KPI’s for CBM / PDM programs can and should be a useful tool, but used incorrectly or manipulated to make things look better than they are can destroy a program. I am not going to cover all of the KPI’s that are out there in the maintenance world but more how they interact with each other and the effects of overlooking one in favor of another one.

• Route adherence: This can be expressed as either was the route completed (yes or no) or how much of the route was completed (expressed as a percentage completed) or how many of the assigned routes were completed for a given time frame.
• Repair Orders Generated: How many of the repair orders in the CMMS system were generated by the CBM / PDM methods.
• Repair Orders Completed: How many of the repair orders were completed in a given time frame.
• Utilization Rate: How many of the available hours was the process running (also known as “uptime”).
• Down Time Cost: All costs from the downtime (lost production, lost sales, lost customers, customer satisfaction losses, repair cost, man-hours, and any other cost incurred from the process not running).
• Maintenance Costs: This can get tricky very quickly, the indicator that we are looking for here is the cost of “planned and scheduled” work generated by the CBM / PDM programs to include the PM (preventive maintenance) expressed as a ratio of asset value to maintenance cost.
• Maintenance Man Hours: What is tracked here is the man-hours spent on the scheduled work generated by the CBM program vs. emergency, unscheduled or general work.

These are just a few of the KPI’s that are used to “evaluate” a CBM/PDM program. For these figures to have any real meaning they should be used as supporting data and how they interact with each other affecting the overall picture. But, they also need to be defined as to what the figure represents.

For example:

Route adherence: Defined as the percentage of completion for each route or how many routes were completed regardless of each routes’ % completed. Here is an opportunity for manipulation of the data to make things “appear” good. Data collected on machines that are not running under “normal” loads, or at all, but collected anyway, making the percentage higher! (Not a good practice but does happen to make the numbers look good.)

Repair orders generated: How many “real” repair orders are created. Here again, is an opportunity for manipulation of the data to make things “appear” good. Orders to repair defects found during CBM activities are valid. Repair orders to increase monitoring or requesting additional testing would not be valid.

Repair orders completed: How many “quality” repair orders were completed in a given time frame (30 days, 60 days, 120 days)? There should not be any over 120 days. Why? Are the 120+ day orders “bad” or not valid? Those orders need to be evaluated and closed as not appropriate or resubmitted with fresh data. Here is another opportunity for manipulation of the data to make things “appear” good. Orders can just be closed with no work actually being completed.

Utilization rate: How many of the available hours does the unit actually produce a product that can be sold for profit. Not rework, not the setup time, not change over time, not clean up time, not break time, not repair time, just time actually producing product vs. how many hours are available to run. If the plant runs a 24/7 schedule (24 X 365 = 8760 total available hours). If you can increase this “uptime” while keeping maintenance cost the same, you increase the bottom line. Or, even if the “uptime” cannot be increased but the maintenance cost is reduced, this will increase the bottom line. Here again, is an opportunity for manipulation of the data to make things “appear” good.

The opportunity for the manipulation of the data to make things “appear” good, exists and can be a very easy trap to fall into. This can be one of the biggest problems with KPI’s. for example, some people will manipulate the data to make their parts look better, some will manipulate the data to make others look worse. Human beings are inputting the data, this part of the equation must be controlled, and entered as it truly is, not how we would like it to be. As much as we would like to believe we are above doing something like manipulating data to make ourselves look better. Now do not get me wrong, there are people out there that would not manipulate the data, as a matter of honor, self-respect, upbringing, or some other moral point. It is a variable that all of the “maintenance models” assumes that the data is correct. Keep in mind that KPIs can be useful, but be aware that the data can be manipulated, due to the “human factor.”

We as Reliability Maintenance Specialists need to remain unbiased in our presentation of data to management and or customers.