Don't miss MaximoWorld 2024, the premier conference on AI for asset management!

Experience the future of asset management with cutting-edge AI at MaximoWorld 2024.

Sign Up

Please use your business email address if applicable

From a Different Angle: A Perspective - Predictive Maintenance or Detective Maintenance?

Are you a big believer in oil analysis, vibration analysis, infrared and the use of various kinds of ultrasound? Are you in favor of the tons of new technologies that let you look into aspects of a machine and see its condition in new ways?

So, what is the problem? The problem is with imprecise language.

How about predictive maintenance (PdM)? If you think it’s hogwash, snake oil, you’re not alone. Only in the last year or two, using big data analysis and Watson-like tools, is any of PdM providing predictions worth the paper they are printed on. So, what is the problem? The problem is with imprecise language. The problem is when words are recruited to mean something different than what they actually mean. When you say something where the meaning is not what the words mean at closer look, you will (intentionally) confuse people.

Is infrared scanning really as its dictionary meaning of “relating to or having the effect of predicting an event or result?” Think about it from your experience – you can predict a failure from a scan, but the scan itself is in no way a predictive.

figure 1

Figure 1

Perhaps the name of predictive maintenance should be changed to detective maintenance. What is it that you are actually doing? Let’s take a look at an infrared picture (see Figure 1). Sure, you could argue that the middle wire will fail. It might be true, but it is not accurate.

What if it is said that infrared is a way to detect issues that could lead to problems? So, Figure 1 is actually a condition that might lead to problems. Therefore, you have successfully detected potential problems with no bogus prediction involved.

How many times have you faced down meetings where your toes were held to the fire, so to speak, because a prediction from your predictive maintenance didn’t turn out? Executives love it when the maintenance folks scream the sky is falling because their predictive gear says it.

Consider that scenario versus you have detected an anomaly and plan to investigate, fix it, watch it, sit on it, whatever. It sounds more credible, doesn’t it?

Someday soon, there really will be predictive maintenance systems that will text you, “Get out of the area if you want to live” or something equally clear.

Joel Levitt

Joel Levitt, CRL, CPMM, CRL, CPMM, is the President of Laser Focused Training. Mr. Levitt has 30 years of experience in many facets of maintenance, including process control design, source equipment inspector, electrician, field service technician, maritime operations and property management. He is a leading trainer of maintenance professionals and has trained more than 17,000 maintenance leaders from 3,000 organizations in 25 countries in over 500 sessions. Since 1980 he has been the President of Springfield Resources, a management consulting firm that services all sized clients on a wide range of maintenance issues. Prior to that Mr. Levitt worked for a CMMS vendor and in manufacturing management. 

He is also a frequent speaker at maintenance and engineering conferences and has written 6 popular maintenance management texts and chapters of 2 additional reference books. He has also published dozens of articles on the topic. Mr. Levitt has served on the safety board of ANSI, Small Business United, National Family Business Council and on the executive committee of the Miquon School. He can be reached at or visit

ChatGPT with
Find Your Answers Fast