Cheese producers struggle to keep the quality of their cheeses consistently high. Dutch company Notilyze has developed a data analytics-based solution to increase production efficiency and prevent ingredient waste.

Cheese is cultural heritage in the Netherlands. Our northern neighbors love their cheese so much that there is even a Dutch Cheese Tasting Championship, as we learn from Colin Nugteren. He is co-founder of Notilyze, a Rotterdam company specializing in analytics. Providing advice, support and data to a Dutch cheese manufacturer was a major challenge for Nugteren and his company. The project earned Notilyze a prestigious recognition at the recent SAS Hackathon.

Exact science

Cheese is not a simple product. First and foremost, there are different types of cheese, each with their own preparation method. You don’t have to be an expert to know that a solid block of Gouda is different from a soft cheese. Specifically, there are four factors that manufacturers must perfectly balance to arrive at the type of cheese of their choice: moisture, acid, fat and salt.

“The difficult thing about making cheese is that you are working with living organisms that convert milk into cheese. In addition, there are large differences in milk quality depending on whether the animals live primarily indoors or outdoors. But the conditions in the factory also play a role: the temperature, how long the machines run… As soon as something changes, it also affects the consistency. “In addition, manufacturers have to produce cheese on an assembly line,” says Nugteren. Due to the mix of all these factors, the Yields (a value that expresses the return on your production) is usually low in the cheese industry, resulting in the waste of many liters of milk.

Three problems, three solutions

There is no problem that data cannot solve, thought Nugteren. “We met with the ‘cheese technologists’ and three problems emerged. The first problem was that data analysis took too much time for producers. The second and third problems are the large fluctuations in raw materials and the fact that adjustments in the production process are very resource-intensive.”

“Of course everything starts with the data: if you don’t have it, the story stops,” Nugteren makes an important point. “You can use data to find out things like which types of cheese are made frequently and which are made less. Thanks to our partner we had this available.”

The next step was to make this data tangible. Nugteren explains how his team achieved this: “Through SAS Viya, we set up dashboards to show correlations between variables. Analyzes can be carried out much faster this way because you don’t have to collect data or work with Excel files. We then built a predictive model to predict what happens when you adjust a variable. This allows the producer to make adjustments more proactively without always having to wait for the result to be published. The third level is to convert these predictions and analyzes into a model that can make decisions for the manufacturer that still need to be tested against reality.”

Increase yields

A well-known Flemish TV chef would now ask what we have learned today. After the project, Nugteren is optimistic that the data can lead to higher yields and therefore less waste of milk and other ingredients. “At this stage it is still too early to really talk about returns, but early opinions are certainly enthusiastic. We’ve created a good model for common cheeses, but it could be improved endlessly. If I have to quantify it, an increase in returns of at least seven percent must be feasible.”

Everything starts with data: if you don’t have it, the story stops.

Colin Nugteren, Notilyze

SAS hackathon

Notilyze was named winner in the Manufacturing category at SAS’s annual global hackathon this year. This is the third year in a row that Notilyze has won awards. A great recognition for Nugteren and his colleagues: “This offers us the opportunity to further expand the data process and gain new insights.” We share many customers with SAS, so this “seal of quality” is a nice bonus.”

This editorial article was created in collaboration with SAS.