By 
A pair of MIT Sloan Executive Education graduates, translated teachings from an MIT course into operational improvements at Heineken Mexico.
It is well-known that manufacturers’ ability to increase and maintain production capacity is critical to long-term competitive success. You may be surprised that you can significantly increase your production without purchasing new equipment.
Heineken, a global beer manufacturer, is the second largest brewer globally. The company was founded in 1864 and had over 160 breweries across more than 70 countries. It also sells more than 8.5 million barrels of beer in the United States. The company is a well-respected brand because of its sustainable earnings and significant social and environmental responsibility. A pair of MIT Sloan Executive Education alums helped the company solve a considerable production bottleneck at its Mexico plant. This allowed it to unlock hidden potential through data-driven development and AI augmentation.
Little’s Law: Big Payoffs
Federico Crespo is the CEO of Valiot.io, a fast-growing industrial technology company. Miguel Aguilera is the supply chain digital transformation manager at Heineken Mexico. They first met at the MIT Sloan Executive Education program Implementing industry 4.0: Leading change in manufacturing and operations. The senior lecturer at MIT Sloan’s System Dynamics Group, John Carrier, taught the course. It gave Crespo and Aguilera the skills they needed to start a significant improvement process at Mexico’s largest brewery.
They would eventually use Valiot’s AI-powered technology to optimize their scheduling process in the face of unpredictable events. This would significantly increase the brewery’s throughput and improve worker experience. It all began with Little’s Law, allowing for a thorough diagnosis.
Little’s Law, also known as the First Law of Operations, is often referred to after John D.C. Little, a professor at MIT Sloan and an MIT Institute Professor Emeritus. The three most important properties of any system, throughput, lead-time, and work in progress, must follow Little’s simple relationship.
Little’s Law is an excellent tool for identifying and quantifying bottlenecks in any system. It is also one of the core frameworks of Carrier’s Implementing Industry 4.0 course.
Crespo and Aguilera used Little’s Law to go backward through the entire production cycle, looking at cycle times to determine wait times and identify bottlenecks.
They discovered a bottleneck in the filtration stage. Due to various interruptions and upsets in the facility and real-time demand-based production updates, beer was often delayed as it moved from maturation to filtration to bright beer tank (BBT) tanks.
This usually initiates a time-consuming, manual rescheduling process. Operators had to locate handwritten production logs to determine the current state of bottling lines. They also had to enter the data into spreadsheets on a local computer manually. A couple of hours were lost each time a line went down.
The facility quickly identified the problem and took immediate action to fix it.
Bottlenecks create habits that evolve into a culture.
The next step after identifying bottlenecks is to eliminate them. This can be difficult because persistent bottlenecks can change how people work in the system. They become part of the worker identity and the reward system.
Carrier says that culture can reject any technological advancement, regardless of its benefit to the system. “But culture can be a powerful tool for change and a problem-solving instrument.”
Carrier advises that the best way to introduce a new technology is to start projects early, which reduces human effort and invariably leads to overall productivity, reliability, safety, and quality improvements.
Heineken Mexico’s digital transformation
Federico and his Valiot.io team worked with Aguilera, the Monterrey brewery crew, to connect the enterprise resource plan with in-floor sensors and digitize the brewing process. The application communicated with Valiot’s data monitors for complete data quality. Using real-time data, machine learning was used to filter and optimize daily-optimized production times. BBT and filtration times were decreased in each cycle. The monthly brewing capacity increased by a significant amount. Within the first month of implementation, the return on investment was evident.
Heineken Mexico has used digital technology to allow real-time visualization of all bottling lines and filtering conditions for each batch. The technology optimizes efficiency at every production stage by continuously monitoring and learning from ongoing production. The real-time visualization tools allow factory workers to make real-time adjustments without slowing down or stopping production. The COVID-19 pandemic has also made it possible for operators to work remotely. This is a significant benefit.
These are the critical practical aspects.
The Valoit team had to be on the floor to decode the operators’ actions. The algorithm also had to be continuously tested for performance. Sergio Rodriguez Garza, from Heineken Mexico’s vice president of the supply chain, said that success was ultimately determined by Valiot’s approach affecting the profit and loss and not just counting the use cases.
Garza says, “the people who create the algorithms don’t always know where the value is in the facility.” Garza says that digitization is a complex process, and it is crucial to establish a bridge between those in charge and those in authority. The method of diagnosis still needs to be standardized. Each plant has its bottleneck and requires its diagnosis. The process of diagnosis is systematic. Each plant manager is responsible for diagnosing the backup.
Carrier says that a unique diagnosis is critical. A quality diagnosis is based upon a deep understanding of systems thinking.
Comments