Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

Factories generate data, but why can’t they use it?

The future of manufacturing requires a new approach.

[Photo: Getty Images]

Profile photo of Hugo Dozois-Caouette

CTO and Co-founder · MaintainX

If manufacturing is the world’s most data-rich sector, why do so many factories still lean on Post-it notes, color-coded paper, binders, and employee intuition?  

Manufacturers generate more data than any other sector, nearly double that of the next-most data-rich industry, banking and financial services. However, IBM research indicates that only 28% of industrial organizations derive insights from data generated by their equipment, processes, and systems. Fewer than 20% have real-time access to important enterprise-wide manufacturing data.  

Instead, frontline workers are leaning on handwritten notes stuck to half-million-dollar pieces of machinery, printing priority repairs on red paper to flag critical items, and risking the loss of institutional knowledge with each employee who quits, is let go, or retires. According to the World Economic Forum, that’s leaving $100 billion of potential value creation on the table.  

Though other industries would salivate at the opportunity, little has been done to seize it. Up to this point, the primary approach has displayed data around the workplace. But frontline manufacturing teams still often operate as if they’re in a low-data environment, jokingly referring to those futuristic displays as “expensive wallpaper.”  

The strange dichotomy of an industry that is simultaneously data-rich and data-underutilized—even within the same four walls—is what I label the “manufacturing data paradox.” Leaders need to understand how we got here and how to fix it.  

Why data doesn’t deliver  

This huge volume of data rarely arrives when and where it’s needed most, so we first need to understand how it’s generated.  

In a typical manufacturing setting, machines and sensors are the biggest data producers. These constantly pick up signals from the floor, along with stacks of reference documents such as manuals, SOPs, maintenance records, and compliance reports.  

Despite the huge volumes of data being produced, most aren’t delivered to the right person at the right time to make a difference. Silos keep data trapped in control systems, ERPs, and clipboards.  

But that’s just the beginning. Industrial data often takes on its own language, or at least its own dialect, making it harder for those who don’t work with it every day to understand and interpret it. For example, enterprise systems are often built for managers and analysts, not the frontline operators that ultimately interact with them.  

Frontline staff are also constantly dealing with information overload. So much is constantly coming at them that it’s hard to know what’s worthy of attention. Those fancy dashboards may even contribute to the problem.  

A culture of “The way it’s always been done” leaves the world’s most data-rich sector struggling to use its informational assets.  

When data assets can walk  

Machines and manuals often contain the highest volumes of data, but the most important information lives in the hardest-to-reach place.  

In this industry, frontline workers possess the most relevant data and understand the operation’s nuances, intuition, and day-to-day patterns. Though that knowledge is of the highest value, it’s the least likely to be captured and easiest to lose as record numbers retire from the workforce, each carrying decades of experience and know-how out the door.  

AI can’t fill this void. Most manufacturers fail to appreciate this institutional knowledge, thereby missing the opportunity to capture it before it’s too late. We even saw one client rehire retirees to support their knowledge-capture strategy.   

Despite sensors, systems, and algorithms applied to the factory floor, the next generation of industrial intelligence can only realize its full potential if it preserves the insights of the previous generation of frontline staff.  

Build the future’s cyborg workforce  

In an ideal world, we will pair human expertise and machine intelligence to leverage their unique strengths and unlock new value.  

We will capture data in ways that don’t feel like data entry—such as voice-to-text, smartphone-captured photos or videos, or intuitive mobile apps connected to manufacturing machinery—and build insights in the background of day-to-day workflows.  

This ideal world can deliver insights in ways that feel less like digging through instruction manuals and more like listening to a tenured colleague who senses when a machine is on the verge of acting up again, or a personal assistant who manages schedules and drafts reports.  

This is what AI promises, and the technology is already delivering in certain corners of the industrial sector. Most manufacturers already have technology-enabling data, but need the will to put it to use.  

Take the next step 

Moving from Post-it notes and color-coded paper to automation and AI-driven efficiency is more a matter of mindset than technology. The tools already exist, but organizations lack an adoption strategy that feels attainable and delivers results.  

To bridge that gap, I recommend that leaders reframe their perspectives. Change your strategy from collecting data to connecting it: 

  • Measure data’s impact in tangible terms, such as downtime avoided, worker safety, and hours saved.  
  • Treat your frontline staff like customers and sell your vision to them accordingly.  
  • Make tools and platforms as intuitive as the consumer products they use every day. 
  • Create open forums and empower technicians to raise questions and concerns about AI or suggest ways it can make a real difference in their day-to-day.  
  • Approach your tenured employees like the walking knowledge repositories they are.  

Automation should bring simplicity, safety, and reliability to industrial settings. Despite collecting the necessary data, many organizations haven’t seen significant change on the floor.  

Moving from the world’s most data-rich sector to the most technologically advanced sector requires utilizing insights. The best way to do so is by building connections between systems, shifts, and human insights and machine intelligence.  

Hugo Dozois-Caouette is the CTO and cofounder of MaintainX.