How to Avoid Subjectivity in CMMS Decision Making

Subjectivity is the bane of precision data. And bad data can affect your ability to implement automated features which improve efficiency. In this overview I have identified 7 scenarios where subjective data is routinely used. I have also identified ways to overcome this subjectivity.

A. Work Order Priority

This value is sometimes entered by the operations staff. This field may have a choice list of 1 to 5 where 1 is emergency and 2 is urgent. Thus, the values of 3,4,5 are used to rank the entire backlog. For some organizations, the maintenance staff is allowed to override this value but this seldom happens. In the end you have a backlog of work that is poorly ranked. This means the planners are not really sure what to conduct formal planning on, thus they too end up guessing. The better approach is to set up a work order matrix. This matrix derives a numeral ranking based on the vertical categories, intersected by asset criticality. And by setting up a nightly cron task, an automation script can re-rank the entire backlog accommodating the latest work.

B. Rough Estimate

The person who first receives the new work from operations, could be multiple roles depending on the organization. I will call this role the gatekeeper. One of his functions is to enter a lead craft and a rough estimate. The accuracy of this estimate can be improved by reviewing past work for the same "type" and craft/skill. An average value (automation script reading historical actuals) can be displayed for the gatekeeper to apply or enter his own value.

C. Work Order Plans

The nature of the job requires a planner to estimate, which is subjective. But ... by using the following designs, a planner can dramatically increase accuracy. Planner accuracy depends a lot on this persons training, familiarity with facility/assets, job plan library, and access to work order history. Some planners were previously in the maintenance organization and therefore have a strong background. Planners can visit the job site to get a better idea of scope. A planner can also request a tech-assist whereby a technician visits the job site and reports back. Planner accuracy can be improved with regular work order feedback as to task wording, missing parts/tools, and hazard/precautions. Accurate job plan libraries and spare part listings can greatly improve work package creation. Lastly, the planner should use queries on the database to review completed work as to the size of the job, (a metric which looks at actuals versus planned where the ratio is significantly greater).

D. Weekly Maintenance Schedule

Some organizations have the scheduler sit with each craft supervisor mid-week to build next weeks schedule. The supervisor picks the jobs he wants to do (subjectively) and the scheduler operates the drag-drop scheduling tool. The supervisor also picks the day-of-the-week and assigns the worker. This process is essentially making 5 daily plans, in advance, which is a bad practice. This process is repeated for each craft supervisor. At end of the week, the scheduler conducts a final review of the weekly schedule and issues it. In summary, because the work order priority system is so weak it can't be relied upon for sorting/selecting work. And because there is no (automatic) resource-leveling tool in the CMMS, the weekly schedule build process is a resource consuming monster. The better solution is to implement an automatic resource leveling tool inside the CMMS combined with WO ranking.

The Truth About Scheduling

E. Asset Criticality

Many organizations just have a 1 through 5 ranking. The better approach is to create an application which calculates a business risk exposure (BRE) value. The first step is to identify critical systems and then conduct this BRE calculation. Asset criticality is also part of the WO matrix design.

F. Maintenance Strategies

The PM application inside the CMMS should not be confused with the "maintenance strategy". The best-in-class organizations will conduct RCM/PMO on their critical systems. The results of this analysis is sometimes kept outside the CMMS in proprietary software or Excel. The better approach is to create a new application inside the CMMS so that this can support a living program for RCM. Plus, you can achieve a 3-way match from the WO failure mode to the RCM app to the PM records. Maintenance strategies are not a one-and-done situation. They should be continuously refined and improved.

G. Chronic Failure Analysis

Chronic means recurring. It has been stated that up to 80% of all repair maintenance costs could be due to recurring failures. If this is true, reliability leaders need a way to find these bad actors. You can't work on everything at once. By creating a decent sort metric, you can find these worst offenders and then drill-down on the failure mode. My suggestion for the sort-metric is......... Average Annual Maintenance Cost divided-by Replacement Cost

It's up to you to be wise on how you use your core systems. The above designs provide the best results for decision making. Otherwise you are just guessing.