Foundations for a Data-Driven Approach to Predictive Maintenance 

Anybody serving in an operations and maintenance (O&M) capacity knows the saying, “schedule downtime or your equipment will schedule it for you.” Down equipment leads to loss of revenue, access to critical resources, and potentially catastrophic system failure.  

It’s no wonder downtime is the word most-feared word by maintenance managers. The implications go beyond a frenzied, and often expensive, scramble to repair or replace whatever broke. If an office building that can’t regulate the temperature for its employees, a freshwater system that can’t deliver water to residents, or a gas pipeline that can’t bring fuel to a local community, the potential loss of revenue pales in comparison to the reputational damage of a downed system. 

The answer to the plague of downtime – regularly scheduled and preventative maintenance. That is the practice of continuously inspecting and repairing systems to keep them running. The goals are to keep the systems optimum and to attempt to predict any future fallout.  

Historically, preventative maintenance has been thought of with a one-size fits all approach using periodic inspections and maintenance based on time in service. Savvier maintenance leaders adopted standard maintenance programs with some consideration of operating environment and active run-time. However, that is still a blanket approach. These practices soak up labor hours and there is potential for human error. Even the most experienced maintenance technician is bound to make mistakes and miss a looming critical event. 

Leveraging data is the next step in optimizing preventative maintenance and reducing downtime. The most progressive operators have adopted a data-driven approach to augment their maintenance schedule. By monitoring critical indicators, organizations can automate anomaly detection and determine issues before they arise. This way they can grease the wheel before it starts to squeak. This level of data-driven maintenance requires three things: 

  • 1. Historical data to establish baseline operations 
  • 2. Real-time access to data from the field  
  • 3. CMMS to compare real-time to baseline and raise a flag on issues 


Tosibox helps organizations build their data strategies, particularly in the water and wastewater or oil and gas sectors. In our conversations, we’ve noticed some patterns. Teams with early adopters have some amount of historical data. While it might not be the comprehensive historian they’d like, it’s better than starting from scratch. Almost universally, the other challenges they tell us about are unreliable network infrastructure resulting in frequent data loss, as well as legacy connectivity that limit network bandwidth to only a few high-level points. 

Without reliable data pathways that let teams consistently monitor all pertinent data at a site, there’s no foundation for a predictive, data-driven maintenance program. Luckily, the gap from legacy infrastructure ill-equipped to support your data strategy to robust connectivity isn’t as far as you might think. And Tosibox is here to help.  

Interested in learning more? Speak to one of our specialists to get started.