%201.svg){kind=icon}
Bandai Namco Nexus, a company with a mission to connect entertainment, products, and customers through data, is driving the "Data Universe Initiative" to utilize data across the Bandai Namco Group. While working on in-house development of their data analytics infrastructure, they faced challenges due to the large number of IPs, data sources, and affiliated group companies. To address this, they introduced TROCCO®, aiming to centralize data collection and management across the group, allowing non-data engineers to utilize the data efficiently. We spoke with the project lead to learn about the challenges, outcomes, and future prospects of this initiative.
Yuma Fujii, Bandai Namco Group: The company aimed to enhance both the quality and speed of data analysis by internalizing the data analytics infrastructure.
Fujii:
At Bandai Namco Group, we consider the "Data Universe Initiative" a key strategy. Our goal is to collect, integrate, and analyze diverse data from our entertainment businesses, such as games, e-commerce sites, and events, across different group companies to drive further business growth.
Previously, each group company collected its own data independently, leading to poor data integration across companies. Ideally, data should be utilized at the IP level (e.g., game titles, characters), but because different group companies handled different products and services, seamless data integration was not achieved.
Our vision is for the entire Bandai Namco Group to use data collectively and strategically to maximize fan engagement. Within this framework, Bandai Namco Nexus oversees the entire process, from data collection to analysis.
Fujii:
Before introducing TROCCO®, we did not have an in-house data engineering team and had outsourced the development and maintenance of our data analytics infrastructure. At that time, we had very few data engineers in-house, so a small number of data analysts and data scientists had to handle data analysis on their own.
Since the data infrastructure was externally managed, there was a long lead time from data extraction requests to obtaining insights. Additionally, communicating our company’s intent and requirements to external teams incurred high communication costs. This delay sometimes caused data analysis results to miss key event planning deadlines, affecting management decisions.
To address this, we decided to strengthen data engineering capabilities and bring the data analytics infrastructure in-house.
Fujii:
This initiative was the first project under the "Data Universe Initiative", and we decided to integrate data from the e-commerce sites operated by our group companies. These sites have large memberships and feature unique products that generate significant fan engagement. We believed that effective data utilization could lead to major business outcomes.
Given the large volume of data and the demand for quick analysis, we started considering ETL solutions to internalize data pipeline development.
The decision was based on its user-friendly GUI, allowing both data engineers and business teams to utilize it effectively.
Fujii:
We compared Google Cloud's managed services (which we were already using) with other ETL tools, including TROCCO®. At that time, since the "Data Universe Initiative" was still new and the future direction of data utilization projects was uncertain, we decided to start small with a cost-effective and easy-to-adopt solution.
Fujii:
We planned to extract data from Treasure Data, which consolidates e-commerce purchase data and logistics data. This data would then be transferred to Google BigQuery, and ultimately connected to BI tools like Looker and Looker Studio, as well as data analysis platforms like Jupyter Notebook.
Fujii:
The key reason was that TROCCO® provides a fully GUI-based interface, making it accessible even to non-engineers. Since hiring data engineers is challenging, we needed an ETL tool that could be used by analysts without requiring coding expertise.
Additionally, as our future goal is to integrate data from multiple group companies, TROCCO®'s wide range of connectors played a crucial role in our decision.
With low implementation and operational costs, we were able to secure internal approval smoothly and successfully introduce TROCCO®.
Transferring data from e-commerce sites and amusement facilities. The team function enables usage across group companies.
Fujii: We were able to build pipelines much more smoothly than expected. The setup is almost complete just by clicking through the process, so by the week after starting implementation, we already had one pipeline transferring data from our e-commerce site.
Additionally, we were pleased with how easily we could set up a pipeline to transfer app game data from the App Store Connect API to Google BigQuery, thanks to the prebuilt plugin. Since Bandai Namco Group operates many app games, we initially thought it would be challenging to set up each App Store and ad data source individually, but it turned out to be much easier than expected.
Even when error logs were difficult to read, we were able to resolve issues with support from primeNumber. When we contact them via Slack, we receive responses quickly, which has been very helpful.
Fujii: The data we use the most is from our e-commerce site, which was the main reason we introduced TROCCO®︎. Apart from e-commerce, various departments handling events and campaigns often request data for planning purposes.
Specifically, we consolidate and utilize information on which IP, character, and product are selling and to what extent, along with buyer attribute data and ad distribution data from Google and Facebook.
Additionally, we now utilize data from amusement facilities operated by our amusement business. All game play logs are retained, and we are currently replacing our system to transfer them to Google BigQuery.
Fujii: Since we handle data from multiple group companies, we have confidentiality obligations that prevent us from disclosing certain data or allowing certain operations outside specific group companies. This is where the team function of "TROCCO®︎" becomes invaluable.
This team function has two key concepts: teams and resource groups. It allows us to manage permissions and access at a team level for specific resource groups and the resources within them. Thanks to this feature, account issuance has become much smoother, and we feel reassured even as more group companies or external contractors become involved.
Fujii: We focused on creating manuals. Since we anticipated cases where external contractors would be involved or employees would rotate, we determined that a manual was necessary so that anyone could operate "TROCCO®︎".
Some group companies have unique security requirements that prohibit external data sharing. Therefore, we created a flowchart to determine when TROCCO®︎ should be used versus when an alternative method should be chosen. The manual also includes guidelines for operation.
Fujii: The biggest benefit is that we have successfully built an in-house data analytics platform while controlling implementation and operational costs.
In the early implementation phase, we were able to set everything up with only 1–2 person-months of effort. Now, in the operational phase, we are working with multiple external companies while maintaining operations without increasing the number of in-house data engineers. Currently, we have four full-time data engineers, but without TROCCO®︎, we would have needed at least twice as many, or eight engineers.
The cost savings have also allowed us to allocate more budget to security measures, such as protecting customer information.
Fujii: The lead time from designing to completing data transfer requests has significantly shortened. Previously, when external companies were involved, the process took about a month, but now we can complete similar requests in just two weeks.
Some requests that wouldn't have been feasible without TROCCO®︎ are now possible. As a result, analysts can start working with data much faster and feel more comfortable requesting data transfers.
We've also received positive feedback from analysts, such as:
If you're struggling with data platform development, replacement, or dashboard creation, feel free to reach out for consultation.
Fujii: As of October 2023, we have finally completed the first step. This means that we now have a stable environment where each group company can collect and analyze data independently.
For the second step, our goal is to establish a system where data can be utilized across the entire group. Ideally, this will enable group-wide decision-making based on data.
The third step, which is more of a personal vision, involves leveraging our analytical insights to expand into various areas, such as anime, merchandise, and games, in a data-driven manner.
Fujii: The Bandai Namco Group has many group companies and external collaborations, as well as a large number of products. If we relied solely on data engineers manually working on data utilization, we would hit a scalability limit due to resource constraints.
However, by implementing tools like TROCCO®︎, we can control operational costs while expanding data sources, making it easier to scale data utilization. We aim to use the insights gained from this process to provide value to our group companies.
Additionally, we would like to use TROCCO®︎ to integrate offline event-related marketing data, which is currently collected through analog methods.
Fujii: For companies that lack a large team of engineers or infrastructure specialists, "TROCCO®︎" is an excellent choice. For us, TROCCO®︎ has become the infrastructure for data utilization.
That’s why, when implementing it, it's crucial to clarify who will use it and how. Doing so will ensure that operations run smoothly after deployment.
%201.svg){kind=icon}
.webp)
