How Machine Learning Will Save The Paper And Packaging Industry

Jennifer Scholze

The paper and packaging industry is in the midst of major change. Print circulation for magazines and newspapers has dropped dramatically in the last decade. Craigslist is replacing the classified section. Google and social media are replacing direct mail advertising. Cloud storage is replacing the office filing cabinet.

Despite these changes, however, it’s not all gloom and doom. Demand for paper-based packaging, including packaging made from recycled materials, is growing. Advancements in machine learning are enabling the industry to optimize production and minimize the input costs of assets, energy, and transportation. Machine learning can turn data into insights, unlocking new business opportunities and revenue streams.

Optimizing complex production and supply chain variables to address communication barriers

Paper and packaging companies are continually seeking new opportunities to streamline production. Last year, LNS Research surveyed more than 180 paper and packaging companies about their business goals, technology efforts, and future plans. Nearly half of all companies listed “better operational performance” as their top objective.

With demand stagnating, paper and packaging companies must hold the line on production costs. This can be challenging, however, due to the many variables involved in the production process. Alfred Becker, vice president, Mill Products Industry Business Unit at SAP, recently spoke about the importance of process optimization on the S.M.A.C. Talk Technology Podcast.

“If you imagine a simple piece of paper, it has many, many characteristics, like thickness or surface evenness or width or length, or it may be coated, and all of that has to be processed somehow, which is not a simple thing,” says Becker. “IT for a paper product is much more complex than the product itself may seem to be.”

In addition to variables in the physical product, there are also a number of different production variables, including assets, energy, and transportation. These variables can be costly compared with the value of the finished product. In the past, Becker says that technical restrictions made it difficult to optimize the production process and the supply chain.

The final part of this challenge is the disconnection between paper supplier and paper converter. Currently, many business models are based on two disconnected production processes: The paper supplier creates the raw paper goods. The paper converter, a separate business, then converts this paper into packaging.

“Paper suppliers and paper converters are disconnected,” says Becker. “They never can [improve] beyond a certain threshold because of their lack of information.”

Machine learning optimizes performance and improves data communication

One of the most fundamental applications of IoT sensor technology is maximizing machine uptime. Sensor data can make it easier for companies to engage in predictive maintenance and avoid unplanned downtime due to machine failure. Becker argues that machine learning can take this process optimization one step further.

“There’s a spin on this specifically to paper production,” says Becker on the S.M.A.C. Talk Technology Podcast. “A paper product is not as simple as one would expect, and [neither] is the process to produce it. The process is always fluctuating… At the moment, [companies] would like to run their known business without any disruptions [or] surprises, so they need really this stable core to sweat the assets, to produce just as planned to deliver in time. Paper companies currently invest significant time into analyzing their production processes and understanding what can be immediately improved. At the same time, [these companies] have to do something on top of what they’re doing today to become better and to develop new business models.”

To truly get ahead of the curve, companies must address the communication disconnect between paper supplier and paper converter. This starts with pooling data for advanced synthesis.

“It’s this synthesis that will lead into future, and that’s one of the concepts that we are preaching,” says Becker. “We tell [companies]: Don’t sell products, sell performance in the future.”

This can be really achieved through disruptive technologies like machine learning. Machine learning enables computers to “learn” without being explicitly programmed. Machine learning is capable of understanding a billion pieces of data in seconds. By instantly analyzing information from thousands of remote sensors across multiple companies, machine learning not only streamlines production, it also elevates performance.

Creating new business models and revenue streams

Machine learning is critical to optimizing performance. This optimization should not be limited to a company’s day-to-day production processes, however. To realize machine learning’s full transformative potential, companies must break down data silos. Pooling data for advanced synthesis across companies is key to creating new, performance-based business models.

Companies that view digital transformation as an opportunity rather than a threat can capture additional market share with these new business models. This will help offset stagnate growth in traditional markets.

For more information on how digital transformation can revolutionize the paper and packaging industry, listen to the S.M.A.C. Talk Technology Podcast.

Learn how to bring new technologies and services together to power digital transformation by downloading The IoT Imperative for Energy and Natural Resource Companies. Explore how to bring Industry 4.0 insights into your business today by reading Industry 4.0: What’s Next?


About Jennifer Scholze

Jennifer Scholze is the Global Lead for Industry Marketing for the Mill Products and Mining Industries at SAP. She has over 20 years of technology marketing, communications and venture capital experience and lives in the Boston area with her husband and two children.