Stop Chasing Data for Data's Sake: What Manufacturers Really Need to Track
Key Highlights
- Labor shortages, inconsistent throughputs and missed delivery targets, not data itself, are the main factors that keep manufacturing leaders up at night.
- Recognize that different roles in manufacturing need different metrics. Targeted data delivered to the right people drives tangible results.
- Simple, frictionless data collection beats sophisticated analytics when adoption is the real barrier to success.
After years working on shop floors across manufacturing facilities of all sizes, Molly Garrison noticed that, despite widespread investment in data collection systems, manufacturers universally feel they're still not getting it right.
"From my days working on the plant floor at General Mills to working with really small two-line manufacturing facilities, everyone kind of feels behind the ball when it comes to data," said Garrison, now product manager at Formic, a robotics-as-a-service provider. "We're told data is a tool and that it's supposed to help us. So why does it feel like data is actually more of a burden?"
The answer, Garrison argued during her presentation at Pack Expo 2025, lies not in collecting more data or implementing more sophisticated analytics, but in fundamentally rethinking what manufacturers track and why.
The problem with traditional manufacturing metrics Most manufacturers track a familiar set of metrics: units produced per shift, labor hours for the week, downtime incidents and overall equipment effectiveness (OEE). These metrics have become expected staples of manufacturing data collection.
The manufacturing data challenge isn't about collecting more information or implementing more sophisticated analytics platforms. It's about stepping back and asking what problems you're actually trying to solve, determining who needs what information to solve them and how to deliver that information to the right people with minimal friction.
"However, they don't really get us to a point of impact and action," Garrison said. "Frequently, when we get a nugget of information from this list, the follow-up is that we need to investigate what the data is trying to tell us."
Manufacturing facilities are incredibly busy environments with constant fires to fight. So, when data feels like it's working against manufacturers rather than for them — by creating a need for more investigative action rather than remediation — the focus on data can fall down the priority list at a time when it should be rising to the top.
Start with what keeps you up at night
When Garrison set out to develop a data tool for Formic's customers to help target specific actions that can be taken on the plant floor to improve operations, she deliberately avoided asking manufacturers about the kinds of data they wanted to see. Instead, she asked engineers, operators and plant managers at more than 100 facilities a different question: "When you close your laptop at the end of the day and head out to your car, what is still nagging at your brain and keeping you up at night?"
Despite the wide variety of people and facilities interviewed by Garrison, four themes emerged consistently in 85-90% of her conversations:
Labor shortages. Whether hiring, retaining or simply having enough bodies on the line, labor challenges pervade manufacturing.
Inconsistent throughputs. Despite regularly conducting time studies, following OEM specifications and implementing careful scheduling, "one little thing can happen and all hell breaks loose on the plant floor and then the plan goes out the window,” Garrison noted. “This feels like an everyday occurrence for many manufacturers."
Missed production targets. When you commit to producing 10,000 cases but only produce 7,000, the gap must be filled somehow, typically through schedule disruption or expensive overtime.
On-time, in-full delivery. Customer orders drive everything. Missing deliveries can mean fines, exorbitant logistics costs and damaged customer relationships.
Garrison pointed out that, in this list generated from her conversation with manufacturers, "you will notice data wasn’t mentioned a single time. That’s because data is not the thing keeping people up at night. These four things are the highest impact factors manufacturers should be focusing on."
Developing persona-based metrics
Once Garrison identified the real business challenges manufacturers struggle with, her next insight emerged: Different roles need different data to address their specific challenges. Operators care about hitting targets and staying on schedule. They need to see trends by hour and SKU that can prompt calls to maintenance or material quality checks.
Manufacturing facilities do not thrive if something is highly disruptive to their way of working.
Supervisors focus on line consistency and labor utilization. Hourly performance data, daily trends and line benchmarking help them make decisions about labor allocation, training and troubleshooting.
Plant managers worry about staffing, output versus plan and overall plant performance. Line and SKU performance metrics, downtime by highest-volume product and plant benchmarking guide them toward continuous improvement projects and better resource allocation.
Executives concentrate on order fulfillment, labor costs and turnover. Cost per case and labor per order metrics drive their decisions on capital projects, resource planning and business growth.
