A 30-year field service veteran at Xerox discusses his early (and decidedly low-tech) experiment with predictive maintenance nearly two decades ago.

Recently, I read an article about how GE is using predictive maintenance in its power plants to maximize output and to keep downtime at a minimum.

It reminded me of an experiment I did nearly 20 years ago. I flipped through the log books at each site of the machines I was responsible for servicing. Averaging the time between service calls for each copier, I planned an impromptu visit a few days before my simplistic analysis told me there was a good chance I’d hear from that customer. The results were stunning.

Some customers accused me of being psychic because they “were just getting ready to place a service call.” In no time, I was doing more self-initiated service calls than customer-initiated calls. My low-tech experiment proved two things: predicting service calls is possible, and customers love when it happens.

In no time, I was doing more self-initiated service calls than customer-initiated calls.

I realize my situation was unique. I worked on machines that had a high call-rate and were maintenance-intense, so there was always something I could work on when I arrived. The odds were in my favor. Advancements in technology have made machine components immensely more reliable, which means fewer service calls and less downtime. Let’s take a look under the hood and see what makes it work.

Sensors and Software on the Cheap

Sensor technology and predictive software have taken my idea way beyond human limitations. A sensor can detect an increase in heat generated that indicates bearing failure. Software can analyze the slight drop in efficiency that a failing motor exhibits before it seizes up. This isn’t breaking news. Engineers have designed machines with sensors that indicate failure for a long time and the same goes for predictive software, but the problem has been that the cost of these solutions has been prohibitive. The real news is that costs are coming down as the demand for sensors and robust analytics skyrockets.


“The good news is that the cost of many new technologies, such as remote monitoring equipment and analytical software, are continuing to fall,” writes energy journalist Heidi Vella.

A Big Bump in Customer Satisfaction

As I found in my experiment, when you show up at a customer’s door and tell them you’re replacing the motor in a machine that could bring their business to a screeching halt before it fails, customer satisfaction goes through the roof. Combine sensor technology with analytics and you can get an accurate enough forecast to give the customer enough time to plan for the downtime. But increased customer satisfaction is not the only benefit.

Where Service Providers Will Benefit

Monitoring of sensors and software is an open door to service contracts. Service providers, and the customers they serve, will be the true winners of this coming trend. Why buy a product loaded with sensors and software that can predict when it’s about to fail if you’re not going to pay to be connected to a network that can alert someone who’s trained to fix it? Predictive maintenance becomes a great way to convince a customer to sign up for a service contract and join your IoT network—for a slight fee, of course.

Predictive software and hardware soon will become the norm, opening the door for manufacturers to offer innovative new services. Inevitably, customers will become so accustomed the technological advancements that they will begin to take predictive maintenance for granted. Manufacturers and service providers alike will have to respond, but that’s just part of keeping up with customer expectations.