Was Predictive Maintenance Overhyped? ??

That's a fabulous article Olivier Bloch !! Thank you for sharing it .

Many valuable comments already. I think it’s also due to the nature of trying to predict an event that might occur in the (far) future. Unlike predictive quality, you might see the real impact of a working predictive maintenance not before a year or longer has passed. And that makes it tough to justify to pay a fee now for something that might not even detect early signs of a failure. Long story short, I think PdM would really benefit from an insurance like business model where you use PdM to guarantee uptime.

Olivier that's a realistic take on #predictivemaintenance. You are totally correct that, generally, (extensive) failure data is required to come up with adequate models. But so much more is required to make it a solution that brings real ROI. The following way of thinking worked in some cases I observed (treat it as a journey over digitalization/smartification maturity): 1. (I)IoT connectivity to collect relevant signals and provide easy access to the relevant users. Easy access to that data was a crucial part. This alone can provide unexpectedly useful insights - like that a pump was being unplugged every weekend. 2. Easy-to-use framework for the relevant users to define rules for alerts based on their data - so that they apply their experience and try getting upfront signals when something might go wrong. 3. Framework for specialized users (like data scientists) to provide more advanced "rules" to enable actual predictive maintenance. That being said, jumping directly to this third point could be that very hype.

It wasn't the fault of predictive maintenance in my opinion. Predictions alone don't mean production and performance will skyrocket. What's needed is an integrated program to go along with the predictions and prevent failures. That requires money and effort. There are no silver bullets.

Great read!? Actionable insights are the missing link in many IoT projects. The key is connecting IoT data with the broader company ecosystem to generate meaningful outcomes. In one of our cases, we integrated IoT signals with service manuals, which then directly links to spare parts.? The integrated approach allows technicians to understand what's wrong immediately, see an exploded view of the relevant part, and access all the critical information. IoT is just one piece of the puzzle, it's a technology, not the end goal. Its real value comes from how it integrates with other technologies to solve user problems effectively.