Having a formal scheduling method should improve wrench time dramatically. Formal scheduling answers the question "How much work should we do next week?" Scheduling is a matter of control. The responsibilities of management can be divided into Plan, Organize, Staff, Direct and Control. Unfortunately, as with most management, maintenance management does a much poorer job with the responsibility of control than with its other areas of responsibility.
The notion of control simply means comparing one's performance against a standard and making adjustments, if needed. In maintenance, we've abdicated this responsibility by not having a standard for productivity. We have expected "craft persons to work hard to keep the plant running and keep the backlog under control." This expectation is somewhat of a standard for quality, but worthless for productivity. Nearly everyone on the site knows whether we are doing a good job with plant availability, a quality standard. Indeed, we would rather have everyone sit around and the plant run than everyone be running around and the plant sit unavailable. Nevertheless, after quality, productivity must be addressed. First effectiveness, then efficiency.
Making a schedule for "How much work should we do next week?" provides an excellent and easy-to-use standard for productivity. The schedule figures how much work we should complete and then compares this quantity against the amount of work we actually complete. Companies can make this schedule creation and comparison easy or extremely complicated. Many maintenance organizations that advance beyond planning to scheduling falter here when they make their scheduling too complex. This column and the next few will discuss the proper principles to prepare and use a schedule as a control standard to improve maintenance productivity.
The first scheduling principle is the prerequisite of having a planned backlog. Maintenance planning supports maintenance scheduling by providing estimates of job hours and craft skill levels required for every work order. Planners must stay ahead of the crafts by planning enough work to provide a week's worth of work for scheduling. Typically, by adjusting the level of detail poured into job plans as more or fewer work orders come into maintenance, planners can plan nearly the entire backlog on an ongoing basis.
The estimated hours planned on work orders are entirely adequate for advance scheduling. Interestingly enough, this is in spite of the time estimates not being very accurate on individual jobs due to the nature of maintenance. A "simple" job planned for half a day might run into problems with rusted bolts or who knows what and take all day. On the other hand, many times technicians complete these same jobs in an hour or two. Maintenance is simply not as easily estimated as assembly line work.
Nonetheless, these actual hours vs. estimates have a very normal distribution (statistically speaking), and a stack of work orders encompassing a week's worth of work for a crew gives a very accurate estimate. The bottom line we need to understand is that the actual hours worked for a 10-person crew on 400 hours of planned work might be accurate to within 10 percent of the job plan estimates.
In addition to simply having the hour estimates, the planner must identify these hours by lowest qualified craft skill level. Instead of simply "20 hours," the estimate states "10 hours for mechanic, 10 hours for helper." This allows the schedule to know which crafts to assign in the scheduling process. Furthermore, note the crafts identified in this example are not "two mechanics for 10 hours each," even though this is a mechanical job. By identifying one of the craft persons as a "helper," the planner gives the scheduler and crew supervisor the flexibility by allowing anyone to be the second person on the job. If the scheduler came to the point of including the job and only had a mechanic and a welder left as a resource, yes, the welder could be the helper.
The first principle of scheduling simply states that maintenance plans must contain planned estimated hours for the lowest-needed craft skills. Well, I must admit that, so far, this is pretty simple. I've been told that having written the industry handbook on planning and having talked exhaustively about it for more than 13 years, we should expect a little rocket science. Sorry. While predictive maintenance might be high tech and high value, planning and scheduling remains low tech and high value. Rest assured, though, that it is high value.
This first principle is the first part of the framework we are establishing to build an advance schedule. We'll then use this schedule as a productivity standard to dramatically improve productivity.