A common phrase is “What’s measured is managed”. What’s also true is that measuring the wrong things can drive unintended behaviors. We want to be deliberate in choosing the metrics that we track for maintenance. We want to communicate what our objectives are, such as increasing efficiency, schedule attainment or improving repair quality. It is important to make sure we pick the right metrics so that we can measure how well we are achieving these objectives.
The first step is understanding the types of metrics
There are 3 general types of metrics: Indicators, Process Indicators and Key Performance Indicators:
- An Indicator is a measurement used in a work process that gives information about the process. For maintenance this could be the number of actual hours available for work in a past week.
- A Performance Indicator (PI) is a measurement that gives information about the relative performance of a work process. For maintenance this could be size of the maintenance planner’s backlog. Along with other PI’s, this PI gives insight on the appropriate number of planning resources.
- A Key Performance Indicator (KPI) is a measurement that gives information if the work process is achieving a desired result. For maintenance this could be the amount of rework (premature failures due to repair quality). This KPI is a measure of the quality of field repairs.
Metrics are also defined as either Leading or Lagging. A Leading indicator predicts the outcome of the maintenance work process. A Lagging indicator measures the outcome of the maintenance work process. An example of a Leading indicator is Preventive Maintenance (PM) Compliance. PM Compliance describes how much of a site’s PM program is completed as scheduled. A high PM Compliance would tend to indicate lower amounts of anticipated unplanned downtime. It doesn’t correlate exactly as the PM’s have to be effective in both content and execution. Unplanned Downtime is a Lagging indicator as it shows actual performance.
The next step is to review the site’s maintenance objectives
Sites typically have a description of their maintenance journey showing the current state and future state with key objectives along the way. As an example, a site may say that their current state is that maintenance execution basics are in place and the future state is maintenance execution is delivering first quartile cost performance. To achieve the future state, the site’s objectives include improved planning, improved scheduling, and a reduction in maintenance demand due to equipment failures.
We want to select a mixture of Indicators, PI’s and KPI’s to give us both the granularity to identify issues and the overview to understand overall performance. We also want a mixture of Leading and Lagging indicators so that we can understand potential issues that are developing as well as issues currently occurring.
The final step is selecting metrics that will drive behaviors which are aligned with those objectives
In our example, we have an objective of improved planning. A potential metric for maintenance planning is the # of plans/planner. The intent of this metric is to help identify issues in planner productivity. The metric does not account for the amount of time that is needed to plan different types of maintenance activities. An instrument troubleshooting activity takes significantly less time to plan compared to retubing a heat exchanger. If we decide to use this metric, we need to set different targets based on the types of activities being planned. An undesired behavior that we could incentivize with this metric is producing plans quickly without the desired quality. Therefore, we would also need to have metrics to measure the quality of the plans. Potential metrics for plan quality include the # of plans requiring changes during execution, estimating accuracy or the # of materials consumed that were not identified in the job plan.
After selecting the appropriate metrics, the proper audience for each metric needs to be determined. This is the first step in building maintenance dashboards. Matching dashboards to the appropriate audience and how to use metrics and metric trends will be the subject of another blog.
In Becht’s Maintenance Guide, we’ve identified 7 KPI’s and an additional 29 PI’s that are commonly used in industry for the maintenance work process with definitions, intended behavior, frequency, expected trends and stewardship. If you have trouble envisioning how to get started, Becht has an experienced team of maintenance professionals who can help with assessing your situation and helping you define the path forward.