Every speck of raw material that fails to become finished pet food carries a cost. Many manufacturers focus on optimizing formulation or increasing throughput. However, Néstor de Miguel of Nutriciel pointed out that some of the largest opportunities for improving profitability come from systematically identifying and reducing losses throughout the production process. De Miguel spoke at Foro Mascotas in Guadalajara, Mexico on July 15. Nutriciel, developed by Actemium Nantes Nutrition, is a production management and automation platform for the animal nutrition industry that integrates process control and manufacturing management.
De Miguel’s presentation examined loss points from ingredient receiving through finished product packaging. Just as the Sahara Desert is made of many tiny grains of sand, he emphasized that many small losses accumulate into significant problems. Automation, data collection and artificial intelligence can help manufacturers identify deviations before they become expensive problems.
The first opportunity to reduce losses occurs before ingredients even enter production. De Miguel estimated the financial impact of impurities in common ingredients. Using example values, corn containing an average of 3% impurities represented an estimated loss of MXN210 per metric ton received, while soybean meal at 2% impurities represented MXN120 per metric ton.
Every speck of raw material that fails to become finished pet food carries a cost. Many manufacturers focus on optimizing formulation or increasing throughput. However, Néstor de Miguel of Nutriciel pointed out that some of the largest opportunities for improving profitability come from systematically identifying and reducing losses throughout the production process. De Miguel spoke at Foro Mascotas in Guadalajara, Mexico on July 15. Nutriciel, developed by Actemium Nantes Nutrition, is a production management and automation platform for the animal nutrition industry that integrates process control and manufacturing management.
De Miguel’s presentation examined loss points from ingredient receiving through finished product packaging. Just as the Sahara Desert is made of many tiny grains of sand, he emphasized that many small losses accumulate into significant problems. Automation, data collection and artificial intelligence can help manufacturers identify deviations before they become expensive problems.
1. Start with incoming ingredients
The first opportunity to reduce losses occurs before ingredients even enter production. De Miguel estimated the financial impact of impurities in common ingredients. Using example values, corn containing an average of 3% impurities represented an estimated loss of MXN210 per metric ton received, while soybean meal at 2% impurities represented MXN120 per metric ton.
"What is not controlled when receiving raw materials becomes a permanent loss,” he said.
Practical recommendations included:
- Review acceptance specifications negotiated with suppliers.
- Verify moisture, protein, impurity levels and particle size during receiving.
- Ensure truck scales are properly calibrated.
- Document incoming moisture levels for each delivery.
2. Moisture loss can quietly erode inventory
Storage losses often receive less attention because they occur gradually rather than during a visible production event.
De Miguel explained that grains and cereals continue losing moisture during storage. Under storage temperatures above 33 degrees C, grains may lose as much as 3% moisture in less than three months.
The presentation identified several ingredients as particularly susceptible, including corn, rice, wheat and soybeans
Using an example of 1,000 kilograms of grain, a 3% moisture reduction lowered inventory weight to approximately 970 kilograms. Although the nutritional value of the material may not change proportionally, the saleable weight decreases.
His recommendations included:
- Monitor silo ventilation.
- Record moisture at receiving.
- Continue monitoring moisture throughout storage.
- Evaluate storage time alongside temperature.
3. Small dosing errors become large financial losses
Another major focus involved ingredient dosing.
De Miguel divided dosing losses into two categories:
- Operator-related errors when weighing ingredients using scales with inadequate precision.
- Process variability caused by automated systems, particularly "freefall" after dosing gates close.
The presentation illustrated how material continues falling after a valve closes, meaning the displayed weight may continue increasing before stabilizing. If the next ingredient begins dosing before stabilization, actual ingredient inclusion can differ from the target.
The presentation stressed an important distinction between three scale characteristics:
- Resolution
- Display (visualization)
- Accuracy
Using a 25.384-kilogram example bag, De Miguel explained that a scale with 10-gram resolution may display 25.38 kilograms, but resolution alone does not guarantee measurement accuracy. Calibration and instrument quality ultimately determine measurement error.
Recommended actions included:
- Verify scale resolution.
- Verify scale accuracy.
- Check load cells regularly.
- Document performance for every dosing station.
- Compare measured performance against expected process losses.
4. Off-specification product deserves another look
Rather than treating all rejected product as waste, De Miguel presented rework as another opportunity for recovery.
His example identified three common sources of off-specification material:
- Start-up and shutdown product, averaging 200 to 400 kilograms per production run.
- Product outside density or dimensional specifications, representing roughly 1% of production.
- Fines removed during polishing, typically 1% to 3% of a production run.
Instead of discarding these materials, the proposed rework system converts rejected product into an aqueous slurry that can be reintroduced through the extruder's water line.
The process includes:
- Collect rejected product from the extruder.
- Mix it with water in a rework tank.
- Create an aqueous solution.
- Feed that solution back into the extrusion process through the process water system.
According to the presentation, this approach allows manufacturers to recover material that would otherwise become waste while reducing disposal costs.
5. Cost visibility creates accountability
Beyond identifying loss points, De Miguel emphasized building a cost structure that operators and managers can use daily.
Rather than tracking dozens of accounting categories individually, he recommended consolidating costs into a smaller number of operational indicators, including:
- Labor
- Maintenance
- Electricity
- Depreciation
- Laboratory analysis
- Natural gas and lubricants
- Water
- Freight
- Administrative costs
- Packaging materials
The presentation recommended establishing acceptable variation limits.
World-class manufacturers, according to the presentation, often target normal variation within ±5%. Deviations beyond that threshold should trigger investigation, while larger deviations require immediate corrective action.
Comparing current performance against both budget and the same month in the previous year helps identify whether rising costs result from changes in operations or broader market conditions.
6. Automation and artificial intelligence extend process control
The presentation concluded by looking beyond traditional manufacturing controls.
According to De Miguel, pet food plants have increasingly standardized critical processes such as ingredient dosing, extrusion and traceability through automation. Once these systems are integrated with enterprise resource planning and supervisory control systems, manufacturers gain better visibility across production.
He also described emerging uses of artificial intelligence, including:
- Monitoring critical process variables in real time.
- Detecting process deviations earlier.
- Predicting equipment jams before they occur.
- Automatically adjusting operating parameters.
- Supporting operational decision-making using historical data.
However, De Miguel cautioned that artificial intelligence should complement, not replace, operator expertise.
“Clients tell me, ‘ChatGPT said this or that,’ and I say, ‘but does ChatGPT know your formulations and machines as well as you do?’” he said. “AI only knows what it has learned from us. When I make an observation or suggestion, it is from real-word experience.”
He recommended continued training of real-live humans by flesh-and-blood equipment manufacturers or process specialists.
De Migueal framed loss reduction as a plant-wide discipline rather than a series of isolated projects. From tighter receiving controls and moisture monitoring to improved dosing accuracy, structured cost tracking, rework systems and predictive automation, each improvement may appear incremental on its own. Combined across the manufacturing process, however, those incremental gains can translate into substantial reductions in raw material losses, lower production costs and improved profitability.
