Why the most common maintenance planning KPI is hurting your department
Key performance indicators (KPIs) provide control to management. Are we, in fact, doing what we need to do? Nonetheless, not all KPIs are useful. The most common maintenance planning KPI, Plan versus Actual (PVA), is actually counterproductive. By using it, we cripple maintenance.
KPIs give “control,” telling us if we are doing what we need to do. But KPIs are not the only form of control. We could do everything ourselves to ensure things are done correctly. We could have frequent meetings to make sure everyone is on the right track. We could also directly supervise everyone to make sure they do a good job. Another way to control would be to hire skilled persons who know what to do and turn them loose to do what we want done.
Each of these forms of “control” has its place, but reliance on KPIs is best used at the upper management level across different company areas. Among choices, the best form of control for planning itself is having skilled persons who know what to do. Nevertheless, PVA is commonly used with maintenance planning.
This KPI compares the estimated labor hours for a work order to the actual labor hours charged. A work order estimated for 5 hours has 10 hours charged to it. Should we deduce from PVA that the crafts take too long, or that the planner was not good at estimating? Therein lies the rub.
On one hand, if we are “grading” planners, planners that give extra time for all jobs get better scores because of Parkinson’s Law (PL: The amount of work assigned expands to fill the time available.) If a planner thinks that a job honestly should take 5 hours, the planner might plan it for 8 hours “just in case” something goes wrong or if the supervisor assigns a normally slow mechanic. Then, upon assignment, a normally fast mechanic gets an allowance of 8 hours to do a 5-hour job. Parkinson’s Law kicks in and the mechanic takes 8 hours. The planner gets a great score!
On the other hand, if we are grading craftspersons to meet the estimate, craftspersons get a better score if they ignore needed steps taking them past the allotted time. So, if we grade planners, we compromise productivity because planners inflate time estimates and we cannot assign more work. But if we grade craftspersons, we compromise quality because craftspersons cannot afford to take extra time when needed to do jobs properly. What a lose-lose KPI!
The PVA KPI is popular for two reasons. First, it seems obvious that planners should accurately estimate labor hours for jobs. Second, it is difficult to imagine another measurable planner KPI. But hyper-accurate labor estimating is not needed. It does not help us complete more work than normal. It does not help us do higher quality work.
Instead, let’s address the purpose of planning and scheduling. The purpose of planning, per se, is not to have hyper-accurate labor estimates. Instead, the purpose of planning is to help us have more helpful job plans over the years to improve work quality. Planners accomplish this bettering of job plans by incorporating craft feedback to make plans more helpful over the years. A better KPI would be ensuring planners use the job plan module of the CMMS to make living plans when they plan jobs. Do they?
And the purpose of scheduling is not to have hyper-accurate labor estimates. Instead, the purpose of scheduling is to help us complete more work than we would normally complete. This increase in productivity is accomplished by fully loading schedules to defeat Parkinsons’s Law. A better KPI would be ensuring schedulers fully load weekly schedules. Do they?
Nonetheless, we do need time estimates to manage work. We need them to help us build schedules and assign work. One method of estimating would be to build up the estimates considering job factors, such as how far from the shop, how many bolts, what motions are required, and so on. Adjustments could be made for factors, such as temperature and height. But we still do not know if a slow craftsperson or a fast craftsperson would be assigned. And we still do not always know the exact scope for reactive work. Another method would be to consider history. But Parkinson’s Law tells us that historical actual labor hours might be too high if enough work was not assigned at the time.
Instead of building up estimates or merely averaging history, planners simply making a reasonable judgment usually suffices for scheduling and assigning work. The planner presumes an assigned craftsperson generally has the skills and there will be no unusual problems with the job. Planners use their own experience, talk to certain craftspersons, or scan history for an idea of labor time. But in the end, planners quickly make a simple judgment. These time estimates vary widely in accuracy for individual jobs, but as many go over as under.
In the long run they average out for “accurate-enough” estimates making them useful for generally assigning work and building schedules. (This is what we wanted in the first place.) But because the accuracy varies widely, the time estimates do not replace the need for active supervision in the daily churn of maintenance with many jobs taking longer or finishing sooner than expected. We conclude that we need time estimates to manage work and planners can make useful time estimates. But there is no need for super-accuracy and there is great danger in holding planners or crafts accountable to exact times.
Let’s return to the issue of controlling planning itself. The primary control mechanism for planning itself should be to hire the right persons and turn them loose to do what we want done. Hire planners with great communication skills and data organizing skills. Great craft skills would be a plus to give them “street credibility.” Let planners freely and quickly make judgments for time estimates.
If we must measure planners with KPIs, grade whether they are using the job plan portion of the CMMS to make living plans. Thereby planners make “better” plans over the years to improve work quality. Also grade whether schedulers fully load schedules to 100% of the next week’s labor capacity. Such schedules defeat Parkinson’s Law to improve productivity.
Complete more work and better work than ever with proper planning and scheduling control. Don’t settle for being good plant. Be a great plant!
