Shimamura Music Improves Business with Demand Forecasting and Automated Ordering Using Amazon Forecast

Originating as a small music school in Japan in 1962, Shimamura Music was founded in 1969 when it started selling musical instruments to students. With a mission to "create even one more person who can enjoy music," it currently has retail stores and music schools in 39 prefectures nationwide. Known as Japan's largest musical instrument retail store and one of Japan's leading music schools, the company provides support to musicians including repairing instruments, planning and holding events and concerts, and operating music studios.

Since 2002, the company prioritizes achieving the highest quality of services over expansion. To improve its services, it built a point of sale (POS) system in all stores to improve its inventory management system and increase the efficiency of its customer-led product lineup. Using an on-premises demand forecasting system, it strengthened its logistics to automate standard product orders, including products that were in low demand.

In 2020, the business environment changed suddenly and drastically due to the coronavirus pandemic. Stores and schools, mainly located in large commercial facilities, were closed and consumer demand for musical instruments changed. 

Standard product sales changed, sales of typically low-demand products increased, and the use of its online stores increased significantly. Since in-person music schools were impacted, online lessons suddenly came to the fore. Shimamura Music needed to respond rapidly and accurately to changes in its business environment by forecasting demand based on data, rather than relying solely on intuition and experience.

With the demand forecasting tool previously being an on-premises system, Shimamura Music knew that an upgrade was necessary. The license was expiring, and there was a high operating load on the server environment. When attempting to create a new service in response to future changes in business and demand, it decided to build a cloud-based system. 

Shiggy Nakane, executive corporate officer and general manager, Logistics Department at Shimamura Music, had heard about Amazon's inventory management and demand forecasting technologies, so he decided to try AWS. Shimamura Music already had experience using AWS in some of its systems, and had realized the benefits of building on the cloud. The team discovered Amazon Forecast, the time-series forecasting service, which is available for trial and has a lot of information including hands-on videos available. 

Using monthly forecasts from Amazon Forecast, Shimamura Music automates ordering for about 2,100 out of approximately 5,000 items. Product turnover remains at the same level as with the on-premises demand forecasting tool. The shortage rate improved by 0.7 percent over the span of 4 months despite an unstable supply due to the coronavirus pandemic. In addition, using Amazon Forecast reduces manual orders by 50 percent by expanding the number of products that are automatically ordered by 500 items. 

With the previous system, the team needed to manually check many items visually. This was burdensome as its staff had a large number of items to handle. Increased automation using Amazon Forecast significantly reduced the manual workload and improved the team's efficiency. The effect has been dramatic as the logistics department can now take on processing orders for other departments.

"Although we weren't completely familiar with AWS, we were still able to use Amazon Forecast to implement automated ordering. The team successfully migrated from our on-premises demand forecasting tool and databases. This allows them to continue working on our long-standing project to improve business. I've been impressed with how the system has evolved while using AWS," said Rumi Aoyagi, Logistics Division, Shimamura Music Co.

In fact, while testing Amazon Forecast, Aoyagi was worried about the complexity of the user interface (UI) and was wondering if an external vendor would need to be hired to develop the system. However, the update significantly improved the UI so that Aoyagi can use it without issue. At the same time, the predictive model has been improved to obtain more accurate results. Aoyagi was surprised at how quickly the UI improved using AWS. 

"The improvement was really good. We wanted to use the system as-is, since adding customized features could pose a problem when we need to update the system in the future. AWS offers a lot of videos describing the product and can give accurate answers to our questions. Originally, I wasn't even thinking about developing an in-house tool, but I was convinced that we could do it with the comprehensive support that AWS offers. Developing the system in-house has resulted in keeping costs down and made it possible to rapidly make improvements. In the future, I would like to share this experience with the team and work on reforming our business even further," said Aoyagi. 

Since building the cloud-based demand forecasting and automated ordering system, Shimamura Music now has the time to look at the data in detail. Using data frees the company from relying on intuition to handle manual orders and allows the implementation of high-quality logistics. In the future, the company plans to verify the resulting impact on inventory and ordering efficiency based on various data metrics and make further improvements based on the results.

The logistics department will continue working on improving the system using services and security features such as Amazon API Gateway (a fully managed service that makes it easy for developers to create, publish, maintain, monitor, and secure APIs at any scale) and Amazon QuickSight (a serverless, cloud-native business intelligence service that helps organizations understand data with interactive dashboards and more). Shimamura Music is also considering deploying a comprehensive logistics system on AWS.