The most sustainable factory may be the one that wastes the least

For years, manufacturers treated sustainability as something to be measured after the work was done. Operations produced, maintenance kept assets running, finance tracked cost and a separate team assembled emissions data for disclosure. That model could produce a report, but it rarely changed what happened on the plant floor.

The shift now taking place is more fundamental. Sustainability is being pulled into the same operational systems used to improve throughput, asset availability, energy intensity and margin. Mounir Boemond, Global Director, Sustainability Value at AVEVA, argues that this changes the economics of the conversation because environmental improvement is no longer presented as an additional cost layered onto production.

“When sustainability is treated purely as a reporting requirement, it is seen as a cost,” he says. “When it is treated as an operational challenge, the business starts looking for ways to improve efficiency, reduce energy use and become more profitable. The reporting and emissions reduction then follow from better operational performance.”

That distinction matters as manufacturers contend with volatile energy prices, tighter disclosure requirements and growing demands for supply chain transparency. Yet the strongest case for action may still be the simplest one. Wasted energy, unnecessary maintenance, poor asset performance and excess material use are environmental problems because they are operational problems first.

The value is in the operating detail

The most immediate opportunities are rarely hidden in a corporate sustainability dashboard. They sit inside the way equipment is scheduled, controlled and maintained, and they often become visible only when manufacturers can see where, when and why resources are being consumed.

At Kellogg’s Battle Creek facility in Michigan, the starting point was a detailed picture of electricity, natural gas and water use. Meters and standardized AVEVA PI Server tags created the baseline needed to analyze consumption across the plant. That visibility exposed systems that were effectively working against each other, including HVAC controls that heated air to protect coils from freezing and then cooled it again before discharge.

Retrofitting those controls produced savings worth several hundred thousand dollars a year. Engineers also found that six HVAC units were bringing in and heating fresh air separately. Reconfiguring the system to supply the units together and mix fresh air with return air saved $40,000 in heating costs in a single February week. Across the plant, the program has delivered annual savings of $3.3 million, alongside $1.8 million in rebates supported by the underlying data.

The significance is not that Kellogg’s built a more sophisticated sustainability report. It built the operational evidence required to find and validate waste, then used the same data to support energy targets, capital decisions and environmental disclosure. “Asset efficiency and energy consumption are where sustainability and profitability most clearly pull in the same direction,” Boemond adds. “Kellogg’s was able to identify when and how energy was being consumed at a granular level, and that led to substantial annual savings. The operational improvement created the sustainability outcome, while also providing the information needed to report it.”

Real-time visibility matters, but visibility alone is not enough. A plant can collect millions of readings and still struggle to decide what action to take. The problem is usually not a lack of data, but a lack of context connecting a reading to the asset, process, production order, environmental condition or business consequence that gives it meaning.

Context turns measurement into improvement

Henkel’s Laundry & Home Care business shows what happens when that context is created across multiple sites rather than within an isolated project. The company developed a digital backbone using AVEVA System Platform, AVEVA Historian and AVEVA Manufacturing Execution System Performance to bring together consumption, emissions and production information across its global operations.

By the end of 2020, more than 4,000 physical and virtual sensors were measuring electricity, fossil fuels, compressed air, steam, water and sewage, generating more than one million data points each day. The information was available to more than 2,000 users and could be analyzed from the factory level down to individual production areas and technologies. Production teams could discuss performance using shop-floor displays, while corporate teams could compare sites and identify where improvement practices should be replicated.

The architecture also connected line and machine status with SAP job data and correlated energy use with external temperature and humidity. Henkel could therefore understand not simply how much energy a site consumed, but how production mix, operating conditions and weather influenced it. The gains extended beyond utility reporting, with energy cost savings exceeding €37 million, annual savings reaching €8 million against the base year and overall equipment effectiveness improving by 15 percent after two years. The same foundation also reduced filling-line waste and virtually eliminated incorrect labeling.

