Maintenance Mindset: Why oil analysis fails for most — but automation can change that
Welcome to Maintenance Mindset, our editors’ takes on things going on in the worlds of manufacturing and asset management that deserve some extra attention. This will appear regularly in the Member’s Only section of the site. This week's column features guest contributor Michael D. Holloway, President of 5th Order Industry.
Every year begins with great intentions. Dreams are formed, resolutions set, and plans meticulously laid out. Yet, as February arrives, those dreams fade, resolutions break, and plans shift. Life happens. The treadmill or exercise bike, once a symbol of commitment, becomes an obstacle in the garage or basement. The idea was great—until it wasn’t.
Ideas without habitual reinforcement and clear value rarely stick. Oil analysis shares a similar fate. It’s an incredibly valuable tool—when used consistently and correctly. There are regional as well as global oil analysis laboratories who celebrate their technical leadership and analytical prowess. They deliver exceptional insights into used lubricating oils, offering benefits such as extended equipment life, reduced operating costs, and minimized waste. The return on investment is undeniable.
Yet, when a global leader at one of these companies asked me, “How many of our customers actually use our work to improve reliability?” I replied honestly: “Around 15%.” He was surprised, even frustrated. But I stood by my answer, backed by years of observing customer behavior and motivations.
Oil analysis, much like an exercise bike, is only effective if it’s used regularly and with purpose. The leader’s follow-up question still lingers: “Then why do we do it? Why are we in this business?” My answer was simple: “For the 15% who do use it.”
I’ve been to many oil condition laboratories across the globe. These labs function like factories: dirty, used oil serves as raw material; sophisticated scientific instruments act as production equipment; and the final product is data paired with diagnostic commentary—what we might call “expert opinion.”
If tasked with designing the next generation of oil analysis businesses, the answer becomes clear: automation. Automate sample processing, automate analysis, and automate diagnostics. Robotics can already handle sample intake and analysis with precision and consistency. It’s not a question of if but when automation will dominate the industry.
Envision this: a lab where a receiving clerk unpacks samples, and an articulated robotic arm registers and queues them for analysis. Samples move seamlessly through automated systems, results are instantly processed by AI, and data is delivered directly to the customer.
On the front of the diagnostics, AI tools like ChatGPT have already demonstrated their capability to generate clear, accurate, and insightful reports. With fine-tuning, AI can match or even surpass seasoned diagnosticians and deliver results exponentially faster. The benefits are clear: reduced operational costs, faster turnaround times, and increased accuracy.
This isn’t speculation: it’s already being implemented, and adoption is accelerating. I suspect within 24 months; most global labs will rely on AI for diagnostics. Within five years, nearly all will approach full automation. Onsite oil analysis, once viewed as perpetually “thirty years away,” is now within reach.
Fifteen years ago, I worked on onboard oil analysis technology for fighter jets. More recently, I’ve been involved in developing portable analysis units for factories and mobile equipment. These systems are becoming both technically feasible and economically viable. Within a short time, they could replace up to 25% of lab-tested samples. Automation and AI in oil analysis aren’t futuristic, they’re inevitable. The industry is transforming, and those who embrace this shift will lead the next era of reliability and efficiency.
For what it’s worth, I still use my bike—both on the road and indoors—and yes, I still test my car’s engine oil. I guess I’m one of the 15%!