Milk Moovement: Digitizing and optimizing raw milk delivery

Transporting raw milk from farms to processing plants is a daily occurrence that seems simple at first glance. But it gets complex quickly when time is of the essence and milk volumes vary.

Every day in dairy is different, but some events are fairly routine. Cows on farms produce raw milk. Trucks travel to farms to pick up the raw milk. Those trucks then transport the raw milk to processing plants where it becomes the dairy products we buy at a grocery store. It seems simple enough, but it can get complex when transportation time is of the essence, and milk volumes and quality components (such as butterfat and protein) vary. Collectively, transporters are looking to get the right milk to the right place at the right time.

Milk Moovement, a Canadian dairy supply chain software company, is solving this problem by digitizing, integrating, and optimizing the production and movement of raw milk. For example, they know raw milk has to move from a farmer to a processor in a set amount of time in order to comply with industry regulations. They also know the quality components of milk may determine if it should go to a processor for yogurt, cheese, or fluid milk. Milk Moovement is looking to modernize these workflows by providing supply chain monitoring and management software to dairy producers, cooperatives, transporters, and processors.

"Our customers can log in from anywhere in the world and see what's happening in real time," said Robert Forsythe, Milk Moovement co-founder and CEO. "They can see where a truck is, how much milk is on the truck, and how quickly routes are traveled. Milk Moovement connects all players in the dairy supply chain which allows users to make decisions in seconds, not days.”

One key aspect of their value proposition is providing their clients with efficient milk transportation routes. Traditionally, weekly milk production volume estimates used historical data to manually plan transportation routes. While this ultimately achieved its goal, it was not optimal — trucks followed circuitous routes to pickup and drop-off milk, trucks arrived at processors with unused capacity, or milk volume inconsistencies caused delays.

For example, at the beginning of the week, a processor might require an additional 60,000 liters of raw milk. A truck is then dispatched. When the truck completes its route, they find out the collection of farms actually produced 65,000 liters. “Now you need to figure out how to get that additional 5,000 liters onto another truck and to another processor,” explained Brandon Drake-Goobie, CTO at Milk Moovement. “It’s also possible the truck completes its route and finds there are 30,000 liters. Now you’ve got a truck that’s only half full. This is one of the problems we’re trying to solve.” But Milk Moovement’s expertise is in the dairy industry, not transportation and routing algorithms or operations research. That’s where Nextmv comes in.

Example of milk transportation delivery and monitoring using Milk Moovement. Customers can plan routes and receive real-time route data such as driver location, the volume of milk on the truck, unused truck capacity, and total distance traveled. Image courtesy of Milk Moovement.

Nextmv’s decision stack gives the Milk Moovement team the building blocks they will need to quickly automate and optimize transportation routes that align to their business’ KPIs. For example, they will be able to maximize the volume of milk that transporters service while minimizing unused truck capacity, total driver time on the road, and distance traveled. The team can also use Nextmv to account for constraints unique to their business. This includes multiple pickup and drop-off locations, milk volume, milk quality, availability of trucks and drivers, and truck capacity.

To get started with Nextmv, the Milk Moovement development team started with a base fleet model. They then leveraged the core concepts to develop a custom model that satisfies the unique requirements of their business. With Nextmv, the team will have the flexibility to both plan out the week and respond to variances as they occur. It will be easy to ensure routes are still optimal by running their model multiple times per day on a per driver basis.

“The speeds are incredible,” said Drake-Goobie. “We can optimize routes for the morning in nanoseconds.” In production, Milk Movement runs the Nextmv decision stack in a dedicated EC2 instance that reads from several databases and feeds the necessary input data to Nextmv in order to generate optimal routes. And because Nextmv models are binaries that read and write JSON data, it's easy for the team to deploy to EC2 or other environments such as AWS Lambda as their architecture evolves.

Working with Nextmv is accelerating Milk Moovement’s ability to move to production with the digitization and optimization they needed for their business. And because they have the building blocks to customize and scale their decision stack without having to be logistics algorithm experts, the team can focus on expanding their business and product.

Video by:
No items found.