Want to reduce food waste? Start with your data

Nov 13, 2026
  • food

A renowned Belgian ice cream producer had an overfilling problem. Product was leaving the line heavier than it needed to be – quietly, consistently, expensively. The fix started with a dashboard. Once the team could see the distribution of fill weights in real time, the root cause became obvious, so they could act. The result: 15 to 20% savings on raw materials. Jeroen Coussement - Founder & CEO of Factry and Rens Bonnez Lead Architect & Tech Strategist at delaware explain the architecture behind results like these – and why most food manufacturers are still missing it.

Data everywhere, insight nowhere

Temperature sensors on the line, quality checks logged in a spreadsheet, batch records in the ERP, energy consumption tracked by the sustainability team: in food production, data abounds. The problem is that that data sits in silos, each one speaking a different language, none of them talking to each other. 

“Ten years ago, the IT and OT worlds were completely separate,” says Jeroen Coussement, founder and CEO of Factry – a platform that helps manufacturers gain control over their operations by collecting, analyzing, and visualizing production data. “When we started, getting to the right data in a factory was technically difficult and operationally even harder. That gap is still the root cause of most food waste problems we encounter today.”

Where the gap shows up

The consequences show up on the production – and the bottom – line. When data from a filling line isn’t connected to batch records, overfilling goes undetected. When temperature deviations in a Clean-In-Place cycle aren’t visible in real time, product quality issues only surface in a lab result two days later. When a packaging line runs without a live feed to the quality system, mislabels compound before anyone notices. 

“Food waste rarely has a single cause,” Jeroen continues. ”It emerges from dozens of small touchpoints across the production process: a slightly miscalibrated filler here, a temperature fluctuation there, a batch that ran five minutes too long. Tackling it starts with visibility, not algorithms.”

You can't manage what you can't measure

“The unique strength of Factry is the ability to generate insights from real-time floor data,” Rens Bonnez says, “and make them available to the people who need them, in a format they can actually act on.” 

“Step one is always measuring and visualizing,” Jeroen explains. “Not to have a beautiful dashboard, but to support a hypothesis. Once the problem has a number, you can act on it.” At ice cream manufacturer Jacques Ice, that meant instrumenting the filling lines and making overfill visible in real time. The intervention started small, with a dashboard on the production floor, yet led to double-digit savings on raw materials. 

Step one is always measuring and visualizing - not to have a beautiful dashboard, but to support a hypothesis. Once the problem has a number, you can act on it.
Jeroen Coussement, Factry

From noise to signal

Behind every useful dashboard lies an architecture decision that most food companies underestimate. A modern production environment generates data at a rate that ERP systems were never designed to handle. Think sensor readings every 100 milliseconds – thousands of data points per machine per hour. Feeding that raw stream directly into a business intelligence tool produces noise. 

 

“On the shop floor you need all the details,” Jeroen explains. “But at ERP level you need aggregates. Factry sits in between: we concentrate raw sensor data into information with context: which batch, which recipe, which conditions, which operator? That contextualization makes the data useful for analysis, and it makes AI possible.” 

 

Large language models and predictive algorithms need clean, structured, contextualized data to produce reliable outputs. “Without proper contextualization, there is no AI in manufacturing,” Jeroen says. “Factry does the preparation work. AI tools assume good data. We’re the ones who make sure that assumption holds.” 

The thousand solutions trap

For food manufacturers trying to build this architecture, the market offers no shortage of options. That abundance is itself a problem. "Companies make choices without a thorough strategy," Rens explains. "They buy solutions for specific problems and end up with ten systems that don't talk to each other. The question of how to collect data from older machines alone is enough to derail a project before it starts." 


The answer is not to avoid technology choices, but to make them in the right order. Strategy before platform. Architecture before implementation. You need a clear picture of what data is needed, where it lives, and how it needs to flow before a single line of code is written, or a single sensor is installed. 

 

That’s where delaware comes in. “delaware is IT-focused by origin, but we now do much more on the operations side, because Industry 4.0 depends on that operational data being available and usable," Rens says. "We help clients build a program at board level to eliminate waste, connecting the strategic ambition to the technical architecture that makes it real." 

Companies make choices without a thorough strategy. They buy solutions for specific problems and end up with ten systems that don't talk to each other.
Rens Bonnez, delaware

One dataset, many audiences

A well-designed architecture doesn’t just feed management dashboards. Its most important users are the people closest to the production process.  

 

That understanding shapes how Factry builds its interfaces. The operator on the line needs one clear screen with the alerts that matter at that moment. The supervisor needs trend data and root cause analysis. The process engineer needs the full historical record. Each layer draws from the same contextualised data, presented at the right level of detail for the person reading it. 

 

“An operator shouldn’t have to look at twenty different screens,” Rens adds. “One unified interface with the right alerts is what drives behavior change on the floor. The deeper data is there when you need it, but it shouldn’t get in the way.”

Read more on why people make the difference in reducing food waste 

We see too many companies spend months on assessments and slides, only to land on the wrong tools. We advise starting small. Pick one line, one use case, and just begin."
Rens Bonnez, delaware

Think big, start small

For food companies, the temptation is to wait until the architecture is complete before doing anything. Or they start too big. That is the wrong instinct. 

  • "We see too many spend months on assessments and slides, only to land on the wrong tools, without ever involving the people who actually work on the floor," says Rens. His advice: “Think big, start small, with one line or one use case.” 
  • “It often begins very simply,” Jeroen confirms. “Put up a dashboard in production to make people aware of losses. That moment of visibility, when a team sees for the first time how much product is leaving the line as waste, is often the spark that changes behavior.”  
  • delaware’s approach mirrors this. “We help clients get started, hold their hand through the early stages, and build towards a point where they can take over themselves,” Rens explains. “The goal is always independence. A client who understands their data, and owns their architecture, will be able to keep improving without us needing to be in the room.”  

That continuous improvement is, of course, a must. "Food waste reduction is not a one-time project," Jeroen says. "It's a commitment. As production lines evolve and new use cases emerge, the data foundation will need to grow with them.” 

The companies that invest in that foundation today are the ones that will be ready to act when the next opportunity, or the next source of waste, reveals itself.

Fight Food Waste Week

Every day, raw materials, water, energy, time, and profit disappear from your production process. You know it's happening, but you don't know exactly where, why, or what to do first. 

delaware's Fight Food Waste Week helps you make the invisible visible.

Rens Bonnez

Lead Architect & Tech Strategist at delaware

Rens Bonnez specializes in designing data and technology architectures for industrial environments. He helps organizations connect production and business systems, turn raw operational data into meaningful insights, and create the digital foundations required for continuous improvement, smarter decision-making, and future AI-driven innovation.

Connect with Rens on LinkedIn

Related content