How can innovation be brought when big data encounters the beer industry?


In December 2018, Deschute Brewing Company, the tenth l […]

In December 2018, Deschute Brewing Company, the tenth largest brewery in the United States, announced plans to lay off 10% of its workforce. The company blamed it on the decline in sales and sales, which is the norm after the craft beer market has reached saturation.

For most brewers, such large-scale layoffs also mean sales and production cuts. But Deschute did not have this plan, because just four years ago, the company decided to integrate sensors connected to the Internet in the production process.

Of course, beer brewing is still a manpower-centered industry. Traditionally, workers manually sample and analyze beer during the production process to determine whether the beer should be transferred from one brewing stage to the next. This process is called stage transfer, and there are 9 stages in this stage. One. Premature or late transfer will affect the quality of the finished beer.

Like many winemakers, Deschute is headquartered in Bent, Oregon, and it maintains these samples and analysis records. Subsequently, the company decided to use the effectiveness of these records, and use Microsoft and OSISoft to parse the data in the cloud to predict the number of transfers in the production process. The resulting conclusions help streamline the brewing process and bring more results to the company.

Deschute's winemaker Brian Favre said: "When there are production problems or layoffs, the company is not willing to invest in other assets. We are no longer recruiting new people for around-the-clock operations. Usually, we have to do Sacrifice, and in the past it was at the expense of quality, productivity or employee happiness... What we can say now is that we are full of confidence in this model."

The predictive analytics framework has been implanted into all of the company's approximately 50 brewing tanks, ranging in capacity from 100 barrels to 1,000 barrels (3,150 gallons to 31,500 gallons). Currently, the company will move through the manpower phase after the winemaker confirms that a certain brewing stage is completed, but Fav said that the company is exploring the automation of this process.Craft Beer

What is the net effect of the brewing process data analysis? Deschute reduced the fermentation process for each batch of beer to 48 hours, a reduction of 24 hours from the previous one. This also allows the company to increase its annual production without purchasing additional equipment.

So far, Deschute is the only company that uses sensors and data analytics to help brew, but Favre said that some winemakers like Sierra Nevada have come to understand this model.

The project also spawned an open source data collection project for craft brewers who shared their history of the length of each phase of beer brewing.

“Most companies may have sensors, but they record the data on paper or in spreadsheets. Our move can help them collect data and build a database and then provide a space for them,” said Favli. They build databases of the above data so they can improve their production in the future."

He continued: "In the crafts industry, it is difficult to make such adjustments. Nowadays, people are gradually adapting to this approach. They see it as a tool, not a robbing of their jobs."

As workloads decrease, Deschute is now looking for new ways to leverage its data analysis tools. The company is still thinking about more industry-specific applications while exploring its application to day-to-day operations, such as predictive analytics for early warning when equipment is about to break.

One of them is to use a spectrometer to measure beer flavor.

“We put all the recipes in the database,” Favre said. “Now we have found matching data for these formulas, so we will conduct a test analysis to measure the various compounds in the beer. This is me. One step to do is to use this data to try to find out if you can find a proportion of the compound that gives the beer a certain quality. It is because of these characteristics that consumers will be interested in our beer, our beer can Different."

This analysis does not mean eliminating the human factor in beer brewing, it is more like the acceleration of the brewing process. Sometimes, when a brewing company finds a successful formula, they have repeatedly brewed more than 100 times or more in order to find the specific flavor they are pursuing. However, with technology, the number of tests can be reduced to 10 times.

It is undeniable that for some brewing companies, this will make the winemaking work lose some fun. One-third of the best brewers are credited to their crazy scientific experiments, which try to mix hops and malt with things that are beyond their imagination, and often give rise to a delicious new flavor.

To ensure that this tradition continues, Deschute set up a testing plant that brews only one barrel of beer each time and continually tests it, then receives feedback from its tasting room to understand the reaction of the beer fans to the new flavor. (These small batches of beer data are recorded in the same way as mass-produced beer, and maybe a test will be a big success.)

“The craft beer requires a lot of effort and energy. Therefore, separating this content from beer is a sensitive topic. In this respect, companies must create trust and engage in dialogue.” One move is a tool. We don't intend to give beer brewing to machines and engineers. It's just a means to improve the efficiency of brewing."

Contact Us