Despite these differences in duties and areas of focus, many manufacturers aggregate all their data across these roles and dump it into Excel spreadsheets that no one really understands, looks at or uses effectively.
Garrison shared a telling example from her customer interviews. One facility collects five sheets of paper daily as production recaps. Every Friday, a supervisor manually enters that week's 25 sheets into an Excel spreadsheet. Plant leadership then transfers metrics from that spreadsheet into another one for a weekly meeting with company leadership. With this process, the meeting takes place a full week after the data was collected.
"When I noted that this (aggregated data) must provide a really important metric and asked them how they used it, their response was, ‘Honestly, we’re not really sure; it's just the best we've got," Garrison recalled.
A week of manual data entry, multiple transfers between spreadsheets and leadership meetings all for a metric whose purpose no one could clearly articulate.
Simplicity over sophistication
The temptation to implement cutting-edge software in industry is strong. AI, predictive analytics and advanced modeling represent exciting possibilities in the data space. But for small to mid-sized companies with little to no data infrastructure, "does it feel realistic to go from zero to AI and predictive maintenance?" Garrison asked.
The answer lies not in collecting more data or implementing more sophisticated analytics, but in fundamentally rethinking what manufacturers track and why.
Many companies have process engineers excited about new data tools, but their initiatives consistently fail due to poor adoption. Technology-resistant workforces, especially in manufacturing, won't embrace overly complex systems regardless of their theoretical benefits.
"So instead of going from zero to AI, start with baby steps and keep it simple," Garrison advised. "This keeps it easy to access, easy to understand and easy to act on."
To implement this more straightforward approach, Garrison recommends manufacturers evaluate their current data practices with three simple questions:
- Is it actionable? If the action is "we need to investigate it," that doesn't count. You should have a clear idea where that information leads.
- Is it owned by someone? Not a team — an individual. Clear ownership ensures accountability for acting on the information.
- Does it tie to a business outcome? If you're evaluating information that doesn't connect to your top three or four priorities, it's probably not the best use of time.
"If the answer is no to any of these, that piece of data probably isn't serving you as well as it could," Garrison said.
Make data frictionless
Beyond tracking the right metrics for the right people, the data collection process itself must be seamless. "Manufacturing facilities do not thrive if something is highly disruptive to their way of working," Garrison explained.
This principle shaped the creation of Formic Production Intelligence (FPI), the tool Garrison developed based on her customer insights. For operators, FPI displays real-time production progress, expected rates for current SKUs and visual cues (in red, yellow and green) indicating performance status. It’s also bilingual in recognition that manufacturing's heavily bilingual workforce shouldn't have to struggle with English-only interfaces.
"I trained all three shifts [at one company] on it within 10 minutes, and they have been using it consistently every day since," Garrison reported.
Adam Bragg from Cameron's Coffee, a Formic customer, reinforced this frictionless data approach: "Don't just automate your machinery, automate your process flow and your information flow. Make operators' jobs easier by not making them fill out a piece of paper. FPI is simple and an easy place to start."
Different roles need different data to address their specific challenges.
Simple, accessible data systems also empower continuous improvement initiatives. "I have been a continuous improvement specialist,” Garrison said. “I've done lots of improvement projects in previous roles, and one thing I often saw is that we would implement a project and then there would be no post-mortem on it. With proper data tracking aligned to business outcomes, manufacturers can see the before and after of their initiatives in real time, understanding whether changes actually delivered value.
Practical, not revolutionary
Garrison stressed that the manufacturing data challenge isn't about collecting more information or implementing more sophisticated analytics platforms. It's about stepping back and asking what problems you're actually trying to solve, determining who needs what information to solve them and how to deliver that information to the right people with minimal friction.
"Nothing I'm saying here is revolutionary, but we still struggle to do it, because the temptation to do the fancy, cool, next big thing is there," Garrison concluded. "But that does not serve our teams."
Amid all the excitement around artificial intelligence or predictive analytics, for many manufacturers meaningful results can be achieved by simply tracking the right metrics, presenting them to the right people and making data collection so seamless that it becomes the tool it was always promised to be.
