Bringing DX to the Shrinking Publishing Business

Since the spread of the internet, the influence of printed media has gradually declined. In recent years, there has been a succession of magazine closures and bookstore closures, and it is a fact that the publishing market is shrinking. Along with the shrinking market, the long-standing problems of the publishing industry are becoming more apparent.

In the process of resolving existing issues and creating a sustainable market, the “return rate” has been a major problem. The economic loss (waste loss, logistics costs, etc.) associated with returns is said to be over 200 billion yen annually, and these costs cannot be ignored.

However, these issues are structural and difficult for a single publisher to solve alone. That’s why three major publishing houses, Shogakukan, Kodansha, and Shueisha, along with Marubeni and Marubeni Forest Links, which has been involved in distribution with the three major publishing houses for many years, stepped up. “PubteX,” established in 2022, is trying to change the way publishing is done by bringing DX to the publishing industry through the use of AI and IoT.

There are two main issues that PubteX is approaching. The first is to eliminate the high return rate. While publishing and distribution have been carried out based on the know-how that each company has cultivated, PubteX aims to achieve demand forecasting and supply chain efficiency by utilizing data from the entire supply chain and creating AI models that fit sales characteristics. We are trying to create a flexible system that delivers books that will definitely sell to bookstores that can sell them.

Another issue is the sluggish bookstore management. By attaching RFID tags, which have proven to be highly effective in the apparel distribution industry, to publications, and using the recorded data, we are promoting support for solving problems in publication distribution, mainly improving bookstore operations and management, from inventory efficiency to book recommendation services at sales floors, and even preventing shoplifting.

Turning Publishing Culture Knowledge into AI Models

However, the “use of AI” and “supply chain efficiency” cannot be achieved overnight. Until now, data has been siloed within publishers, distributors, and bookstores, and there has been a situation where data for the entire market is not visible. Even if we tried to conduct data analysis, it was necessary to collect data one by one at the person in charge level, and as a result, decision-making often did not go well.

Therefore, PubteX is first supporting the disclosure and sharing of data, mainly focusing on Shogakukan, Kodansha, and Shueisha. In data analysis and AI model development, we verified the visualization of bookstore sales data and demand forecasting at the individual bookstore level from various perspectives.

There are many challenges that must be overcome in building AI models. Until now, the knowledge of publishers has been utilized in deciding the number of copies published and the number of copies distributed, but there is also a lot of tacit knowledge that has not been clearly stated. As Marubeni and publishers repeatedly discussed, that knowledge was incorporated into logic, and the construction of AI models progressed.

Balancing Economic Rationality and Cultural Value

Thus, after data analysis and AI model construction, PubteX is currently working on demonstration experiments. We are conducting demonstrations of models created from data collected from three companies, and based on those numerical values, we are considering how to transform the existing operations of publishers.

Realizing collaboration that includes not only publishers and bookstores, but also distributors, is by no means easy. The amount of data that needs to be linked is enormous, and it will take time to change the entire publication distribution supply chain. There are cases where data itself is scarce because some bookstores only take inventory once a year, and it is true that data is still scattered. In addition, continuous research and development of AI models is important in order to adapt to the diverse sales characteristics of publications in detail.

In the publishing industry, creating interesting content is a competitive area, but improving the efficiency of the publication distribution supply chain can be said to be a collaborative area. PubteX’s efforts are also opening up areas of collaboration that publishers alone could not achieve. It was realized because Marubeni and Marubeni Forest Links have strengths in commercial distribution and physical distribution, and at the same time, they are advocating a new form of distribution while organizing publishers with different ideas.

There is no doubt that the know-how that Marubeni has cultivated in various industries and retail fields will be useful as the publishing industry shifts from the conventional push-type supply chain logistics to pull-type supply chain logistics optimized according to demand.

Among various businesses, publishing is also said to have strong cultural and social significance, as it is called the intellectual infrastructure. There is no model answer to the question of how to balance economic rationality and cultural value. PubteX’s practice may provide a new answer to this question through DX of the supply chain.