclusion on Shiver's list.
- Storeroom inventory value. A benchmark level is .5% to .75% of replacement asset value. Further, Shiver suggests inventory turns should number two to three per year. "I want to rotate the inventory because it ages," he says.
- Backlog of work orders in man weeks. Because companies calculate the measure differently, it is hard to identify an optimal number. If the correct practices are in place and the backlog is high, staffing may be an issue. On the other hand, operations decisions can drive the metric, in which case "it could could be a maintenance-operations partnership issue."
- Schedule compliance or "Did we do what we put on the maintenance schedule to do?" Aim for 80% to 85%.
- Critical events. "This is not a benchmark metric, but it is something a corporate guy should be looking at," Shiver says. As the description implies, a critical event is about big issues. It may be two-hour downtime on a production line or four hours on a packaging line, but it should spur a root-cause examination.
As a final caution, Shiver advises against rolling metrics up to a corporate level without the ability to examine plant-level details. "We need to be able to drill down and ask, 'Why?' " he says.