Boemond believes this shared foundation is essential because the plant floor and the boardroom do not need identical dashboards, but they do need information derived from the same operational truth. “The corporate team and the plant team may be looking at different priorities, but the source of the information has to align,” he continues. “Real-time data needs to be contextualized and linked with other sources, whether that is production, finance, maintenance or even weather. Without integration and governance, manufacturers may have a lot of data, but they do not have the information needed to make the right decision.”

Many sustainability technology programs lose momentum by moving too quickly toward dashboards and disclosure before deciding which operational decisions the data should improve. The result is a polished view of consumption that remains disconnected from maintenance planning, operator workflows and production control. Sustainability becomes visible, but not actionable.

Moving from visibility to prediction

Once operational data is integrated and contextualized, manufacturers can move beyond identifying waste after it occurs. Predictive analytics and machine learning can detect deterioration, compare operating alternatives and intervene before an emerging problem creates downtime, excess fuel use, scrap or avoidable emissions.

Enel has applied that principle across a large power generation portfolio. AVEVA Predictive Analytics and AVEVA PI System monitor more than 1,285 assets using 4,310 predictive models and more than 47,000 PI System tags. Early detection helps resolve equipment problems before they escalate and require less efficient, more emissions-intensive replacement generation.

Enel estimates that thermal fleet interventions avoided 410,000 metric tons of CO2e over 24 months, while geothermal fleet interventions avoided a further 4,800 metric tons annually. Since 2020, its remote predictive diagnostic center has identified about 461 failures and avoided estimated losses approaching €47 million. The same maintenance decision can therefore protect availability, reduce cost and prevent emissions.

The relationship is equally clear in TC Energy’s compressor operations. Its Compression Optimization Tool combines AVEVA PI System data, digitized compressor performance models, fuel curves, machine learning and custom optimization engines. Every 15 minutes, it evaluates configurations across the company’s US footprint and flags more fuel-efficient alternatives for review by Gas Control supervisors.

That human review is important. Domain expertise is embedded in the models, while operators validate whether recommendations are practical under current system conditions. Between June 2023 and January 2025, TC Energy implemented 180 recommendations and eliminated 54,000 metric tons of CO2e while improving fuel efficiency across a network of more than 1,000 compressor units.

“AI and predictive analytics help manufacturers move from reacting to a problem to acting before it becomes a cost and an environmental issue,” Boemond explains. “They can accelerate the analysis and identify opportunities at a scale that would be difficult manually, but the human still needs to remain in the loop. The purpose is to give people better information and better options, not to remove operational judgment.”

These examples point toward a more useful way of measuring progress. The best indicators are not necessarily those created exclusively for a sustainability function, but the metrics that operations and leadership can both act upon. Energy consumption per unit of output, fuel efficiency, equipment health, process yield, downtime, water use and material loss connect environmental performance with cost and productivity.

That requires ownership beyond a single department. Leadership must set direction and fund common data foundations, while operators, engineers and maintenance teams define how insights will be used. A purely top-down mandate risks becoming another reporting exercise, while disconnected local projects will struggle to scale without standards, governance and executive support.

Manufacturers should also resist framing the transformation as one large sustainability program. A more practical approach is to begin with a visible operational constraint, establish a reliable baseline, connect the data needed to understand it and prove value through a measurable improvement. Once the model works, it can be extended across additional assets, utilities and sites.

For Boemond, the long-term direction is clear. Sustainability should eventually cease to appear as a separate operational category because the decisions that improve environmental performance will be embedded in the normal work of running the plant. “The evidence is already there,” he concludes. “Kellogg’s saved $3.3 million a year by understanding its energy use, Henkel combined €37 million in energy savings with a 15 percent OEE improvement, and TC Energy reduced emissions through better compressor decisions. These companies did not pay for sustainability as a separate burden. They created value by operating better, and sustainability was one of the outcomes.”

The question is no longer how much productivity a manufacturer must sacrifice to reduce its environmental impact. It is how much cost, waste and emissions remain because the operation is still not visible, connected or optimized enough. Closing that gap will do more for sustainability than producing another dashboard disconnected from the plant floor.